Don't add include files to the archive. Avoids build failures for
Windows such as:
the input file 'libvpx_g.a(x86_abi_support.asm.o)' has no sections
Change-Id: If9c8e70c0ec913b7ad7dd6a08d4fa19011114ad2
No need to specify default behaviour. The original change introducing nasm:
7be093ea4d
mentions requiring 2.0.9, which was the first release to default to this behaviour:
http://www.nasm.us/doc/nasmdoc2.html
"The -Ox mode is recommended for most uses, and is the default since NASM 2.09."
Change-Id: Ia914c4deede5aa447277b5189bb4fcf7e54c338d
nasm does not accept x64
yasm has accepted (and appears to prefer) win64 at least as far back as
1.0.0:
http://yasm.tortall.net/releases/Release1.0.0.html
Change-Id: Ied881b1df0570da256b1bd7e131e7817e47f768f
Set num_inter_modes based on ref_mode_set_svc, which is
smaller set than ref_mode_set (which may use alt-ref).
No change in behavior.
Change-Id: I31169bb09028db230552c6fca0a86959d1ade692
1. Delete unnecessary zero setting process.
2. Optimize the method of calculating SSE in vpx_varianceWxH.
Change-Id: I58890c6a2ed1543379acb48e03e620c144f6515f
Avoids duplicate computation of UV predictor.
Bit-exact when static_threshold is zero.
Small/neutral difference on RTC set with nonzero static_threshold
(since UV predictor won't be skipped with this change).
Small speed gain, ~1-2%, at speed 8.
Change-Id: Iba8d22a307768b391e29d63c9826aac5a4d9c285
this is only meant for testing. along with --enable-experimental
--enable-spatial-svc require VPX_TEST_SPATIAL_SVC to be defined rather
than bumping the encoder ABI.
Change-Id: I7f34d9f60300fa31ccf22e1a4aa619392c391b2e
For 1 pass cbr SVC: GOLDEN is the spatial reference,
better not to check for encoder_breakout on this reference.
Small positive ~0.075% (mostly neutral) gain in avgPSNR/SSIM metrics.
No observed change in encoder speed.
Change-Id: Ib337f16d6771105bf06384c6a23ad047fc690418
For the case when the number of temporal layers > 1,
the buffer levels (starting/optimal_buffer_level,
and maximum_buffer_size) were not scaled properly.
In vp9_update_layer_context_change_config():
when setting the layer-buffer levels, fix is to scale
the layer-target_bandwidth by the target_bandwidth
(which is the full stream bandwidth) instead of the
spatial_layer_target.
This is needed because prior to the call
vp9_update_layer_context_change_config(), set_rc_buffer_sizes()
is called which sets the buffer levels based on target bandwidth
(which is the full bandwidth for the SVC stream).
This fix properly sets the layer-buffer levels based on the
layer-bandwidth, and leads to better rate targeting.
Small/neutral change in avgPSNR/SSIM metrics on RTC set.
Change-Id: Ic0f4f7f3487c37b9a9adb4781ae5edfed7140a57
Control Flow Integrity [1] indirect call checking verifies that function
pointers only call valid functions with a matching type signature. This
change eliminates function pointer casts to make libvpx CFI-safe.
[1] https://www.chromium.org/developers/testing/control-flow-integrity
Change-Id: I7e08522d195a43c88cda06fa20414426c8c4372c
For reference frames: enable scale partition for
superblocks with low source sad or if bsize on lower-resoln
is at least 32x32.
Keep feature disabled for base temporal layer.
Small regression in avgPNSR/SSIM metrics, ~0.5-1%.
Speedup ~2-3% on mac for SVC (3 spatial/3 temporal layers) at speed 7.
Change-Id: I5987eb7763845b680059128b538bb5188be0cca5
When allow_partition_search_skip is set the two pass code
can optionally skip the partition search in the rd loop if the image
appears static (based on selection of 0,0 motion).
Unfortunately 0,0 motion does not necessarily mean that there are
no meaningful changes or that motion or intra modes will not be selected
in the second pass.
Disabling "allow_partition_search_skip" may hurt the encode speed a little
for a small number of clips but can have a big impact on compression.
The most notable example of this in our test sets is "bridge_close_cif"
where this change gives a gains of 18%, 12% and 16% in opsnr, ssim and
psnr-hvs.
Change-Id: I765e288b5c0cd82bce00a148e7653a21e9203024
Enable partition copy on boundary and scale blocks along the boundary.
Rename copy_partition_svc to scale_partition_svc.
Do not copy if the block crosses the boundary.
Change-Id: I37a04d48f11b15c4ea67facd7631193ec2f62150
Fixes a build issue when relocation is not allowed:
relocation R_X86_64_32 against '.rodata' can not be used when making a shared object
Change-Id: Ica3e90c926847bc384e818d7854f0030f4d69aa0
Removal of parameters to and code in calc_frame_boost() that is no
longer required.
No change to results from previous patch.
Change-Id: Ic92da35613fdc247d22fddf24d09679fc5329017
The decay accumulator clause covers similar ground to the
new clause that tests the accumulated second reference error
so it has been removed to reduce complexity.
Change-Id: I4ec1cce32d72bd4ee463ad7def2831a68447d525
Add a clause to the breakout test for alt ref groups that
examines the size of the accumulated second reference
frame error compared to the cost of intra coding.
This clause causes a reduction in the average group length for many
clips. Alongside the change to the group length the minimum
boost is increased.
On balance the results are positive for psnr and psnr-hvs
but is negative for ssim/fast ssim for the smaller image formats.
Strong gains on some harder clips (eg ducks take off (midres) ~20%,
husky (lowres) 6-17%. Most of the negative cases are lower motion
clips. Subsequent patch hopefully will help with those.
Change-Id: Ic1f5dbb9153d5089e58b1540470e799f91a65dc4
Fix/cleaup the conditioning for usage of the reuse-lowres
partition feature.
Replace the non-reference condition with the top temporal
layer, and put this condition in the speed feature.
This prevents doing update_partition_svc() on every
VGA frame, instead it will now only do update for VGA in
the top temporal layer frames.
Also this makes it easier to test/enable this feature
for lower layer temporal frames.
Change-Id: Ia897afbc6fe5c84c5693e310bcaa6a87ce017be5
For new VP9 only content type adjust the rate distortion and ARF
filter based on the relative spatial variance of the source and
reconstruction.
In regards to the RD loop the method favors modes where the
reconstruction variance is similar to the source variance. However it
is currently only applied to regions where the source variance is quite
low.
For very low variance blocks it applies a further bias against intra
coding and large prediction block sizes (the later in particular limit
the usefulness of the loop filter).
The final part of this change is to lower the strength of the ARF
filter for blocks where the source has very low spatial variance, to
encourage some low amplitude texture or noise to pass through
the filter.
This change improves the retention of film grain and fine noise /
texture in spatially flat regions, but as expected causes a significant
drop in PSNR on many clips. This is to be expected because similar
but misaligned noise or texture will give a lower PSNR than a flat
noise free reconstruction. However, it is worth noting that most clips
show a strong gain in FAST SSIM.
The features are enabled on the vpxenc command line by setting
--tune-content=film.
VPX_ENCODER_ABI_VERSION bumped for this change and cvbr.
Change-Id: I26a4e4edfa3dc5cacead82fa701fe7a9118ccd0a
Removed three parameters that are no longer needed in calls
to calc_arf_boost() and associated minor changes.
No impact on encode results.
Change-Id: Ieaf31d0d2e1990b99cf69647170145a1bbfbb9fb
For choose_partitioning (speed >= 6): avoid computation
of minmax variance for non-reference frames in SVC.
Existing condition only avoided this for speed >= 8.
Combine that existing logic with non-reference condition.
Small speedup (~0.5-1%) for 3 layer SVC,
neutral change on avgPSNR/SSIM metrics.
Change-Id: I3e9f3a1af0647b15e475cf170d9402908d672ee5
Release frame buffers for non-ref when the decoder is destroyed.
Enable the non ref test.
BUG=b/68819248
Change-Id: Id87ef3b0a62318f9812e927cd957c05c859047fa
For SVC with 3 spatial layers:
Add feature to copy/upscale partition from middle spatial layer
to the upper/highest resolution, when superblock sad is not high.
Enabled for speed >= 7 and only for non-reference frames.
Speedup ~3-4%, small loss in avgPNSR/SSIM of ~1%.
Change-Id: I7f0a2716c0fde28bade0f86159d11b7e31d6ab8d
For a chosen interval "i" the existing arf boost calculation examined frames
+/- (i-1) frames from the current location in the second pass.
This change checks to make sure that the forward search does not extend
beyond the next key frame in the event that the distance to the next key
frame is < (i - 1).
Small metrics gains on all our test sets but these are localized to a few clips
(e.g. midres set psnr-hvs sintel -2.59% but overall average was only -0.185%)
Change-Id: I26fc9ce582b6d58fa1113a238395e12ad3123cf6
The new test will run a SVC bitstream which has non ref frames.
It checks the number of buffer acquired and released to make sure all
external frame buffers are released.
Add a new test bitstream:
vp90-2-22-svc_1280x720_1.webm
which has 400 frames in total, and 1 spatial layer and 2 temporal layers.
There is one non ref frame every other frame.
Disabled for now. Will be enabled with the fix.
BUG=b/68819248
Change-Id: I0515336fd9809a9e1fceba90e4dce53dabaf53a5
Added command line control of Corpus VBR.
The new corpus vbr mode is a variant of standard
VBR (end-usage=0) where the complexity distribution
mid point is passed in rather than calculated for a specific
clip or chunk.
The new variant is enabled by setting a new command line
parameter --corpus-complexity to a zero value. Omitting
this parameter or setting it to 0 will cause the codec to use
standard vbr mode.
The correct value for a given corpus needs to be derived
experimentally using a training set such that the average
rate for the corpus is close to the target value.
For example our using our low res test set with upper and lower
vbr limits of 50%-150% and a corpus complexity value of 650
gives a similar average data rate across the set to using standard
vbr. However, with the corpus mode easier clips will be allocated
fewer bits and harder clips more bits rather than having the same
rate target for all.
Change-Id: I03f0fc8c6fb0ee32dc03720fea6a3f1949118589
For nonrd_pickmode: if early_term is set there should be
no need to include UV in rdcost (when color_sensitivity is set).
Neutral change on RTC and RTC_derf metrics, for speed >= 5.
No change for ytlive metrics.
Very small speed gain (~0.5%) on some clips with strong color content.
Change-Id: Ifc00928ecd935fc71e94935ceef0ae7481249f07
Allow for compound prediction mode in nonrd_pickmode for ZEROMV.
For real-time encoding, 1 pass with non-zero lag-in-frames.
Added speed feature to control the feature.
Enabled for speed >=6 for now, under VBR mode.
avgPSNR/SSIM metrics positive on ytlive set, for speed 6:
some clips up by ~3-5%, some clips neutral gain, average gain
across clips is ~1%.
Small/negligible decrease in speed.
Change-Id: I7a60c7596e69b9a928410c5ee2f9141eecd8613d
Even though frame_size is calculated in uint64_t, it winds up in an int
size value.
This was exposed with the msan test because the memset is called with
(int)frame_size, leading to a segfault.
Change-Id: I7fd930360dca274adb8f3e43e5e6785204808861
Change type of sum_square_error from int64_t to uint32_t.
Change type of sum_error from int64_t to int32_t.
This reduces the stack usage from ~131K to ~87K.
BUG=b/68362457
Change-Id: I147d7c7b226bceb4f0817bb86848e1fa9d9ac149
swap '{' and c-style comments removing a few redundant ones along the
way; covers most leftovers from the clang-tidy run against an
x86_64-linux config.
Change-Id: I67a45596f80a12389faca49c5be440875092a7df
Changed the intrinsics to perform summation similiar to the way the assembly does.
The new code diverges from the assembly by preferring unsaturated additions.
Results for haswell
SSSE3
Horiz/Vert Size Speedup
Horiz x4 ~32%
Horiz x8 ~6%
Vert x8 ~4%
AVX2
Horiz/Vert Size Speedup
Horiz x16 ~16%
Vert x16 ~14%
BUG=webm:1471
Change-Id: I7ad98ea688c904b1ba324adf8eb977873c8b8668
Set adaptive_row_thresh_mt = 1 at speed >= 7,
for svc when multi-threading is used with row-mt.
This allow the adaptive_rd_thresh feature to be used
in the nonrd-pickmode.
~1-2% speedup for SVC encoding with small quality
loss (< 0.6%) on RTC set.
Change-Id: Iab9878dff117bccdaef3e4d0645165db9808cdfc
Disable cyclic refresh if ROI is used and add flag to properly handle
the static_thresh deltas.
Remove the ROI test for cyclic refresh (it's allowed but disabled if ROI
is used).
Add an example in vpx_temporal_svc_encoder.c. Turned off by default.
BUG=webm:1470
Change-Id: Ief9ba1d7f967bc00511b412b491c3f70943bfbda
Note this change will trigger the different C version on SSSE3 and
generate different scaled output.
Its speed is 2x compared with the version calling vpx_scaled_2d_ssse3().
Change-Id: I17fff122cd0a5ac8aa451d84daa606582da8e194
Small inncrease the sad_thresh1, avoids some false
detection of possible scene changes within lag.
Small improvement in few clips on ytlive, otherwise neutral change.
Change-Id: Ia79b53bb657bbce65a7aac7d20666b6373d5af8b
Expose the threshold for setting key frame on cut,
and increase it for speed 5.
Also small adjustment to min_thresh.
No change in overall metrics or fps.
Small quality improvement and lower encode time on scene cuts.
Change-Id: I36e06ff3b26b6c29aede39c23fce454525fc9026
Small increase in threshold for the 1 pass VBR datarate tests.
Needed due to commit:
<017257a Adjustment to scene detection and key frame>
Change-Id: I28b3bd7db2192a8cc2bccc3cb0e3b8dbb910ca16
The initial allocation of bits in the two pass code to each frame
should be within the min max limits on the command line. However,
when forming an ARF group the cost of the ARF is shared by frames
in that group such that the residual bits for a frame could drop below
the min value. This change prevents the minimum being re-applied
after the cost of the ARF has been deducted as this may otherwise
cause low rate sections to overshoot their target.
Test runs comparing to a baseline run with min and max section pct
0-2000% vs one closer to the YT use case (50-150%) suggest that
this fix not only results in better rate control but also gives a better
rd outcome.
For example the HD set vs 0-2000% baseline (opsnr, ssim).
Old code (50-150): +0.751, +1.099
New code(50-150): +0.241, -0.009
Change-Id: I715da7b130bf53ba8aa609532aa9e18b84f5e2ef
Let it test extreme inputs and all filter types.
In the future ConvolveTest should test regular 8-bit functions in
high bitdepth mode.
Change-Id: I1042564d1d390589ca203070fe332c6da3315d75
For 1 pass vbr: use higher threshold on avg_sad
and force key frame under scene cut detection if
above the threshold. Allow it for speed >= 6 for now,
since it does not use the full nonrd_pickmode partition
(as in speed 5).
Improves quality somewhat on scene cut frames.
Neutral on overall metrics and fps for speed 6 on
ytlive set.
Change-Id: I12626f7627419ca14f9d0d249df86c7104438162
Change to the bit allocation within a GF/ARF group.
Normal VBR and CQ mode allocate bits to a GF/ARF group based of the mean
complexity score of the frames in that group but then share bits evenly between
the "normal" frames in that group regardless of the individual frame complexity
scores (with the exception of the middle and last frames).
This patch alters the behavior for the experimental "Corpus VBR" mode such that
the allocation is always based on the individual complexity scores.
Change-Id: I5045a143eadeb452302886cc5ccffd0906b75708
This patch makes further changes to support an experimental
corpus wide VBR mode that uses a corpus complexity
number as the midpoint of the distribution used to allocate bits
within a clip, rather than some average error score derived from the
clip itself.
At the moment the midpoint number is hard wired for testing and
the mode is enabled or disabled through a #ifdef. Ultimately this
would need to be controlled by command line parameters.
Change-Id: I9383b76ac9fc646eb35a5d2c5b7d8bc645bfa873
vpx_convolve8_avg works by first running a normal horizontal filter then a
vertical filter averages at the end.
The added vpx_convolve8_avg_avx2 calls pre-existing AVX2 code for the
horizontal step.
vpx_convolve8_avg_vert_avx2 is also added, but only uses ssse3 code.
Change-Id: If5160c0c8e778e10de61ee9bf42ee4be5975c983
This reverts commit 9311ef18b4.
Reason for revert:
Notice small regression in some clips.
Will revisit in another change.
Original change's description:
> Speed >=5 real-time: add TM intra mode for high_source_sad.
>
> Small/neutral change in metrics or speed for ytlive.
> Some improvement in quality on frames with big content change.
>
> Change-Id: Ib3b0703a5f28ea6710e90324436e27598ab7384d
TBR=marpan@google.com,builds@webmproject.org,jianj@google.com
Change-Id: I9d8ec5195bb05ddf329d325699355185affb9b13
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
For 1 pass vbr: increase min_thresh slightly, and also add
condition on golden/arf update for using full nonrd_pick_partition.
Reduces possible false detection for scene cut detection.
Neutral/small change in metrics or speed for speed 5.
Change-Id: I388f4d9a56e3cc763e0148338c1bc0381e58ad76
Small/neutral change in metrics or speed for ytlive.
Some improvement in quality on frames with big content change.
Change-Id: Ib3b0703a5f28ea6710e90324436e27598ab7384d
Lower SAD threshold to select non_rd pickmode partition
at superblock level more often.
Small gain in metrics, small/negligible decrease in speed.
Change-Id: I0f728236b91a604e4ca7e02039adc54d5985c4dc
For 1 pass vbr speed >= 6: when REFERENCE_PARTITION is selected,
avoid doing the full nonrd_pickmode based partition.
No change in overall metrics or speed.
Reduces encode times on scene cuts by 10-20%.
Change-Id: I0310b1610cc1c83793a509e0a9059840e8f18308
For 1 pass vbr mode:
On no-show_frame/ARF: instead of skipping alt_ref_frame
completely in mode testing, allow for checking (0, 0) on alt_ref.
Small gain in metrics, ~0.18%, no change in speed.
Change-Id: I32a3c24faca64ab70dd5091071a0dc301db7dd1e
For 1 pass vbr: when significant content/scene change is detected
(high_source_sad = 1) reduce/turnoff the additional qdelta on the
active_worst_quality. This helps somewhat to reduce the occurrence
of large frame sizes and large encode times.
Allow it only when use_altef_onepass is enabled.
Neutral/no change on metrics.
Change-Id: I1dd97dd2ab892d65f707b841b27a5de300b714ea
For speed 6 real-time mode: use adapt_partition
on ARF frame instead of REFERENCE_PARTITION (which is slower).
This requires enabling compute_source_sad_onepass for no-show_frames.
Speedup of ~3-5% on some clips that heavily use ARF,
small loss (~0.2%) in quality on ytlive set.
Change-Id: Ib50acc97df06458244a6ac55d2bd882c30012536
Speed comparing with the one calling vpx_scaled_2d_neon()
~1.7 x in general
~2.8x for BILINEAR filter
BUG=webm:1419
Change-Id: I8f0a54c2013e61ea086033010f97c19ecf47c7c6
Scale 3x3 block instead of 16x16 block in each loop. Disabled by
default.
Benefits:
1. Reduced number of different phase_scaler from 16 to 3.
Optimization code will be smaller and faster.
2. Maximum phase_scaler drifting will be reduced from 5/16 to 1/24.
(The drifting is 1/(3*16) in each step.)
BUG=webm:1419
Change-Id: I59a1f7496d89a1b090498c935d30cfcf1d0c282b
For real-time mode. Move the switch to fixed partition
for is_src_frame_alt_ref so all speeds may use it
if use_altref_onepass is set.
Improves metrics by ~2% for ytlive set at speed 4
(where use_altref_onepass is currently used).
Change-Id: I033240386598c9dbd0364da89ccbcca64bc663ee
Only has effect when sf->use_altref_onepass is enabled,
as in that case scene detection is skipped for non-show frame
and so high_source_sad does not get reset to 0.
No change in metrics or speed.
Change-Id: I421f066d239341449c18826089e1810b9fc5967f
Add stats for past ARF usage, and use it to disable
ARF usage based on some conditions.
Overall improvement on ytlive set, reduces the regression
on the problem clips for this feature.
Only affects when sf->use_altref_onepass is enabled
(currently off by default).
Change-Id: I66267f227ea132dc86acb730e9882f85bead2cdb
This reverts commit 535b7b915a.
This is actually used in CBR to reset the rate control if high source sad is detected.
Original change's description:
> Remove the speed condition on scene detection in 1 pass code.
>
> Scene detection is used for VBR mode and for screen_content mode.
>
> It was also enabled for CBR mode via the speed condition,
> but currently the analysis in the scene detection is not used
> in CRB mode (similar computations are done locally at superblock level
> when the source_sad feature is enabled).
>
> For 1 pass code.
> No change in behavior. Small speed gain, ~0.5%.
>
> Change-Id: I59991d7ef2af320bea7af4b907596e057affa42f
TBR=marpan@google.com,builds@webmproject.org,jianj@google.com
Change-Id: Ib4e6b02047f75632503e7b0fc870af97fa9291c3
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Scene detection is used for VBR mode and for screen_content mode.
It was also enabled for CBR mode via the speed condition,
but currently the analysis in the scene detection is not used
in CRB mode (similar computations are done locally at superblock level
when the source_sad feature is enabled).
For 1 pass code.
No change in behavior. Small speed gain, ~0.5%.
Change-Id: I59991d7ef2af320bea7af4b907596e057affa42f
'iter' parameter is being checked for NULL in every call to
decoder_get_frame which is quite pointless because it is always
going to be NULL unless the application changed it. The code works
as described only because vp9_get_raw_frame returns -1 on all
subsequent calls after the first.
Change-Id: Ic736b9e8fe36fc1430fc11d6a9b292be02497248
* changes:
Remove the unnecessary cast of (int16_t)cospi_{1...31}_64
Remove the unnecessary upcasts of (int)cospi_{1...31}_64
Change cospi_{1...31}_64 from tran_high_t to tran_coef_t
Add the condition frames_since_golden > 0 to the
early exit check for ARF usage in nonrd_pickmode.
This improves quality of first frame following ARF, where
frame_since_golden = 0.
Small/neutral gain in metrics for speed 6, neutral change in speed.
Only affects when USE_ALTREF_FOR_ONE_PASS is enabled.
Change-Id: I82e73e6ff6fc849e5ca5448563cb8a0515fe0cdc
A new bug was introduced in a80bdfd "Change sinpi_{1,2,3,4}_9 from
tran_high_t to int16_t". Reverted the change in this file.
BUG=webm:1450
Failed test C/TransHT.AccuracyCheck/26.
Change-Id: Id001f57aad811803ef7d367d2b2bc008d8499991
Modify simple_block_yrd condition in nonrd_pickmode for SVC:
allow it to be used also on base temporal_layer, only when
spatial_layer > 1 and block size < 32x32.
Speed up of about ~2% for 3 layer SVC, with little/negligible
loss in quality.
Change-Id: I7734bdae51cf51f22b96f6b2b27da20ea1d84344
Fix the setting to frames_till_gf_update_due, and
adjust the limit value.
Only affects when USE_ALTREF_FOR_ONE_PASS is enabled.
Neutral change to metrics and speed for ytlive.
Change-Id: I266d9a00b36221bc8602fa2746d4e8a8f7d4dfae
Only when USE_ALT_REF_ONE_PASS is enabled (off by default).
Force fixed partition to 64x64 when is_src_alt_ref_frame is true,
and don't force early exit for some modes in nonrd_pickmode
for ARF noshow frames.
Small gain ~0.2% on ytlive metrics for speed 6.
Neutral speed difference.
Change-Id: I27eb6622d0453c09a06ccdc3b16368762474d11d
Add datarate test, for both VBR and CBR mode, with the
frame_parallel_decoding mode disabled (and error_resilience off).
Change-Id: I54feec3248a68ecff4bef8d9a31bb1616fab77df
In the new AUTO mode, restrict the minimum alt-ref interval and max column
tiles adaptively based on picture size, while not applying any rate control
constraints.
This mode aims to produce encodings that fit into levels corresponding to
the source picture size, with minimum compression quality lost. However, the
bitstream is not guaranteed to be level compatible, e.g., the average bitrate
may exceed level limit.
BUG=b/64451920
Change-Id: I02080b169cbbef4ab2e08c0df4697ce894aad83c
Removed inline for GP load-store in case of (__mips_isa_rev >= 6)
Created one define LD_V for vector load and ST_V for vector store
Change-Id: Ifec3570fa18346e39791b0dd622892e5c18bd448
Also add column headings so that the output can still be parsed if the
set of headers changes later.
Change-Id: I4beaf266521e093db4acf5f715b18fdfb7e3d1cd
This reverts commit 8c42237bb2.
Because ssse3 code is used for the reference, the qcoeff and dqcoeff
reference buffers must be aligned.
Original change's description:
> quantize avx: copy 32x32 implementation
>
> Ensure avx and ssse3 stay in sync by testing them against each other.
>
> Change-Id: I699f3b48785c83260825402d7826231f475f697c
Change-Id: Ieeef11b9406964194028b0d81d84bcb63296ae06
Scale 3x3 block instead of 16x16 block in each loop.
Benefits:
1. Reduced number of different phase_scaler from 16 to 3. Optimization code
will be smaller and faster.
2. The maximum phase_scaler drifting will be reduced from 5/16 to 1/24.
(The drifting is 1/(3*16) in each step.)
BUG=webm:1419
Change-Id: Ibb9242a629ddb03e1ff93b859bece738255e698c
The intra mode rd penalty was implemented as a rate penalty.
Code was added to scale the penalty according to block size but
this was not done correctly for the SB level or sub 8x8.
The code did a weird double scaling in regard to bit depth that
has been removed. Given that it is a rate penalty the bit depth
should not matter.
This bug fix improves average metrics on our standard test
sets by about 0.1%
Change-Id: I7cf81b66aad0cda389fe234f47beba01c7493b1e
Move class VpxScaleBase to new file test/vpx_scale_test.h.
Add new file test/vp9_scale_test.cc with ScaleFrameTest.
BUG=webm:1419
Change-Id: Iec2098eafcef99b94047de525e5da47bcab519c1
This header doesn't build on g++ v6 as it's a C and not C++ header
(_Atomic is not a keyword in C++11). Since the C and C++ invocations
cannot be guaranteed to point to the same underlying atomic_int
implementation, remove support for them and use compiler intrinsics
instead.
BUG=webm:1461
Change-Id: Ie1cd6759c258042efc87f51f036b9aa53e4ea9d5
Makes main thread wait for the filter level to be picked to avoid a race
between the LPF thread and update_reference_frames(). This also
re-enables the failing tests under thread_sanitizer where this data race
was detected.
BUG=webm:1460
Change-Id: I7f5797142ea0200394309842ce3e91a480be4fbc
Fixes issue on iPad Pro 10.5 (and probably other places) where threads
are not properly synchronized. On x86 this data race was benign as load
and store instructions are atomic, they were being atomic in practice as
the program hasn't been observed to be miscompiled.
Such guarantees are not made outside x86, and real problems manifested
where libvpx reliably reproduced a broken bitstream for even just the
initial keyframe. This was detected in WebRTC where this device started
using multithreading (as its CPU count is higher than earlier devices,
where the problem did not manifest as single-threading was used in
practice).
This issue was not detected under thread-sanitizer bots as mutexes were
conditionally used under this platform to simulate the protected read
and write semantics that were in practice provided on x86 platforms.
This change also removes several mutexes, so encoder/decoder state is
lighter-weight after this change and we do not need to initialize so
many mutexes (this was done even on non-thread-sanitizer platforms where
they were unused).
Change-Id: If41fcb0d99944f7bbc8ec40877cdc34d672ae72a
Neutral on rtc set for speed 8. Neutral on ytlive for speed 5.
Saves some computation cycles but no speed gain observed on Pixel.
Change-Id: I34c4642cd543aa89c5b9c4bff6b7113577c64c91
This reverts commit df9ce12259.
Reason for revert:
Re-enabled tests still fail tsan in high bitdepth.
Original change's description:
> Re-enable disabled tests under TSan.
>
> These tests point to an already-fixed bug, this should no longer have a
> data race.
>
> BUG=webm:1049
>
> Change-Id: Iaedc5db8df99362bdc501b70ff7fdebf8756fdb8
TBR=jzern@google.com,pbos@chromium.org,builds@webmproject.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: webm:1049
Change-Id: I232f1f7726bf795b301abfb2e07cad6756642e53
Rev d147771 fixed the test failure. So remove the resolution condition
for using source_sad in speed 6.
BUG=webm:1452
Change-Id: I1efba97e1ef5bd4de5f886299f6fcb907187abcd
Enable adapt_partition for vbr mode for speed 6.
This allows the usage of the pickmode-based partition
(used in speed 5), but only selectively for superblocks
with high source sad, otherwise the faster variance based
partition scheme is used.
For speed 6 on ytlive set: avgPSNR/SSIM metrics up by ~0.6%,
several clips up by ~1.5%. Small/negligible decrease in speed.
Change-Id: I12f3efef6b3e059391de330fdbe5a44c2587f1f8
For SVC at speed >= 7: only use the improved mv search
on base spatial layer, if top layer resolution is above 640x360.
~2.3% speedup
Small/negligible loss in avgPSNR metrics on rtc set.
Change-Id: Iaef75a57ebf1c248931bc1aa28d20b7fecac1851
This reverts commit f60d1dcd3d.
Reason for revert: <INSERT REASONING HERE>
Failures in AVX/VP9QuantizeTest in nightly tests.
Original change's description:
> quantize avx: copy 32x32 implementation
>
> Ensure avx and ssse3 stay in sync by testing them against each other.
>
> Change-Id: I699f3b48785c83260825402d7826231f475f697c
TBR=slavarnway@google.com,johannkoenig@google.com,builds@webmproject.org
Change-Id: Ibd38636212269328317dd0721be9d25452113d1c
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
For speeds < 7, increase threshold that controls the split
of 16x16->8x8 blocks, for resolutions 720p and higher.
Minor change for speed 5 (since it uses reference partition scheme
which only uses variance partition as first step).
For speed 6: ~0.5% increase in avgPSNR/SSIM metrics on ytlvie set.
No change in speed.
Change-Id: I5126580973201538d8ca26a9256b93c4d11d685b
Still does not pass tests. Does match the previous assembly, although
saving the sign before multiplying is dubious.
Change-Id: Ia163f18c755aba542d6e93f7bf7343184660df5a
Adds an early exit based on ptest. Slightly slower than ssse3 in the
full case because of the extra check, but potentially faster if lots of
rows can be skipped.
Very close in speed to the assembly.
Can run in 32 bit, unlike the assembly. Allows reworking the function
prototype to use structs.
Change-Id: If80e2b9ba059370a4cad3c973196e82a97b4330e
Add 1 if negative to get dqcoeff to round towards zero.
10-15% faster than converting to positive before shifting.
Change-Id: I01a62fd0c9bca786b6885b318bd447bb9229903d
About 4x faster when values are below the dequant threshold and 10x
faster if everything needs to be calculated.
Both numbers would improve if the division for dqcoeff could be
simplified.
BUG=webm:1426
Change-Id: I8da67c1f3fcb4abed8751990c1afe00bc841f4b2
This feature is used for the CBR RTC encoding mode
at speed >= 6. This change will exclude it for VBR mode.
For speed 6 live encoding (VBR):
avgPSNR/SSIM metrics on ytlive set up by ~1% (few clips up by 2/3%).
No change in speed.
Change-Id: I1a0dd94c334f7df309ab5a48d477d7e25355b798
* changes:
quantize: ignore skip_block in arm
quantize: ignore skip_block in x86
quantize fp: ignore skip_block in arm
quantize fp: ignore skip_block in x86
This should probably be handled before vp9_regular_quantize_b_4x4 even
gets called.
Fixes an assert resulting from removing skip_block from the quantize
functions.
BUG=webm:1459
Change-Id: I7f52b53f959b4654b3d4517ebda31a678f4d0fde
This condition is handled before this code is reached. The ssse3 version
of the function has always crashed when attempting to handle the
skip_block condition.
Add assert() and comments regarding the usage of skip_block.
Removing the parameter is a fairly involved process so leave it be for
the moment.
Change-Id: Ib299f6fc6589d7ee102262cc74a7aeb60110bc5a
Despite abs_coeff being a positive value, all the other implementations
treat it as signed which simplifies restoring the sign.
HBD builds cast qcoeff to avoid a visual studio warning. Match
vp9_quantize.c style of casting the entire expression.
Change-Id: I62b539b8df05364df3d7644311e325288da7c5b5
Having a very low "lag_in_frames" value could cause the encoder to create
incorrect / corrupt ARF groups including displayed frames that update the
ARF buffer and false overlay frames that are coded at low rate but are not
actually overlays of a real ARF frame.
This is linked to a reported unit test "slow down" where the chosen parameters
(lag of 3 frames) gave rise to such "broken" ARF group(s).
See also BUG=webm:1454
Change-Id: If52d0236243ed5552537d1ea9ed3fed8c867232c
Having a very small value for "lag_in_frames" can result in
corrupt arf groups including displayed frames that update
the arf buffer and fake overlay frames that are not in fact
overlays of real arfs but are nevertheless starved of bits.
Leaving lag_in_frames at the default of 25 for these 5 frame two
pass VBR tests should now give rise to a valid ARF coding pattern
as follows:- K(ey), A(rf), N(ormal), N, N, O(verlay).
This change is part of a response to BUG=webm:1454 where broken
arf groups interacted badly with a change that corrects for large rate
misses. However, it may still in some cases increase encode time by
virtue of the fact that the unit test now codes a correct coding pattern
with "hidden" ARF frames.
Change-Id: Ifd0246a4c1d0be247247c754024d7a4ed5f66a6b
Some clips in nightly unit test exhibiting significant encoder slowdown which
appears to bisect to Change-Id: I692311a709ccdb6003e705103de9d05b59bf840a.
The above change allowed for emergency iterations of the recode loop and
adjustment of the Q range if there is a large rate miss.
This patch disables the above adaptation for cases of cpu_speed >= 3 or more
specifically where cpi->sf.recode_loop >= ALLOW_RECODE_KFARFGF.
For speeds >= 3 the code does not currently run a dummy bit pack operation
inside the recode loop. Without this dummy pack operation there is no up to
date estimate of the current frame's size to use as a basis for assessing the
requirement for a recode. In practice it was using the previous frames size (or 0
for the first frame) which could cause odd behavior.
If we require the emergency rate correction added in Change-Id: I6923.. for
the higher speed settings it will be necessary to enable the dummy pack
which will in turn hurt encode speed.
BUG=webm:1454
Change-Id: I4fb3c6062ca9508325a6f31582f8e80f1a9b126f
Change legacy vp8/9_write_yuv_frame to vpx_write_yuv_files.
Delete some flags that can be enabled during build.
To enable writing denoised YUV, use the following command line:
CFLAGS='-DOUTPUT_YUV_DENOISED' ./configure
--enable-vp9-temporal-denoising
For skinmap, use CFLAGS='-DOUTPUT_YUV_SKINMAP'
Change-Id: I236974ac8b3cf279d20c4dc7f6162d8b480b6528
The result of the xor operation is unsigned. If coeff was negative,
this results in an unsigned value - INT_MIN.
Change-Id: I1f1edeaa6de1f4c68b848e8a82a666d390b749f0
Actual frame size and bitrate is all 0 when using SVC sample encoder
with sl = 1 because the stats are set in parse_superframe_index which
will not caculate properly when sl = 1 since there is no superframe.
Use pkt->data.frame.sz instead when sl = 1.
Change-Id: I93f5e98a4c779e32b007e1564ba5396af9e34ad6
Use input with a narrow range because the filter only applies when the
frames are similar.
Run CompareReferenceRandom more times. Especially before narrowing the
input range, the filter frequently did not apply.
Change-Id: Ie249bedf6d0d33dfa5884611cb1835788e418b38
this test fails with the configuration similar to the assembly prior to:
d52cb5972 quantize: copy ssse3 optimizations to intrinsics
BUG=webm:1458
Change-Id: Idc5c0b84c0598259fc49609a9f0756de531d3baf
Change the denoiser frame buffer management for SVC to more generally
handle the layer patterns in SVC (where last is not always refreshed).
This change is only for SVC with denoising and is bitexact.
Change-Id: Ic2b146a924cdf6e7114609158afa3d4880fe3fae
Testing of 4k videos encoded with a fixed arbitrary chunking interval
uncovered a bug where by if a chunk ends 1 frame before a real scene cut,
the next chunk may be encoded with two consecutive key frames at the start
with the first being assigned 0 bits.
This fix insures that where there is a key frame group of length 1 it is
at least assigned 1 frames worth of bits not 0.
See also patch Change-Id: I692311a709ccdb6003e705103de9d05b59bf840a
which by virtue of allowing fast adaptation of Q made this bug more visible.
BUG=webm:1456
Change-Id: Ic9e016cb66d489b829412052273238975dc6f6ab
Created inline functions highbd_butterfly_cospi16_sse2()
and highbd_butterfly_cospi16_sse4_1()
BUG=webm:1412
Change-Id: Icbc53a73712b6207379872a5e88d0a4d09e2322a
With skip block the neon is about twice as fast as C.
The neon has no shortcut for coeff < zbin so it always takes the
same amount of time. Even if the C can take the shortcut, it is over
twice as fast in neon. If it can't, that gap increases to over 10x.
BUG=webm:1426
Change-Id: I400722146c1b5a5f6289f67d85fd642463d2bfc6
Fairly minor differences from sse2. pabsw and psignw are the big gains.
Also re-uses some values in eob calculation to avoid an extra pcmp.
Fixes test failures in HBD and OS X builds.
Allows using it in 32bit builds, where it is about 40% faster than sse2.
Substantially faster than the assembly for skip_block. 10-20% faster the
rest of the time.
Change-Id: If783bb3567e561e47667e10133b9c84414a334e2
When adapt_partition_source_sad is enabled (currently only at
speed 6 for resoln <= 360p): use lower subsize (8x8 instead of 16x16)
for nonrd_select_partition on 32X32 blocks.
And force avoiding rectangular partition checks in
nonrd_pick_partition for speed >= 6.
Small increase ~0.5 in metrics for speed 6 on rtc_derf,
no change in speed.
Change-Id: Id751bc8f7573634571b2d6f5e29627cd5cebccae
Prepare for high bitdepth 16x16 idct sse4.1 code.
Just functions moving and renaming.
BUG=webm:1412
Change-Id: Ie056fe4494b1f299491968beadcef990e2ab714a
vpx_sub_pixel_variance32xh_avx2() and
vpx_sub_pixel_avg_variance32xh_avx2
see:
17fae3a Change to use correct check for halfpel
Change-Id: Ib0741c5c2fd011e9650ca62b76009f1b59fdbe4c
Enable fast adaptation of Q when there is a large overshoot
for the #ifdef AGGRESSIVE_VBR test case.
AGGRESSIVE_VBR is not currently enabled by default.
Change-Id: I7240bb6589795964b6b0b66df4468e4f21504e0f
Originally, for the purpose of keeping a fast first pass, the first-pass
stats between row_mt_mode = 0 and row_mt_mode = 1 are not bit exact, but
that difference is very small that doesn't cause a mismatch between the
final bitstreams. However, if the encoder changes, this minor difference
may cause a mismatch. Thus, this patch always forces the first pass to
be bit exact.
BUG=webm:1453
Change-Id: I2b67cf529dee81f660f9d9e7fe9a60ea3c7b12b8
For 1 pass CBR mode:
Apply the logic for dropping (and re-adjusting rate control)
due to large overshoot to the case of non-screen content when
drop_frames_allowed is enabled.
For the non-screen content case: add additional condition that
rate correction factor is close to minimum state, and flag to
constrain the frequency of the dropping.
Also handle the case of temporal layers and multi-res encoding.
Add some flags/counters to the layer context for temporal layers.
For multi-res: drop due to overshoot is checked on lowest stream,
and if overshoot is detected we force drops on all upper streams
for that frame.
This feature is to avoid large frame sizes on big content
changes following low content period.
No change in behavior for screen_content_mode = 2.
Change-Id: I797ab236cbbf3b15cad439e9a227fbebced632e6
This replaces commit aa1c4cd, which has a bug and was reverted in
commit 3c73e58.
The bug is caused by rounding -step1[5] in highbd_idct8x8_12_half1d().
Change-Id: I37b3a5f0d91815f2dc570209091dc6626fd178a8
With skip block or coeff < zbin it is about twice as fast as C.
If most coeff values are > zbin it is about 10-15x as fast as C.
BUG=webm:1426
Change-Id: I5d3c007b014a372d5ef0882b39bb48983b4131c7
When the superblock partition is based on the nonrd-pickmode,
we need to avoid the denoising. Current condition was based on
the speed level. This change is to make the condition at the
superblock level, as the switch in partitioning may be done at
sb level based on source_sad (e.g., in speed 6).
Change-Id: I12ece4f60b93ed34ee65ff2d6cdce1213c36de04
This reverts commit c9266b8547.
Disable source_sad when resolution > 1080P. The test should
pass now.
BUG=webm:1452
Change-Id: I72dde88e66590ff9e41da5e5dd83f5550a83f082
left shifting a negative value is undefined; quiets a ubsan warning.
this is applied to a constant, no change in the generated code.
Change-Id: I595f0ff7904ef025e07bb80234293d958dc9f254
This reverts commit aa1c4cd140.
This fails the following tests with extreme input coefficients:
SSE2/InvTrans8x8DCT.CompareReference/0
SSE2/InvTrans8x8DCT.CompareReference/2
previously the optimized path was skipped in this range
Change-Id: I9af015a46eba96208834a219fafd651d37556a80
Move the source_sad feature to speed 6 (from speed 7), and
add speed feature to switch from the variance-based partition
to reference_partition (which uses nonrd-pickmode for bsize selection)
if source_sad is high.
Currently used only for speed 6 for resoln <= 360p.
About 4-5% improvement on 360p in RTC set.
Some speed slowdown, but still ~30% faster than speed 5.
Change-Id: Ib0330ee5fe9fdd2608aed91359a2a339d967491c
This reverts commit 03f5e300d6.
This causes test failures under OSX:
SSSE3/VP9QuantizeTest.EOBCheck/0
SSSE3/VP9QuantizeTest.OperationCheck/0
Change-Id: I122732717ead1f7af5b04c529a6948e382e5e59b
allow the right shift to operate on 64-bits, this matches the rest of
the implementations
previously:
b0f1ae147 vpx_get16x16var_avx2: correct cast order
Change-Id: I632ee5e418f3f9b30e79ecd05588eb172b0783aa
allow the right shift to operate on 64-bits, this matches the rest of
the implementations
missed in:
6acd061aa variance_avx2: sync variance functions with c-code
Change-Id: Icae436b881251ccb9f9ed64fcbf8d358c58a4617
For 8-bit the subtrahend is small enough to fit into uint32_t.
For 10/12-bit apply:
63a37d16f Prevent negative variance
previously:
47b9a0912 Resolve -Wshorten-64-to-32 in highbd variance.
c0241664a Resolve -Wshorten-64-to-32 in variance.
Change-Id: I181c85f0b9a03da37c2e8b89482d48aa3dbc0aee
Avoid unsigned overflow warning:
unsigned integer overflow: 19974 - 32703 cannot be represented in type
'unsigned int'
Change-Id: Ifebee014342e4c6f3b53306c0cad6ae0b465ac12
Backend specific optimization for PPC VSX reads 16 bytes, whereas arm neon /
sse2 only reads <= 8 bytes. Although the extra bytes read are actually never
used, this is not a warrant for groping around. Fixed by allocating more when
building for VSX. This is reported by asan.
Also note - PPC does have assembly that loads 64-bit content from memory - lxsdx
loads one 64-bit doubleword (whereas lxvd2x loads two 64-bit doubleword) from
memory. However, we only have "vec_vsx_ld" builtins that mapped to lxvd2x, no
builtins to lxsdx. The only way to access lxsdx is through inline assembly,
which does not fit well in the origin paradigm.
Refer:
vsx:
vpx_tm_predictor_4x4_vsx @ third_party/libvpx/git_root/vpx_dsp/ppc/intrapred_vsx.c
neon:
vpx_tm_predictor_4x4_neon @ third_party/libvpx/git_root/vpx_dsp/arm/intrapred_neon_asm.asm
sse2:
tm_predictor_4x4 @ third_party/libvpx/git_root/vpx_dsp/x86/intrapred_sse2.asm
BUG=b/63112600
Tested:
asan tests passed.
Change-Id: I5f74b56e35c05b67851de8b5530aece213f2ce9d
Keep optimized code out of the reference implementation. This matches
the style of the other sub calls.
Change-Id: I3da6acd4f2c647b029c420e22ac9410a18259689
0.007% regression on rtc and 0.004% gain on rtc_derf.
1 thread on QVGA,VGA and HD has ~0.2% speed regression while 2 threads has
~0.2% speed gain on Google Pixel.
Change-Id: Ia4a6ec904df670d7001e35e070b01e34149d23dc
Officially the quant structures are 8 elements, with one dc element and
7 repeated ac elements. The low bit depth optimizations take advantage
of this to fill the xmm registers. The high bit depth version manually
duplicates the values.
If all the optimizations were unified, the structure sizes could be
greatly reduced.
Change-Id: Ibd7a0337a7832ce2a1a05ee433c310077e1059ae
Use only valid values for quantize inputs. These were determined by
looping over vp9_init_quantizer and looking for max and min values.
This allows extending the test to the low bit depth functions which were
not designed to handle all possible inputs but only valid inputs.
Change-Id: I94e1d8863a49ac227845b65c6b50130e10e6319e
To fix valgrind issueis with SVC tests.
SVC encoding uses prune_evenmore which is causing uinit value.
Will re-enable later when issue is resolved.
Change-Id: I257ff878cf78197ddd813db056582a4d5fe94f44
When content_state_sb is set to LowVarHighSumdiff, don't reset
it to VeryHighSad. Visually better on clips with strong lighting changes.
Small/negligible change in RTC metrics and speed.
Change-Id: I20c383e3c4cf8d1149de5f9260449c0b7cf7c6aa
When int_pro_motion_estimation is done for superblock in
choose_partitioning, use it to avoid the full_pixel_search
for NEWMV mode, if bsize is >= 32X32.
For speed > 7.
Small/neutral change on RTC metrics.
~1-2% speedup on arm on high motion clip.
Change-Id: I3cfe6833ff4bf75d4afa83eaf058ad45729de85b
Although the low bitdepth functions are identical (excepting the need
for larger intermediate values) they do not pass these tests. This
improves the error output to aid debugging.
Simplify buffer usage with Buffer and removing unnecessarily aligned
variables.
eob is a single element and never written using aligned instructions.
BUG=webm:1426
Change-Id: Ic95789a135cf1e8a3846d85270f2b818f6ec7e35
Reduces memory usage, and speeds up encoding for some difficult clips.
No impact on output or metrics.
Ported from aomedia patch:
https://aomedia-review.googlesource.com/c/14501
Change-Id: I26ec69af8336f9e80da486a1cfbfc89a3596954d
This reintroduces the fix:
https://chromium-review.googlesource.com/c/422807/
and later reverted here:
https://chromium-review.googlesource.com/c/447843/
BUG=webm:1355
This time behind a compile time flag :
configure --disable-always_adjust_bpm
configure --enable-always_adjust_bpm
This should make side by side testing easier and let users of the
lib pick which way they want to go.
Change-Id: I7d7b37b83015dc001810af84c132cbc1e71ba8d6
For fixed pattern SVC: keep track of denoised last_frame buffer
for base temporal layer, and if alt_ref is updated on middle/upper
temporal layers, force an update to denoised last_frame buffer.
This allows for improved denoising on top temporal layers.
Change-Id: Icbd08566027d4d2eabc024d3b7a0d959d2f8c18b
This code is unused in vp9. Only vp8 still contains references to
vpx_sad_NxMx[3|8] and only for sizes 16x16, 16x8, 8x16, 8x8 and 4x4.
Remove the remaining sizes and all the highbitdepth versions.
BUG=webm:1425
Change-Id: If6a253977c8e0c04599e25cbeb45f71a94f563e8
Denoiser is used in real-time mode which does not use alt-ref.
Reduce memory usage when denoiser is enabled.
Change-Id: I54ba3bcaeeb1818bbdf718ef90e97d4897ff793d
* changes:
sad neon: avg for 64x[32,64]
sad neon: macroize 64xN definitions
sad neon: avg for 32x[16,32,64]
sad neon: macroize 32xN definitions
sad neon: avg for 16x[8,16,32]
sad neon: macroize 16xN definitions
this has been set to max since:
f5c36a5ce VP9: turn on tile-columns and frame-parallel-mode by default
~v1.4.0
Change-Id: Ic796fc05abe73a58700ec50e3f8e72d3462898ec
In the content_state for a superblock is set to HighSad,
use that to bias some decisions in variance partition and
nonrd pickmde: use int_pro_motion for sad computation in
choose_partitioning, and set large_block in pickmode based
on the content_state_sb.
Only affects speed >= 7.
Immprovement for high motion content.
Small gain (~1%) in RTC metrics.
Speedup of ~5 for high motion clip on android (speed 8, 1 thread).
Change-Id: I5774c4854f012b89c8e969f6129b60988c2ce11c
this has been on by default since:
f5c36a5ce VP9: turn on tile-columns and frame-parallel-mode by default
~v1.4.0
Change-Id: I52017ab0157feaf429dce3d9e1af8a53bb5c1b65
the file was empty after the struct removal. the only remaining use was
within vp9_dx_iface, but the wrapper became unnecessary after the
removal of frame_parallel_decode.
BUG=webm:1395
Change-Id: I515ab585d701e77d388d12b2802d844c424f9bcd
This patch attempts to address a bug reported for 4K video.
https://b.corp.google.com/issues/62215394
In this instance a perfect storm of a moderate complexity section
followed by a much easier section where a CGI overlay helped to
suppress film grain noise, followed by a much harder and very grainy
section at the end, cause a massive local rate spike that pushed a chunk
over the upper allowed rate limit.
This patch detects cases where the rate for a frame is much higher than
expected and allows, in this special case, for rapid adjustment of the active
Q range.
For the example chunk in the bug report the target rate was 18Mb/s and the
observed rate was over 37 Mb/s with a surge for the last few frames to over
100Mb/s. This patch brings the overall chunk rate right back down to ~18.2 Mbit/s
and almost completely eliminates the rate spike at the end. (See graphs appended
to bug report)
Also see I108da7ca42f3bc95c5825dd33c9d84583227dac1 which fixes a bug
unearthed during testing of this patch and also has a bearing on high rate
encodes such as 4K.
This patch does have a negative impact on some metrics. Most notably there are
clips in our standard test set where it hurts global psnr (though in many cases it
conversely helps SSIM, FAST SSIM and PSNR-HVS). It is also worth noting that
the clips (and data rates) where there is a big metric impact, are almost all cases
where there is currently a significant overshoot vs the target rate and overall rate
accuracy is greatly improved.
Change-Id: I692311a709ccdb6003e705103de9d05b59bf840a
Local application of:
https://github.com/google/googletest/pull/1066
Suppress unsigned overflow instrumentation in the LCG
The rest of the (covered) codebase is already integer overflow clean.
TESTED=gtest_shuffle_test goes from fail to pass with -fsanitize=integer
Change-Id: I8a6db02a7c274160adb08b7dfd528b87b5b53050
left shifting a negative value is undefined; quiets a ubsan warning.
this is applied to a constant, no change in the generated code.
Change-Id: Ia17a7672d4832463decbc4afd6cd42974d02698e
Finish the calulations in neon registers. This avoids a potentially
expensive move from neon to gp and allows at least clang to store
directly to memory.
BUG=webm:1424
Change-Id: Idef25eec95f7610947167818e9194bde8b00d282
this makes the function compatible with high-bitdepth and fixes test
failures since:
5ac88162b partial fdct test
Change-Id: Ib630694608237f0c515948942e05dbea259ba338
testing::Range does not include the end parameter in the set of values.
also adjust the start to 2 as the single threaded case is already
covered in another instantiation
Change-Id: Iae3bf3ed4363dd434eccfa5ad4e3c5e553fbee60
For nonrd_pickmode: add condition for checking
intra mode if the sb content state is VeryHighSad.
Reduces artifacts when sudden change in content.
Metrics on RTC/RTC_derf neutral (small gain).
No speed loss observed.
Change-Id: I07006d28fd2dc06c1d06b07630102b0fece50c40
the last frame_worker_owner, row and col references were removed in:
131bd06e6 remove vp9_dthread.c
BUG=webm:1395
Change-Id: Ia7fb2e8782b12a58d2a2263849d20a8abf06aef6
and the related prototypes in vp9_dthread.h. the last references were
removed in:
09dabc58d VP9_COMMON: rm frame_parallel_decode
vp9_dx_iface.c still uses FrameWorkerData
BUG=webm:1395
Change-Id: Ica8e98ae776fc0105f1fbbed9e0a729808980810
creating a thread associated with the sole worker isn't necessary when
only execute() is being used after the removal of frame_parallel_decode.
BUG=webm:1395
Change-Id: I2255ce72607321e5708bc82a632dc6825d4eff5c
Add a method to acm_random.h to generate ranges of values
Add a way to call that method to buffer.h
Adjust dct_[partial_]test.cc to use it.
Change-Id: I8c23ae9d27612c28f050b0e44c41cb4ad2494086
this field has been 0 since:
01d23109a vp9: make VPX_CODEC_USE_FRAME_THREADING a no-op
BUG=webm:1395
Change-Id: I15448e9401e15329b54c6878dda033b17be5ec6b
VPX_CODEC_USE_FRAME_THREADING was made a no-op in:
01d23109a vp9: make VPX_CODEC_USE_FRAME_THREADING a no-op
and the tests in this file have been disabled since:
6ab0870d4 disable VP9MultiThreadedFrameParallel tests
BUG=webm:1395
Change-Id: I2c7a250acb65cf9522cf8a7bb724bb92070e41c6
this was made a no-op in:
01d23109a vp9: make VPX_CODEC_USE_FRAME_THREADING a no-op
and the test hitting this branch has been disabled since:
6ab0870d4 disable VP9MultiThreadedFrameParallel tests
rename the test to VP9MultiThreaded to exercise the tile-based threading
BUG=webm:1395
Change-Id: I35564a75eb5a7d7f7ccb923133b1b07295201f4c
Always return an int32_t. Since it needs to be moved to a register for
shifting, this doesn't really penalize the smaller transforms.
The values could potentially be summed and shifted in place.
BUG=webm:1424
Change-Id: Id5beb35d79c7574ebd99285fc4182788cf2bb972
For the 8x8_1, the highbd output fit nicely in the existing function. 12
bit input will overflow this implementation of 16x16_1.
BUG=webm:1424
Change-Id: I2945fe5478b18f996f1a5de80110fa30f3f4e7ec
The function was originally written with HBD in mind. Enable it and
configure the tests.
BUG=webm:1424
Change-Id: I78a2eba8d4d9d59db98a344ba0840d4a60ebe9a1
* changes:
sad neon: rewrite 64x64 and add 64x32
sad neon: rewrite 32x32, add 32x16 and 32x64
sad neon: rewrite 16x8, 16x16, add 16x32
sad neon: rewrite 8x8 and 8x16
sad neon: rewrite 4x4 and add 4x8
Test the _1 variant of the fdct, which simply sums the block and applies
a modifying shift based on the block size.
BUG=webm:1424
Change-Id: Ic80d6008abba0c596b575fa0484d5b5855321468
Existing logic was only affecting resolutions above 720p.
Needs more testing for reducing subpel for speed >= 8.
No change on RTC metrics.
Change-Id: I2f4bf9f25891614aafa9a86aa5a5063a3ccfce4d
This could save some cycles since skin detection is used in multiple
places in vp9.
1~2% speed up on ARM.
Change-Id: I86b731945f85215bbb0976021cd0f2040ff2687c
Split to load_input_data4() and load_input_data8().
Use pack with signed saturation instruction for high bitdepth.
Change-Id: Icda3e0129a6fdb4a51d1cafbdc652ae3a65f4e06
this normalizes these tests with the regular variance ones both in
implementation and test list output
Change-Id: I387aea81456f94b8223b8fb2a28cab94bc1aa9d5
Use the scene detection for CBR mode, and use it to reset the
rate control if large source sad is detected and rate
correctioni fact/QP is at minimum state.
Avoids large frame sizes after big content change following
low content period.
Only affects CBR mode for 1 pass at speeds 5, 6, 7.
Change-Id: I56dd853478cd5849b32db776e9221e258998d874
Fix misplaced cast that caused an overflow and incorrect rate adaptation
behavior for high data rates. This in particular will have affected 4k encodes
but could also have come into play for some higher rate 1080p cases.
In our standard test sets the quality impact is small though several high rate
clips show improved rate accuracy. This can also impact the number of recode
loop hits and on one problem 4k clip the encode time for speeds 0 and 1 was
reduced by >25%
Change-Id: I108da7ca42f3bc95c5825dd33c9d84583227dac1
Use it to limit NEWMV early exit in nonrd pickmode
Small change in RTC metrics, has some improvement
for high motion clips.
Change-Id: I1d89fd955e1b3486d5fb07f4472eeeecd553f67f
this is consistent with other threaded tests and ensures gtest_filters
meant to operate on these pick them up
Change-Id: I99ce53720553a22c4b9905a2882273c2be2c031b
and vp8_fast_quantize_b_impl_mmx; this was never enabled in rtcd
an sse2 version exists so there isn't much reason to keep a mmx
implementation around.
Change-Id: I8b3ee7f46ba194ffa0d0a6225a0f299f2a4dea90
use an int to quiet an unsigned rollover warning similar to:
25110f283 Fix an ubsan warning: vp9_quantizer.c
Change-Id: Iedecb79a17249bc18f10c0920f88cf704920f12b
Adjust the threshold for turning off cyclic refresh for high motion,
and avoid testing golden in nonrd pickmode for speed >= 8 if
golden refresh was long ago.
No change/neutral on RTC metrics.
Change-Id: I40959b8d9637f3553e7458bbabd8c6024c2c09c0
vpx_idct32x32_1024_add_ssse3() is actually a sse2 function and faster
than vpx_idct32x32_1024_add_sse2(). Replace the slow one. All are
code relocations, no new code.
Change-Id: I5dac0e98cc411a4ce05660406921118986638d19
'in' is used for the reference fdct. 'coeff' is input to the idct being
tested and 'dst[16]' is output
Fixes a segfault on unaligned memory access on x86.
Change-Id: I3691b1380ed49986897dd89a63ce63a80a0e0962
this was deleted in:
98967645a Remove vpx_idct8x8_64_add_ssse3()
but this was merged in:
9e03eedf6 Merge changes Ib26dd515,Ie60dabc3
after:
a92991133 Merge "dct tests: run all possible sizes in one test"
which added a new reference
Change-Id: I8da4a6c80d27b237a378ff15eead1daab89e7e25
Don't overide max_gf_interval if it's not specified. It will
be assigned with a default value in vp9_rc_set_gf_interval_range().
BUG=b/62803416
Change-Id: Ide46ce00279ed076865fc54ce98c55a994f0c798
Sample encoder change: reduce max-intra-rate to 1000 and
buf-initial to 600. Paramaters affect target size of key frame.
Change-Id: I2be6bc2927f5fa74e19e1efa3fb574d23a503300
Sample encoder change: reduce max-intra-rate to 1500 and
buf-initial to 700. Paramaters affect target size of key frame.
Change-Id: I01e238378b63eeef28dfc2178baadffcd3cc7561
Adjust some parameters in sample encoder: vpx_temporal_svc_encoder.
Parameters adjusted to set lower QP for initial key frame,
and allow for larger target size on subsequent key frames.
Change-Id: I092ad968e5b51b9f495dadb6ee96e810663c910e
Modify fdct4x4_test.cc to support all size combinations. This does not
add any new tests and in fact fails a few. There were minimal changes
made to the tests so it's not entirely surprising that some of the
larger 12 bit transforms are failing since it was initially only used
for 4x4.
In follow up patches the tests in fdct8x8_test.cc, dct16x16_test.cc and
dct32x32_test.cc will be evaluated and moved to dct_test.cc.
BUG=webm:1424
Change-Id: I72a23430f457d7fae8c91e706adc0e77c25abc8f
Set the base_mv_aggressive for temporal enhancement layers (TL > 0).
Under the aggressive mode, skip the NEWMV depending on the
SSE of the base_mv. Also reduce the subpel motion to 1/2 under
aggressive mode if base_mv is good.
Speedup ~3% with small/negligible loss in quality on RTC.
Affects speed >= 6.
Change-Id: I89341b279cad6da2a04b76d5e726016191dacdb8
It's almost identical with vpx_idct8x8_64_add_sse2(), except little
difference in instructions order.
Change-Id: Ie60dabc35eaa6ebae7c755e6cff00a710aad284f
This was ported from the greedy version in AV1, written by Dake He
(dkhe@google.com).
See:
https://aomedia.googlesource.com/aom/+/master/av1/encoder/encodemb.c#137
Greedy version is disabled by default, but can be picked by setting
USE_GREEDY_OPTIMIZE_B to 1.
To be enabled by default later.
This is both faster and better in terms of compression.
Compression Improvement:
------------------------
lowres: -0.119
midres: -0.064
hdres: -0.405
Speed Improvement:
------------------
(Based on encode time of 3 videos of different difficulties at
3 different target bitrates)
With --cpu-used=0: 0.38% to 5.55% faster
With --cpu-used=1: 0.24% to 2.79% faster
With --cpu-used=2: 0.29% to 1.46% faster
Change-Id: Ia7a23b3b244ad8eb253ac9e43cd03c5e021d2635
* changes:
Update high bitdepth load_input_data() in x86
Clean array_transpose_{4X8,16x16,16x16_2) in x86
Remove array_transpose_8x8() in x86
Convert 8x8 idct x86 macros to inline functions
some build systems have trouble with duplicate basenames.
vpx_dsp/skin_detection.[hc] were added in:
658e85425 Merge skin detection code in vp8/9.
BUG=webm:1438
Change-Id: Ieaa70b40bda409ec23e6d179b47a930ac6243b05
Set subpel prune_evenmore only for non_reference frames,
instead of all TL > 0 frames. Gain some quality back at
cost of small speed loss (~1-2%).
Change only effects SVC encoding at speed >= 7.
Change-Id: I5b9f51e51dccfd7050521a66996176b0415ca3f9
the check for error correction being disabled was overriding the data
length checks. this avoids returning incorrect information (width /
height) for the decoded frame which could result in inconsistent sizes
returned in to an application causing it to read beyond the bounds of
the frame allocation.
BUG=webm:1443
BUG=b/62458770
Change-Id: I063459674e01b57c0990cb29372e0eb9a1fbf342
min_gf_interval should be no less than min_altref_distance + 1,
as the encoder may produce bitstream with alt-ref distance being
min_gf_interval - 1.
BUG=b/38450599
Change-Id: Ifb733daa643ebc668d1b23e1ce92db94b66dabe8
Roughly 2x speedup. Since the only change for HBD is to store(), the
improvement appears to hold there as well.
BUG=webm:1424
Change-Id: I15b813d50deb2e47b49a6b0705945de748e83c19
Keep the 1/4subpel for all frames, use SUBPEL_TREE_PRUNED_EVENMORE
for all temporal enhancement layer frames.
Change-Id: Ibc681acbb6fc75b7b3c57fc483fcb11d591dfc9a
It is initialized to be { INT_MAX, 0, ... } in ffe0f9b.
No effect on encoders.
Make it consistent with other initializations.
BUG=webm:1440
Change-Id: Ie2a180d93626b55914c8c4255e466a1986d2b922
visual studio will warn if a 32-bit shift is implicitly converted to 64.
in this case integer storage is enough for the result.
since:
f3a9ae5ba Fix ubsan failure in vp9_mcomp.c.
Change-Id: I7e0e199ef8d3c64e07b780c8905da8c53c1d09fc
For SVC 1 pass non-rd mode:
Force subpel seach off for SVC for non-reference frames
under motion threshold.
Add flag to svc context to indicate if the frame is not used
as a reference.
Little/no quaity loss, ~2% speedup.
Change-Id: Ic433c44b514d19d08b28f80ff05231dc943b28e9
Speed >=8: for resolutions above CIF, and for low motion content,
set subpel_search_method to SUBPEL_TREE_PRUNED_EVENMORE.
Small speed gain (~2%) on vga clips,
RTC metrics up by ~2-3% on average.
Change-Id: Ie26ba0264589652f92dfe74308740debf94cf0cc
x86 requires 16 byte alignment for some vector loads/stores.
arm does not have the same requirement.
The asserts are still in avg_pred_sse2.c. This just removes them from
the common code.
Change-Id: Ic5175c607a94d2abf0b80d431c4e30c8a6f731b6
Split vp8/vp9 implementations on yv12_copy_frame_c.
Remove high-bitdepth codes from vp8_yv12_extend_frame_borders_c.
Clean up vp8 codes usage in vp9.
BUG=webm:1435
Change-Id: Ic68e79e9d71e1b20ddfc451fb8dcf2447861236d
Fix the condition on usage of source_sad for temporal layers.
FIx allows it to be used for the case of 1 temporal layer.
Change-Id: I02b1b0ade67a7889d1b93cee66d27c0951131fc3
Adjust the max_copied_frame setting for temporal layers.
Keep the same setting for non-SVC at speed 8.
This change also enables copy_partiton for non-SVC at speed 7,
but with smaller value of max_copied_frame (=2).
~2% speedup for SVC speed 7, 3 layers, with little/no quality loss.
Change-Id: Ic65ac9aad764ec65a35770d263424b2393ec6780
Unlike x86, arm does not impose additional alignment restrictions on
vector loads. For incoming values to the first pass, it uses vld1_u32()
which typically does impose a 4 byte alignment. However, as the first
pass operates on user-supplied values we must prepare for unaligned
values anyway (and have, see mem_neon.h).
But for the local temporary values there is no stride and the load will
use vld1_u8 which does not require 4 byte alignment.
There are 3 temporary structures. In the C, one is uint16_t. The arm
saturates between passes but still passes tests. If this becomes an
issue new functions will be needed.
Change-Id: I3c9d4701bfeb14b77c783d0164608e621bfecfb1
The sub pixel variance uses a temp buffer which guarantees width ==
stride. Take advantage of this with the 4x and avoid the very costly
lane loads.
Change-Id: Ia0c97eb8c29dc8dfa6e51a29dff9b75b3c6726f1
For aq-mode=3: refactor the condition for turning off
the refresh. Add some adjustments for high motion content.
No/little change in RTC metrics, only affects high motion case.
Change-Id: I7da8eabfb0e61db014be4562806f72ee5ef4a43b
When temporal layers are used, only allow for copy partition
on the top temporal enhancement layer frames.
Change-Id: I5472abdc0f9f6c8dafa75a7a84c615e08ae22af8
Only affects speed 8.
Make changes to copy partition to fix a bug in setting microblock
offset. Avg PSNR shows 0.02% gain on rtc_derf and 0.08% loss on rtc.
Change-Id: I61c3e5914dde645331344388e7437e5638acd4f3
The modified error was a derivative of the "coded_error"
that was used to allocate bits between different frames on the
assumption that the allocation should be linear in terms of this
modified error. I.e. a frame with double the modified error score
should all things being equal get double the number of bits. The
code also included upper and lower caps derived from input
VBR parameters.
This patch improves the initial calculation of the clip mean error
(now called "mean_mod_score" as it is no longer a prediction error)
used as the midpoint for the rate distribution function and normalizes
the output "modified scores" scores such that 1.0 indicates a frame
in the middle of the distribution. The VBR upper and lower caps are
then applied directly to a frame's normalized score.
This refactoring is intended to make it easier to drop in alternative
distribution functions or to base the rate allocation on a corpus wide
midpoint (rather than the clip mean).
Change-Id: I4fb09de637e93566bfc4e022b2e7d04660817195
Continue processing sets of 16 values. Plenty of improvement for 4x8
(doubles the speed) but only about 30% for 4x4.
BUG=webm:1422
Change-Id: Ib8dd96f75d474f0348800271d11e58356b620905
Advise the compiler that the store is eventually going to a uint8_t
buffer. This helps avoid getting alignment hints which would cause the
memory access to fail.
Originally added as a workaround for clang:
https://bugs.llvm.org//show_bug.cgi?id=24421
Change-Id: Ie9854b777cfb2f4baaee66764f0e51dcb094d51e
Add PartialIDctTest::PrintDiff() to help debugging.
In RunQuantCheck, try all combinations of +/-mask_ input for 4x4 idct.
Update PartialIDctTest::InitInput().
Change-Id: I13fd163954a4c1a3a6cfeb5e4a4d3d0e7ff901f4
Most existing first pass stats are stored in a form normalized to a
macro-block scale. However the error scores for intra / inter etc were
stored as frame level values but mainly used as MB level values.
This change fixes that. Normalized per MB values make comparisons
between different formats easier and in any case this is usually what is
wanted.
An change in results should be limited to slight differences in rounding.
*** Change after patch 8 +2 requiring new approval.
Final pre-submit testing showed one 4K clip with above expected change.
Investigation showed this was due to a value used to test for ultra low intra
complexity in key frame detection. This was a per frame not per MB value but
also did not scale with frame size. Replacement with a small per MB value
(based on original per frame value and cif frame size) resolved the KF detection
problem.
Also converted kf_group_error_left to a double in line with other error values
to reduce rounding problems in KF group bit allocation
All clips and sets now show nominal (or 0) change as expected.
Change-Id: Ic2d57980398c99ade2b7380e3e6ca6b32186901f
This reverts commit 0d88e15454.
Reason for revert: chromium builds are failing to locate vpx_rv during dlopen()
dlopen failed: cannot locate symbol "vpx_rv" referenced by "libstandalonelibwebviewchromium.so"
Original change's description:
> Add visibility="protected" attribute for global variables referenced in asm files.
>
> During aosp builds with binutils-2.27, we're seeing linker error
> messages of this form:
> libvpx.a(subpixel_mmx.o): relocation R_386_GOTOFF against preemptible
> symbol vp8_bilinear_filters_x86_8 cannot be used when making a shared
> object
>
> subpixel_mmx.o is assembled from "vp8/common/x86/subpixel_mmx.asm".
> Other messages refer to symbol references from deblock_sse2.o and
> subpixel_sse2.o, also assembled from asm files.
>
> This change marks such symbols as having "protected" visibility. This
> satisfies the linker as the symbols are not preemptible from outside
> the shared library now, which I think is the original intent anyway.
>
> Change-Id: I2817f7a5f43041533d65ebf41aefd63f8581a452
>
TBR=jzern@google.com,johannkoenig@google.com,rahulchaudhry@chromium.org,builds@webmproject.org
Change-Id: I0c2ea375aa7ef5fda15b9d9e23e654bb315c941b
This reverts commit 3704807805.
Reason for revert: <INSERT REASONING HERE>
Does not look to be the cause of the test failures.
Original change's description:
> Revert "vp8: Real-time mode: reduce mode_check_freq thresh for speed 10."
>
> This reverts commit 4a7424adba.
>
> Reason for revert: <INSERT REASONING HERE>
> Possibly causing test failures in roll into chromium.
>
> Original change's description:
> > vp8: Real-time mode: reduce mode_check_freq thresh for speed 10.
> >
> > Reduces quality regression at speed 10 for real-time mode.
> >
> > Change-Id: I9f624bea9ca262dab32ce9de7d6d91175d6becc8
> >
>
> TBR=marpan@google.com,builds@webmproject.org,jianj@google.com
> # Not skipping CQ checks because original CL landed > 1 day ago.
>
> Change-Id: I1defcb74e78a5a3bd29b7d1b21a96a79fa26a457
>
TBR=marpan@google.com,builds@webmproject.org,jianj@google.com
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: I13d86a2a68b8aa8c0c7465e6e58cff0e00bc7862
This reverts commit 4a7424adba.
Reason for revert: <INSERT REASONING HERE>
Possibly causing test failures in roll into chromium.
Original change's description:
> vp8: Real-time mode: reduce mode_check_freq thresh for speed 10.
>
> Reduces quality regression at speed 10 for real-time mode.
>
> Change-Id: I9f624bea9ca262dab32ce9de7d6d91175d6becc8
>
TBR=marpan@google.com,builds@webmproject.org,jianj@google.com
# Not skipping CQ checks because original CL landed > 1 day ago.
Change-Id: I1defcb74e78a5a3bd29b7d1b21a96a79fa26a457
Move the tran_low_t helper functions to a new file. Additional
load/store functions will be added here.
Change-Id: I52bf652c344c585ea2f3e1230886be93f5caefc3
During aosp builds with binutils-2.27, we're seeing linker error
messages of this form:
libvpx.a(subpixel_mmx.o): relocation R_386_GOTOFF against preemptible
symbol vp8_bilinear_filters_x86_8 cannot be used when making a shared
object
subpixel_mmx.o is assembled from "vp8/common/x86/subpixel_mmx.asm".
Other messages refer to symbol references from deblock_sse2.o and
subpixel_sse2.o, also assembled from asm files.
This change marks such symbols as having "protected" visibility. This
satisfies the linker as the symbols are not preemptible from outside
the shared library now, which I think is the original intent anyway.
Change-Id: I2817f7a5f43041533d65ebf41aefd63f8581a452
Increase the partition and acskip thresholds for temporal
enhancement layers.
~1-2% speedup, with negligible loss in quality.
Change-Id: Id527398a05855298ad9ddac10ada972482415627
For SVC 1 pass non-rd pickmode, the interpolation filter for the
upsampling of the golden (spatial) reference was not being explicitly
set and instead was takin gwhatever value was set in the previous
mode/block (which would be either EIGHTTAP or EIGHTAP_SMOOTH).
Fix it to the default EIGHTTAP for now, to be updated/selected
adaptively in a later change.
Minor adjustmemt to rate targeting thresholds in datarate unittests.
Change-Id: I52085048674072c6cfb7163e11e9a2658d773826
A more detailed explanation of the experimentation
leading to this change can be found in:-
https://docs.google.com/a/google.com/document/d/13lsYhxgPyxUHvEess6wg9nikaonIZKY9Ak_Lpafv5Mo/edit?usp=sharing
This change gives gains across all our standard test sets for
overall psnr, ssim, fast ssim and psnr-HVS.
Values expressed as % reduction in bitrate.
Low res set -0.257, -0.192, -0.173, -0.101
Mid res set -0.233, -0.336, -0.367, -0.139
High res set -0.999, -1.039, -1.111, -0.567
NetFlix 2K set -0.734, -0.174, -0.389, -0.820
Netflix 4K set -0.814, -0.485, -0.796, -0.839
Change-Id: Ie981fb3c895c9dfcfc8682640d201a86375db5c8
Speed up for speed 0.
Reduce 10+% of encoding time for hdres in speed 0,
with less than 0.1% PSNR loss.
Compute total difference of previous and current frame context probability
model. If the diff is less than the threshold, skip recoding the frame.
Borg test (positive number means performance loss):
lowres midres hdres
PSNR: 0.030 0.032 0.065
Local speed test: bitrate set at 1200
blue_sky pedestrian rush_hour
Encoding time: -10.0% -16.5% -16.5%
Change-Id: I4e2d200ea3115d48b2c3e890143596b31b8ef9e9
Introduced append situation in Commit 0178d97 which could be
confusing. Clean a little bit and add some comments.
Change-Id: I69ad336f805aca7ce9d45515b8cd237423fadbb2
When the noise estimate is forced off due to large motion,
reset the counter and set smaller window for next estimate.
Change-Id: Ifa4ec95396134173a00d48353ad52f1b6a40c217
Add option in SVC to set the filter type and phase for
the frame level downsampling filters.
For 3 spatial layers: set downsampling filter type to bilinear
and set phase to 8, for lowest spatial layer.
Change-Id: Id81f4b1ba93db19c1cd37b6a46d1281a2c61bc43
Makes more sense to call the corresponding partial idct C function
instead of the full idct C function as the reference.
Change-Id: Ibb7681dd063edd6307ba582c10c26c4c6a4b78c6
Base the condition on the resolution of the spatial layer.
And remove restriction on scaling factor.
Change-Id: Iad00177ce364279d85661654bff00ce7f48a672e
Read in a Q register. Works on blocks of 16 and larger.
Improvement of about 20% for 64x64. The smaller blocks are faster, but
don't have quite the same level of improvement. 16x32 is only about 5%
BUG=webm:1422
Change-Id: Ie11a877c7b839e66690a48117a46657b2ac82d4b
For lowest spatial layer, in 3 layer SVC, set the
downsampling filtertype to get averaging filter.
Needed for reducing aliasing on low-res layer,
small increase in overall encoder time.
Change-Id: Ia31460123bd91b72eca49b46dd924b9f226d4563
An intended behavior change disabling exhaustive searches in speed
feature causes VP9/DatarateTestVP9LargeDenoiser.4threads test failure.
Change the threshold to make it pass.
BUG=webm:1429
Change-Id: Ibcbe2314c6b2525799894f5d7204fc8eb4ec2a1e
Adjust thresholds for noise estimation, for resolutions above VGA.
Tends to push cleaner/low noise clips to LowLow state.
No change in RTC metrics.
Change-Id: I739ca6b797d0a60ccd1c6c6a2775269b1f007e5e
Set noise level to kLowLow for high motion low res clips.
Change the normalization in noise metric for low res.
Reduce the initial time-window for all resolutions.
Change-Id: Iaed39dbb50b205cd9c735dc5b84822304fb01987
Add support for everything except block sizes of 4.
Performance is better but numbers will improve again when the variance
optimizations land.
BUG=webm:1423
Change-Id: I92eb4312b20be423fa2fe6fdb18167a604ff4d80
When a neon version is available it will be called. This allows
decoupling the variance implementations and has no real downside. For
most configurations, the call will be #define'd to the neon
implementation.
Change-Id: Ibb2afe4e156c5610e89488504d366b3e6d1ba712
When the width is equal to 8, process two rows at a time. This doubles
the speed of 8x4 and improves 8x8 by about 20%.
8x16 was using this technique already, but still improved a little bit
with the rewrite.
Also use this for vpx_get8x8var_neon
BUG=webm:1422
Change-Id: Id602909afcec683665536d11298b7387ac0a1207
Some of the mixed sizes were missing. They can be implemented trivially
using the existing helper function.
When comparing the previous 16x8 and 8x16 implementations, the helper
function is about 10% faster than the 16x8 version. The 8x16 is very
close, but the existing version appears to be faster.
BUG=webm:1422
Change-Id: Ib0e856083c1893e1bd399373c5fbcd6271a7f004
Add 31bit pairs before unpacking in x86 block error code
AVX2 code provides a very minor performance improvement.
BUG=webm:1210
Change-Id: I4c82308eaf65741dca2f5c6db9be9c85f905073a
For SVC 1 pass real-time: add condition to skip the
golden (spatial) reference mode in non-rd pickmode.
Condition is to skip golden if the sse of zeromv-last mode
is below threshold. And change order in ref_mode_set_svc
to make sure golden zeromv is tested after last-nearest.
Speedup ~3-4% with little/negligible quality loss.
Change-Id: I6cbe314a93210454ba2997945f714015f1b2fca3
Approximates division using multiply and shift.
Speeds up both sizes (8x8 and 16x16) by 30 times.
Fix the call sites to use the RTCD function.
Delete sse2 and mips implementation. They were based on a previous
implementation of the filter. It was changed in Dec 2015:
ece4fd5d22
BUG=webm:1378
Change-Id: I0818e767a802966520b5c6e7999584ad13159276
Don't force disabling of adaptive_rd_thresh for realtime when
row_mt_bit_exact is set.
Row based adaptive rd is made usable in CL
454882(https://chromium-review.googlesource.com/c/454882) for REALTIME.
Change-Id: Ief023414f0fd6eb86f299dd46ae58f4436875af5
Add tentative max cpb size values for levels 5.2 and up. Otherwise
encoding will fail when targeting for these levels.
Change-Id: Ib7e0ba4b9836ea1ac900b6822543812843d48463
107de19698 changes the encoder alt-ref selection behavior. Assuming
min_gf_interval = max_gf_interval = 4, the frame order would be
frm_1 arf_1 frm_2 frm_3 frm_4 frm_5 arf_2 before 107de19698;
frm_1 arf_1 frm_2 frm_3 frm_4 arf_2 frm_5 after 107de19698.
This patch reverts such alt-ref placement change.
Change-Id: I93a4a65036575151286f004d455d4fcea88a1550
Make some speed setting changes for temporal enhancement layers,
and remove the switch in subpel_force_stop for the aggressive_base_mv
in non-rd pickmode.
Gain some 2-3% speed with little/negligible quality loss.
Change-Id: I3e2a7f80ff45f38c0a6ceb01b34dbca2f53edbf0
For speed >= 8 and color_sensitivity not set, skip the transform
skipping test in UV planes.
Add a new condition to check noise level to skip chroma check
for speed >= 8 if y_sad is high.
1~2% speedup on ARM for speed 8.
Borg tests show neutral results in both rtc and rtc_derf.
Change-Id: Idecd3ff6e28c97757a43bb6f3a7082c85f72109c
Add a low-variance high-sumdiff to the superblock content state
and use it to limit the mv and bias some decisions in non-rd pickmode.
Only affects speed >= 6.
Reduces artifact for lighting changes.
Small/no difference in metrics on RTC set.
Change-Id: Ic84b2379fe0ae3fa71ae826ee6bae3eaf551a25b
This patch followed allow_exhaustive_searches feature modification and
continued to modify the encoder to achieve the determinism in the row
based multi-threaded encoding. While row-mt = 1 and using multiple
threads, the adaptive feature in encoder was disabled, which gave
BDRate gain(at speed 1, -0.6% ~ -0.7%; at speed 2, -0.46% ~ -0.59%),
but some encoder speed losses(7% ~ 10% at speed 1 and 3% ~ 6% at
speed 2). These speed losses were acceptable considering the speed
gains obtained from row-mt.
Change-Id: I60d87a25346ebc487a864b57d559f560b7e398bb
A previous patch turned on allow_exhaustive_searches feature only for
FC_GRAPHICS_ANIMATION content. This patch further modified the feature
by removing the exhaustive search limit, and made it no longer adaptive.
As a result, the 2 counts that recorded the number of motion searches
were removed, which helped achieve the determinism in the row based
multi-threading encoding. Tests showed that this patch didn't cause
the encoder much slower.
Used exhaustive_searches_thresh for this speed feature, and removed
allow_exhaustive_searches. Also, refactored the speed feature code
to follow the general speed feature setting style.
Change-Id: Ib96b182c4c8dfff4c1ab91d2497cc42bb9e5a4aa
The more aggressive settings should only be used when denoise_svc
condition is satisfied (which means top spatial layer).
Change-Id: Ia8e3515b27f31bf21b1976ca80a2fa826daece3a
In non-rd pickmode (speed >= 5), avoid duplication of computations in
model_rd_for_sb_y when the speed feature use_simple_block_yrd is
enabled (or for high bitdepth build under certain conditions).
QVGA, VGA and HD have 1.23%, 2.68% and 1.7% speedup on ARM for speed 8,
respectively.
Encoding results are bitexact for speed >= 5.
Change-Id: I3f9130810c21439f5ad7e159e21cb2243dcd05f1
Re-enable the SVC tests, wrap the non-zero expectation
in GetMismatchFrames around #if CONFIG_VP9_DECODER.
Change-Id: I0e8a2d78b868c32f18fe597540f397d3a1b303b5
Slightly faster, the other dc predictors cannot be faster since
the computation speedup is overwhelmed by the time spent reading
dst to write just the 8x8 part.
Change-Id: I94a0b50500adf8b7b6bb919dbf5c7adf5b9fba66
The allow_exhaustive_searches feature improves the encoding quality
of FC_GRAPHICS_ANIMATION content a lot. For non-FC_GRAPHICS_ANIMATION
content, the quality test result is almost neutral. This patch makes
this feature to be used only for FC_GRAPHICS_ANIMATION content.
The motivation of doing that is to make this feature no longer adaptive,
which will be implemented in the following patch.
Change-Id: Ic911df6dd757402b6480789cc247801e99840369
Replace by CAST_TO_BYTEPTR/SHORTPTR.
The rule is: if a short ptr is casted to a byte ptr, any offset
operation on the byte ptr must be doubled. We do this by casting to
short ptr first, adding offset, then casting back to byte ptr.
BUG=webm:1388
Change-Id: I9e18a73ba45ddae58fc9dae470c0ff34951fe248
The scaling filter with zero shift will give sub-sampling for
2x downsampling. Allow for a phase shift to get an averaging filter.
Usage is for source scaling in 1 pass SVC mode for 1:2 downscale.
Reduces aliasing in downsampled image.
Keep the phase to 0/off for now.
Change-Id: Ic547ea0748d151b675f877527e656407fcf4d51e
Disable the 1 pass CBR SVC tests with temporal_layers > 1.
Issue with the commit 863f860, which will cause encoder/decoder
mismatch due to skipping encoder loopfilter for non-reference frames.
Will re-enable the tests when fixed.
Change-Id: I74918a0045a17976b069c4be63fbeb921974df0d
This condiiton is not needed as key_frame should set the refresh
of the reference frames, but good to have for clarity in condition.
Change-Id: Icf9838e7e4f0ff5cf0a9562ae3b5d6c7e6f78702
Modify the frame flags to update the ARF on top layer,
for the tests:
VP9/DatarateTestVP9Large.BasicRateTargeting3TemporalLayers
VP9/DatarateTestVP9Large.BasicRateTargeting3TemporalLayersFrameDropping
This is needed to fix the encode/decoder mismatches caused by 863f860,
and removed in the revert e9b7f98.
Change-Id: I6b9fecfdd17315fc0179e29949338c77636026c0
Buffers on 32 bit x86 builds only guaranteed 8 byte alignment. Fixed
with "AvgPred test: use aligned buffers" and "sad avg: align
intermediate buffer"
Also re-enable asserts on the C version.
BUG=webm:1390
Change-Id: I93081f1b0002a352bb0a3371ac35452417fa8514
This reverts commit 863f860bfc.
This causes encoder / decoder mismatches in various
VP9/DatarateTestVP9Large.BasicRateTargeting3TemporalLayers tests
BUG=webm:1408
Change-Id: Ic200c39d7ed9c0b0247ef562f5d6f7b2625f7e14
For low resolutions (<= CIF): use quarter-pixel and simple_block_yrd.
~5% gain on RTC_derf.
~6-7% slowdown on ARM.
Change-Id: I4439ebd1116b9decac04786503f978840b68a60c
Fix to avoid getting stuck at very low Q even
though content is changing, which can happen for --min-q=0.
Fix is to more aggressively increase active_worst_quality
when detecting significant rate_deviation at very low Q.
Change will only affect 1 pass VBR for --min-q < 4, so no
change in ytlive metrics for --min-q >= 4.
Change-Id: I4dd77dd7c08a30a4390da0ff2c8bda6fccfa76d7
Useful for SVC, where the top layer enhancement frames may
not update any reference buffers, as is the case for the
patterns in the 1 pass CBR SVC when #temporal_layers > 1.
~3% encoder speedup for SVC patterns with temporal layers
in 1 pass CBR mode.
Updated the SVC datarate tests for the mismatch frames.
Set the frame-dropper off in some tests with #temporal_layers > 1
so we can correctly set #mismatch frames. Adjusted rate target
threshold for tests where frame-dropper was turned off.
Change-Id: Ia0c142f02100be0fed61cd2049691be9c59d6793
Provides over 15x speedup for width > 8.
Due to smaller loads and shifting for width == 8 it gets about 8x
speedup.
For width == 4 it's only about 4x speedup because there is a lot of
shuffling and shifting to get the data properly situated.
BUG=webm:1390
Change-Id: Ice0b3dbbf007be3d9509786a61e7f35e94bdffa8
The MV unit test revealed an integer overflow issue in vp9_mcomp.c.
This was caused if the MV was very large. In mv_err_cost(), when
mv->row = 8184, mv->col = 8184 and ref_mv is 0, mv_cost = 34363
and error_per_bit = 132412, causing the overflow.
BUG=webm:1406
Change-Id: I35f8299f22f9bee39cd9153d7b00d0993838845e
Set adaptive_rd_thresh to 2 when simple block yrd is not used.
Fix regression caused by computing y sad without
int_pro_motion_estimation on low res motion clips.
Overall 0.07% quality loss on rtc_derf.
Change only affects low res on speed 8.
Change-Id: Ic6a188a56529f1034d6431005fb4b0e24e8a7e27
For speed 5, 1 pass CBR: Don't use the nonrd_pick_partition
on the segment, rather use choose_partitioning followed by
nonrd_select_partition (as is done on base segment).
Little/no quality loss on RTC and RTC_derf (< 0.3%),
speedup of at least 5%.
Change-Id: I5273d5f950e60adf5e437b4ca8c4f63964641e83
If the noise estimation is avoided due to large motion,
the last_source for denoising should still be updated.
Change-Id: I67155ea7dbe9ac2785978e64a27bdafd7d57aac0
To reduce refresh on partial super-blocks on boundary,
for noisy input. Reduces some artifacts on noisy input.
Change-Id: I10b5808a296874e08c7f378b3df58466591d8dbe
Edit
Move the condition for effectively disabling the denoising
for speed 5 into the vp9_denoiser_denoise().
This is cleaner, and also moving the condition into vp9_denoiser_denoise
will keep the denoiser buffer updated with the current source.
This allows for more consistent behavior if speed is changed midstream.
Change-Id: Ia001f591c56e454bf724c3ae73c024badb183ef8
To prevent the motion vector out of range bug, added a motion vector unit
test in VP9. In the 4k video encoding, always forced to use extreme motion
vectors and also encouraged to use INTER modes. In the decoding, checked if
the motion vector was valid, and also checked the encoder/decoder mismatch.
The tests showed that this unit test could reveal the issue we saw before.
Change-Id: I0a880bd847dad8a13f7fd2012faf6868b02fa3b4
For 8-bit the subtrahend is small enough to fit into uint32_t.
This is the same that was done for:
c0241664a Resolve -Wshorten-64-to-32 in variance.
For 10/12-bit apply:
63a37d16f Prevent negative variance
Change-Id: Iab35e3f3f269035e17c711bd6cc01272c3137e1d
Temporal denoiser runs in non-rd pickmode, so it is only used
for speed >= 5. Regression exists for speed 5, due to use of
reference_partition (which use non-rd pickmode for partitioning).
Avoid denoising for now at speed 5.
Change-Id: I74a74d2e1404d7cfd33dcf4ec06dd2e503256cf0
Base the low_content_frame metric on the motion vectors,
and adjust the logic for preventing golden update.
Small change in behavior: small positive gain (~0.2-1%) on clips
with high activity.
Change-Id: I0b861c8e9666cd82b45cde5ee57ee8a1e5ab453c
Code cleanup: merged two functions that were doing postencode
update for cylic refresh, remove some unused code and fix comments.
No change in behavior.
Change-Id: I9be0d7e346d34dec29bf4e5bb380a7bf81c8480a
BUG=webm:1397
(yunqingwang)
To verify that this patch wouldn't cause much performance change,
the Borg tests were run. Here was the result:
avg_psnr overall_psnr ssim
hdres: -0.002 0.006 0.013
midres: 0 0 0
lowres: 0 0 0
Change-Id: Iae395ae7b741e0513cf5bab9dcace110b792a67d
The row mt sync read uses sync_range = 1, and wouldn't work if we want
to use a sync_range that is greater than 1. To make it work, this sync
read code is modified. Pass in col instead of col - 1 to make it
consistent with other row mt code in VP9, and then add 1 in "while"
codition.
Change-Id: I4a0e487190ac5d47b8216368da12d80fec779c1a
Issue/bug happens for denoising with spatial layers, where
the golden (spatial) reference is used in pickmode, but
denoising is only done wrt to last (temporal).
Fix is to make sure set_ref_ptrs is set before build predictors
in denoiser.
Change-Id: I793cf441341edf7c4a88b8ab1e1b22b3cb0eb508
Temporary override to condition for disallowing intra-search in SVC,
since golden (spatial) reference is currently suppressed due to
artifact issue.
Change-Id: I28ed7fdddc9fcdbcc0a4175a247a3ecc94c11767
For non-rd variance partition, avoid the chrome check
unless y_sad is below some threshold.
Small decrease in avgPSNR (~0.3) on RTC set.
Small/negligible decrease on RTC_derf.
Change-Id: I7af44235af514058ccf9a4f10bb737da9d720866
Refactor to split the 1 passs source sad computation into scene
detection (currently used for VBR and screen-content mode), and
superblock based source sad computation (used in non-rd CBR mode).
This allows the source sad computation for CBR mode to be
multi-threaded.
No change in compression.
Change-Id: I112f2918613ccbd37c1771d852606d3af18c1388
d207/d63/d45/d117/d135/d153
~9-45% better depending on the predictor on 32-bit ARM, similar range on
x86-64
this matches the non-highbitdepth implementation
BUG=webm:1316
Change-Id: Iddebdf7c58c6f31c47cae04da95c6e5318200e4c
Make the source_sad feature work properly for cases of VBR or
screen_content with SVC.
Added unittest for SVC with screen-content on.
Change-Id: Iba5254fd8833fb11da521e00cc1317ec81d3f89b
tolerate a NULL hist being passed as a result of invalid parameters
passed to init_rate_histogram(). this fixes a divide by zero in
init_rate_histogram() with an invalid fps.
BUG=webm:1383
Change-Id: Id203e0f3b18d67a4a09aaf206abcce4708f966ec
Since y_sad is not computed yet (on the early exit due to source_sad),
no need to check for setting color_sensitiviy.
Only affects speed >=8. No change in behavior.
Change-Id: I3a6f2d20fed38d8b8ec51b75bcacf9a21f2db916
Allow for simple_block_rd for VGA resoln, and reduce
adaptive_rd_thresh to 1.
On average no loss on RTC set, ~4% speedup on mac.
Change-Id: Ib549c4061c853776062b5e34040f839d470fbebc
Change tests to reflect use. Input sizes will be 8 or 16 (but not
necessarily square).
filter_weight is capped at 2 and filter_strength at 6
Speed test, disabled by default.
Change-Id: Idfde9d6c4b7d93aaf0e641b0f4862c15e2a2af7a
Change it to row based array to avoid the slow down cause by sync.
row-mt on, speed 8, 2 threads: ~4% speedup for VGA on ARM benefited
from adaptive_rd_threshold.
Change-Id: I887e65a53af20a6c4f48d293daaee09dab3512cf
- Refer to patch: 48fca113d inv_txfm_ssse3,butterfly: fix win32 abi
compatibility.
- Change four butterfly() calls to butterfly_self(), to simplify the
operations.
Change-Id: Ib2a8cfe6cddcaf0a59e6e6270d8380055ea42ef3
Add additional condition to split to 16x16, for resolutions <= 360p,
reduces dragging artifact near moving boundary.
Small/no change on RTC metrics.
Change-Id: I314694f2166435d918f74e7ab42f002b07f40dae
For each superblock, keep track of how far from current frame
was the last significant content change, and use that (along
with GF distance), to turnoff GF search in non-rd pickmode.
Only enabled for speed >= 8.
avgPNSR on RTC/RTC_derf down by ~0.9/1.2.
Speedup on mac: ~3-5%.
Speedup on arm: 3.6% for VGA and 4.4% for HD.
Change-Id: Ic3f3d6a2af650aca6ba0064d2b1db8d48c035ac7
The sum of tx bloxk eobs is needed in the machine learning based partition
early termination. The eobs are first accumulated during tx search, and
then the value associated with the best tx_size is copied to ctx for later
use.
After the sum of eobs are calculated correctly, re-enabled
ml_partition_search_early_termination speed feature.
Re-did the quality/speed test to check the impact of the fix.
1. Borg test BDRATE result:
4k set: PSNR: +0.183%; SSIM: +0.100%;
hdres set: PSNR: +0.168%; SSIM: +0.256%;
midres set: PSNR: +0.186%; SSIM: +0.326%;
2.Average speed gain result:
4k clips: 21%;
hd clips: 26%;
midres clips: 15%.
The result is in line with the original result.
Change-Id: I4209a95c89be03b4cbfb6a95b16885f89feddbda
Add routine vp9_model_rd_from_var_lapndz_vec and call it from model_rd_for_sb
to model the rate and distortion for MAX_MB_PLANE Laplacian sources in
parallel. The caller ensures that all sources have non-zero variance.
Measured a 18% to 25% reduction in retired instructions, and 17% to 24%
reduction in instruction execution cost with different compilers for the
Laplacian modeling.
No change in behavior.
TEST=Verified that encoded files match bit for bit, with and without this
change.
BUG=b/33678225
Change-Id: I6b76947f21c659a349adb896e13e99f6e3f951e6
Don't denoise spatial layer frames whose base layer is a key frame.
Disallow golden reference for SVC with denoising on frames
that will be denoised (highest layer), as this removes bad artifact.
Will re-enable when issue is resolved.
Change-Id: I87a6597812330500966458172acfce54af65f70f
Fix the update of the denoiser buffer when the base
spatial layer is a key frame. And allow for better/lower
QP on high spatial layers when their base layer is key frame.
Change-Id: I96b2426f1eaa43b8b8d4c31a68b0c6d68c3024a2
Similar issue as Change bc1c18e.
The PartialIDctTest.ResultsMatch test on vpx_idct32x32_135_add_neon()
in high bit-depth mode exposes 16-bit overflow in final stage of pass
2, when changing the test number from 1,000 to 1,000,000.
Change to use saturating add/sub for vpx_idct32x32_34_add_neon(),
vpx_idct32x32_135_add_neon and vpx_idct32x32_1024_add_neon() in high
bit-depth mode.
Change-Id: Iaec0e9aeab41a3fdb4e170d7e9b3ad1fda922f6f
Reduce it from 5 to 4, small/no change in metrics or speed.
Small reduction in dragging artifact near moving head.
Change-Id: Ic3bc5ca67c70bf0c89fc2ed14454840a28ae5b6a
This patch was based on Yang Xian's intern project code. Further modifications
were done.
1. Moved machine-learning related parameters into the context structure.
2. Corrected the calculation of sum_eobs.
3. Removed unused parameters and calculations.
4. Made it work with multiple tiles.
5. Added a speed feature for the machine-learning based partition search
early termination.
6. Re-organized the code.
The patch was rebased to the top-of-tree.
Borg test BDRATE result:
4k set: PSNR: +0.144%; SSIM: +0.043%;
hdres set: PSNR: +0.149%; SSIM: +0.269%;
midres set: PSNR: +0.127%; SSIM: +0.257%;
Average speed gain result:
4k clips: 22%;
hd clips: 23%;
midres clips: 15%.
Change-Id: I0220e93a8277e6a7ea4b2c34b605966e3b1584ac
Fixes an issue when the LAST and golden is not used as a reference,
in which case its possible no encoding mode is set (since intra may be
skipped under certain codtions). Fix is to make sure intra is searched
if no inter mode is checked.
Issue can happen for temporal layer pattern#7 in vpx_temporal_svc_encoder.c
Change-Id: I5ab4999b2f9dbd739044888e0916b5ec491d966b
only the first 3 parameters can be aligned to 16 as required by __m128i,
make them all pointers for consistency.
since:
07c48ccfe Improve idct32x32_34_add SSSE3 intrinsics performance
BUG=webm:1384
Change-Id: I0324f701e723a27cb470036a180693ba8829d01d
shift the bsse[] member of the macroblock struct to the front to avoid
an incorrect offset (0) to the upper half of bsse[0] which leads to a
negative resulting in a crash. restrict this to visual studio versions
before 2015 (the bug was observed with 2013, fixed in 2015) to avoid any
potential cache impact on other platforms.
https://connect.microsoft.com/VisualStudio/feedback/details/2396360/bad-structure-offset-in-32-bit-code
BUG=webm:1054
Change-Id: I40f68a1d421ccc503cc712192263bab4f7dde076
Enable row-mt for SVC for real-time mode, speed >=5.
Add the controls to the sample encoders, but keep it off for now.
Add the control and enable it for the 1 pass CBR unittests.
For speed 7, 3 layer SVC, 2 threads, row-mt enabled gives about ~5% speedup.
Change-Id: Ie8e77323c17263e3e7a7b9858aec12a3a93ec0c1
- Split the inv txfm into three parts to avoid stack spillover.
- Function level speed improves ~12%.
- Use function and macro to remove some repeated code.
Change-Id: I14f5f072334fd766808cb52bf648df792e7379ee
this is similar to the x86 configuration and helps mitigate an issue
with a circular dependency between this function and the ssse3 variant
causing an outsized increase in binary size (~300K for chrome)
chrome.dll:
.text 255B000 -> 252B000
.data 7B000 -> 75000
-221184 bytes
BUG=chromium:697956
Change-Id: Ic95b142ecd62dd4f1795788aa27dd8fab59b708c
The 2 thresholds(i.e. partition_search_breakout_dist_thr and
partition_search_breakout_rate_thr) are used as the partition search
early termination speed feature. This refactoring patch made this
feature to be frame size dependent consistently throughout the code.
Change-Id: Idaa0bd8400badaa0f8e2091e3f41ed2544e71be9
Most are cosmetics changes.
Speed has no change with clang 3.8, and about 5% faster with gcc 4.8.4
Tried the strategy used in 8x8 and 16x16 (which operations' orders are
similar to the C code), though speed gets better with gcc, it's worse
with clang.
Tried to remove store_in_output(), but speed gets worse.
Change-Id: I93c8d284e90836f98962bb23d63a454cd40f776e
Add ppc, ppc64 and ppc64le on all_platforms and ARCH_LIST
Add VSX flags and check for -mvsx
Define empty setup_rtcd_internal
Add Altivec detection based on:
http://freevec.org/function/altivec_runtime_detection_linux
Detect VSX at runtime when enabled
Change-Id: I304f4d8c5fee0ff19b6483cd2e9cc50d6ddec472
Signed-off-by: Rafael de Lucena Valle <rafaeldelucena@gmail.com>
When eob is less than or equal to 135 for high-bitdepth 32x32 idct,
call this function.
BUG=webm:1301
Change-Id: I8a5864f5c076e449c984e602946547a7b09c9fe6
Fix the conditon for getting last_source when denoising is on.
This avoids unneeded scaling in the case of SVC.
No change in quality.
Change-Id: I32c1c2c9085104da51af8535716bcc4d55fb0f42
This may fix the time out failure of valgrind tests in nightly
since more coverages were added on row-mt.
Change-Id: Id9414e66d1a266602c7495243d9f5cb69e17ccdc
clear the entire array on error. the size used previously was equal to
the number of elements.
BUG=webm:1364
Change-Id: I2f2e16ed6e867f41d4774a5a8ac9cedaee11ce46
Reduce the level from 4 to 2.
This gives ~1-2% quality gain on RTC set, with small decreaee in speed (~1-2% on mac).
Change-Id: I7d959731badcee3d45b2f4a08efe378765016a13
- Split the transform into first half and second half.
- Reschedule the instructions to avoid stack spillover.
- Function level speed improves ~16%.
Change-Id: I166889840d23aa8a273eca00f6fbdae8b4566f35
Moves the def from vpx_encoder.h -> vpx_codec.h. The defined value
is changed as part of this move.
Adds the value to decoder capabilities when CONFIG_VP9_HIGHBITDEPTH.
Change-Id: I7d61fc821cda29f1e32bb9b2b9ffd3d83966e419
This reverts commit d3db846cc5.
This change causes a large drop in psnr (4-5db) on low framerate
difficult content (tested at 360/480p)
BUG=b/35804225
Change-Id: I8e90012d3b9c8a0cddb062ba93b01b36c0e0c0a0
From commit:
https://chromium-review.googlesource.com/c/441393/
On non-segment the set_vbp_thresholds() should be called
again to adjust thresholds based on content_state of superblock.
This was the intended behavior from 441393.
Small change in RTC metrics and speed.
Change-Id: I45e5fbdc4af74db76b3cb4f13074fcae0eb2219e
new_mt is a very generic name that will get obsolete soon enough.
Since this is exposed as a codec control, renaming it to row_mt to
signify row level paralellism. Also renaming the ETHREAD_BIT_MATCH
codec control to ROW_MT_BIT_EXACT.
Change-Id: Ic7872d78bb3b12fb4cf92ba028ec8e08eb3a9558
Re-organized the encoder threading tests and grouped tests into
4 parts. Added PSNR checking test to make sure the PSNR variation
is within a small range.
BUG=webm:1376
Change-Id: I09edb990236a87a4d2b2b0e1ceaf6c6435a35eff
vp9_highbd_block_error_8bit_c was a very simple wrapper around
vp9_block_error_c. The SSE2 implemention was practically identical to
the non-HBD one. It was missing some minor improvements which only
went into the original version.
In quick speed tests, the AVX implementation showed minimal
improvement over SSE2 when it does not detect overflow. However, when
overflow is detected the function is run a second time. The
OperationCheck test seems to trigger this case and reverses any
speed benefits by running ~60% slower. AVX2 on the other hand is
always 30-40% faster.
Change-Id: I9fcb9afbcb560f234c7ae1b13ddb69eca3988ba1
Only works for bitdepth = 8 when compiled with high bitdepth flag.
4x speed ups for handling 1:2 down/upsampling.
Validated manually for:
1) Dynamic resize for a single layer encoding
2) SVC encoding with 3 spatial layers
Results are bitexact with the patch and the speed gain (~4x) in the
scaling was verified.
BUG=webm:1371
Change-Id: I1bdb5f4d4bd0df67763fc271b6aa355e60f34712
The reduction showed improvement on RTC when aq-mode=3 is on.
Add that (cyclic refresh enabled) to the condition.
Only affects 1 pass CBR.
Change-Id: I5d0843002d8e31d7c165098a62e7a71146b08664
For speed 8 only.
3% speed up for QVGA and 6.3% for VGA on Nexus 6.
~3% avgPSNR decrease on rtc_derf and 2.9% on rtc.
Disabled for now.
Change-Id: I70133f1f6c804d663d594df437bfe7fdb0030d6a
The output needs to be aligned. Input is read with 'movq' not 'movqda'
so it is not expected to be aligned.
Change-Id: Ibd48a84c1785917a6a97c3689a05322abba486b4
Increase the variance partition thresholds for superblocks that
have low sum-diff (from source analysis prior to encoding frame).
Use it for now only for speed >= 7 or for denoising on.
Small change on metrics for rtc set: less than ~0.1 avgPNSR decrease
on RTC set, for both speed 7 and 8.
Change-Id: I38325046ebd5f371f51d6e91233d68ff73561af1
Use the simple block_yrd under certain conditions.
The optimization code is completed but the speed is still slower
(~6% on 720p) than the low-bitdepth build.
For now, use the more complex block_yrd under certain conditions
(always use it for speed <= 5, otherwise use it on key frames and for
bsize >= 32x32).
This gives about ~2-3% gain in quality for speed 7 on RTC set
(over high bitdepth build), with about the same encoder fps as the
low bitdepth build.
Change-Id: Ibe92a1945d0bd635f880befb4c815727df62d754
Modified the code to facilitate bit-match tests in first pass
Added unit-tests to test the row based multi-threading behavior for bit-exactness
Change-Id: Ieaf6a8f935bb1075597e0a3b52d9989c8546d7df
This change subtracts out low complexity intra regions that are also low
error in the inter domain, in the calculation of the frame prediction decay.
The rationale here his that low complexity regions (such as sky) do not imply
high prediction decay in the same way as high error intra or neutral blocks.
The effect of this is small in most clips but in a few clips it can be > 10%.
(E.g. In to tree)
Change-Id: If67ac23d17fca14285cad2defa464c61c9ea861c
vp9[_highbd]_quantize]_fp[_32x32] and vp9_fdct8x8_quant do not make use
of these parameters.
scan is used for C code and iscan is used for SIMD implementations.
Change-Id: I908a0ff7d3febac33da97e0596e040ec7bc18ca5
* changes:
quantize_fp_32x32 highbd ssse3: enable existing function
quantize_fp highbd ssse3: use tran_low_t for coeff
quantize_fp highbd sse2: use tran_low_t for coeff
- Replace the corresponding assembly code.
- No user level speed performance degrade.
- Unit tests passed.
Change-Id: Idd0c5a4bad4976f1617c34100cb46e75e3b961e5
This was created as part of the quantize_fp_ssse3 change. Both
functions use the same source file with different macro parameters.
Change-Id: I267050a559426a85955d215aa0aaca270439c5ab
The previous implementation confused bit/bytes/elements. It was using
'32' as the multiplier but that was mistakenly adopted because a 32x32
transform embedded the stride.
Change-Id: Ieeb867a332416b9a40580b5e7c9b20088e9e691a
The weight segment needs to only be computed once per frame,
so remove it from the funciton vp9_cyclic_refresh_rc_bits_per_mb(),
which is called within a loop inside vp9_rc_regulate_q.
Change-Id: Ia0e18b89abb97e42c466d4dbc47700d7f76555db
vp9_compute_qdelta_by_rate has almost 2% overhead in profiling on Nexus 6.
Reduce the calling of that function in speed 8 by estimating the delta-q.
Both rtc and rtc_derf show little/no change in avg psnr/ssim.
Encoding speed is 2~3% faster on Nexus 6.
Change-Id: If25933715783f31104a18a5092ea347b1221b5f5
This small change replaces the frame boost check in the arf group
length break out clause with a test against a prediction decay value.
The boost value is in fact partly dependent on the decay value but
this change means that the per frame boost calculation can be adjusted
without influencing the group length calculation.
The value chosen gives a close match on all the test sets with the previous
code (on average) but it was noted that a lower threshold was slightly better
for 1080P and up and a slightly higher value for small image sizes.
Change-Id: I4d5b9f67d5b17b0d99ea3f796d3d6202fd61ee0c
The function scale_sse_threshold() returns a threshold scaled
if necessary for use with 10 and 12 bit from an 8 bit baseline.
SSE error values would be expected to rise for the 10 and 12
bit cases where there are more bits of precision.
Hence the threshold used for the test should also be scaled up.
Change-Id: I4009c98b6eecd1bf64c3c38aaa56598e0136b03d
When eob is less than or equal to 38 for high-bitdepth 16x16 idct,
call this function.
BUG=webm:1301
Change-Id: I09167f89d29c401f9c36710b0fd2d02644052060
(Yunqing Wang)
This patch implements the row-based multi-threading within tiles in
the encoding pass, and substantially speeds up the multi-threaded
encoder in VP9.
Speed tests at speed 1 on STDHD(using 4 tiles) set show that the
average speedups of the encoding pass(second pass in the 2-pass
encoding) is 7% while using 2 threads, 16% while using 4 threads,
85% while using 8 threads, and 116% while using 16 threads.
Change-Id: I12e41dbc171951958af9e6d098efd6e2c82827de
This matches bitdepth_conversion_sse2.asm and produces substantially
better assembly. The old way had lots of 'movzwl' and 'shl' and storing
back to memory before loading into an xmm register.
Change-Id: Ib33e35354dfd691a4f8b1e39f4dbcbb14cd5302b
Clears up static clang analysis warning regarding divide by zero.
Trying to explain to the compiler how it's impossible to avoid
incrementing num_blocks at least once is difficult.
Change-Id: Ibaae43be572e5cd7a689b440dcd341c17d33443b
Where clang static analysis or gcc -Wmaybe-uninitialized warns of
uninitialized values, assign 0 to ints, MB_MODE_COUNT to
MB_PREDICTION_MODE, and B_MODE_COUNT to B_PREDICTION_MODE.
Assert that the modes have been changed from the invalid value by
the end of the function.
Change-Id: Ib11e1ffb08f0a6fe4b6c6729dc93b83b1c4b6350
While the new-mt mode is enabled(namely, allowing to use row-based
multi-threading in encoder), several speed features that adaptively
adjust encoding parameters during encoding would cause mismatch
between single-thread encoded bitstream and multi-thread encoded
bitstream. This patch provides a set_control API to disable these
features, so that the bit match bitstream is obtained in the unit
test.
Change-Id: Ie9868bafdfe196296d1dd29e0dca517f6a9a4d60
broken since:
c3f095c8b Merge "Fix to avoid abrupt relaxation of max qindex in recode path"
5f21aba4b Fix to avoid abrupt relaxation of max qindex in recode path
the original change pre-dated the addition of .clang-format
Change-Id: If5e399d9a805bcad9147360b13b36fbc8c560a7c
VBR method that allows a wider Q range for the first normal frame
in each ARF group and then centers the min - max range for the rest of
the arf group on the chosen Q value for that first frame.
This allows for quite rapid adjustment of the active Q range even if the
initial estimate is poor.
In some cases where the ARF frames themselves are tending to
undershoot but the normal frames are overshooting this can still give
net undershoot. This can be corrected by allowing a larger Q delta for
arf frames but is usually is a sign that the allocation to the arfs was to
high.
Change-Id: Icec87758925d8f7aeb2dca29aac0ff9496237469
Temporary fix until optimization work for block_yrd is completed.
This essentially reverts back to the state before the change:
https://chromium-review.googlesource.com/c/433821/
Compression loss is about ~5-6% on RTC set.
Speed-up (from using this simple/model-based block_yrd) over the low
bitdepth builds (which uses more complex block_yrd) is ~5% on 720p.
Change-Id: Ie0af9eb0d111e5595f587870c44f08317403b8d8
this prevents a rollover when tv_sec is a long:
signed integer overflow: 2776 * 1000000 cannot be represented in type
'long'
Change-Id: I03dc4476ee122b02e2856dad28358a20cf16a9f8
The RunQuantCheck() test on it exposes 16-bit overflow in stage 7 of
pass 2. Change to use saturating add/sub for both
vpx_idct16x16_38_add_neon() and vpx_idct16x16_256_add_neon() for high
bitdepth.
Change-Id: Ibf4c107a887553a52852cc582e28d38a5a5a2712
Add factor to increase varianace partition and ac skip thresholds,
under certain conditions (noise level and sum_diff), to increase
denoiser speed.
Change-Id: I7671140ef3598bf5f114a72623d68792bcd7b77b
Threshold for partitioning only affects VGA and lower res.
0.07% quality regression is observed in borg tests on rtc_derf
and 0.2% regression on rtc.
5.6% speed up for low res and 6.8% for VGA on Nexus 6.
Change-Id: If85a2919b48c991de66059c90f32ed06980452be
Fixed the following issue.
..\test\vp9_ethread_test.cc(69): warning C4805: '|=' : unsafe mix of type 'bool' and type 'int' in operation [C:\src\buildbot\test-libvpx\tests\dveCPjwhBE\.build-x86_64-win64-vs10\test_libvpx.vcxproj]
..\test\vp9_ethread_test.cc(69): warning C4800: 'int' : forcing value to bool 'true' or 'false' (performance warning) [C:\src\buildbot\test-libvpx\tests\dveCPjwhBE\.build-x86_64-win64-vs10\test_libvpx.vcxproj]
Change-Id: I37f897cf12a0b7500d2fcbac9e4615f08a83fdb4
Modified the encoding stage to have row level entry points with relevant
initializations and to access the token information at row level
Change-Id: Ife10e55a7c1a420ee906d711caf75002688d9e39
* changes:
hadamard highbd ssse3: use tran_low_t for coeff
hadamard highbd neon: use tran_low_t for coeff
hadamard highbd sse2: use tran_low_t for coeff
Clears up static clang analysis warning regarding a dead store. Only
declare 'c' when it will be used.
Change-Id: I1ac0fc7f94bc44da63938c63cd1efcd6b95e0eb3
This commit resolves the compression performance regression in
real-time encoding setting when high bit-depth mode is enabled.
The current solution temporarily disables the SIMD implementations
of vpx_satd, hadamard8x8, and hadamard16x16 in high bit-depth mode.
The commit makes the coding results bit-wise identical between
regular coding pipeline and high bit-depth at profile 0.
BUG=webm:1365
Change-Id: Icfb900821733749685370460a1a5a7e07f76f4bf
Clears a clang static analyzer warning where 'cols' is assumed to be
less than 0, preventing the for loop from executing.
The assembly already requires that the size be 8 or 16 (U/V or Y plane)
and cols is a multiple of 8.
Change-Id: Ica4612690ead1638c94cfe56b306e87f8ce644f9
In non-rd pickmode: Allow speed 7 to also use larger block size in
model_rd. Small change in behavior for speed 7.
Change-Id: I8c5523e424308e8f0bc71b3f6324dec42a464cc8
In non-rd pickmode: small change in behavior for speed 6 and 7.
Remove condition on HIGHBITDEPTH flag.
Change-Id: I360a13fcc313d72612fe9b918162ef4bb278cdea
Add Buffer features for:
Setting the buffer to the output of an ACMRandom function.
Copying a buffer.
Comparing two buffers.
Printing two buffers.
Change-Id: Ib53fb602451a3abdcee279ea2b65b51fbc02d3df
Skip denoising for blocks < 16x16, and for block = 16x16
skip denoising for low noise levels and width > 480 for now.
Allow for some speed-up in denoiser.
Change-Id: Ib46cefe4741962d145fa08775defea3a9c928567
(yunqingwang)
1. Rebased the patch. Incorporated recent first pass changes.
2. Turned on the first pass unit test.
Change-Id: Ia2f7ba8152d0b6dd6bf8efb9dfaf505ba7d8edee
Increase the qp-delta, mainly for low resolutions,
excluding case of very low bitrates.
avgPSNR/SSSIM gain of ~3-5% on rtc_derf set.
Small change on rtc set.
Change-Id: Ice03d04bd0340404d1957666ef154fd64fed0606
Affecting only speed 8.
Speed tests on Nexus 6 show 4% faster for QVGA and 2.4% faster for VGA.
Little/negligible quality regression observed on both rtc and rtc_derf sets.
Change-Id: I337f301a2db49a568d18ba7623160f7678399ae1
(Yunqing)
This patch added the missing initialization in temporal filter.
Borg test BDRate results:
PSNR: -0.019%(lowres); -0.013%(hdres);
SSIM: -0.001%(lowres); -0.010%(hdres).
Other q values gave comparable but no better results.
Change-Id: I7ad0c18b39e6f558342688e2fe1e12fdb133ce9b
This currently runs 1000 * 1000 = one *million* times which is quite
unnecessary. It's one of the slowest items in Jenkins and takes over an
hour for each of the larger transforms.
Change-Id: I01653b5e610683e1a2d778ec60cf5065562ab8db
Only for speed >= 7, and affects skipping of intra modes.
Threshold is set low for now, needs to be tuned.
Small/no difference in metrics on rtc clips.
Change-Id: If9bdbd43f08d1f80407cdd2e9e5e96780dcd2424
Added the multi-threaded first pass encoder unit test in VP9. The test is
to check if the new multi-threaded first pass encoder(namely, new-mt = 1)
still generates matching stats. In the unit test, the new-mt mode will be
turned on once the multi-threaded first pass implementation is checked in.
Change-Id: Ic21bb1a55c454f024cfd2b397a4c148cfe638218
For short_circuit set to level 1, skip newmv for 64x64 blocks if the
low temporal variance flag is set. Also modify threshold for 64x64 split
in variance partitioning.
Overall speed-up on noisy clips of 2-4%.
Only affect speed >= 7.
Change-Id: I384b3772007e84de6f8707e480d2ddf1fe1f907d
Avoid quality loss when copying partition of superblock with large motions.
Maximum consecutively copied frames can be set (currently 5).
Change-Id: I11c30575514f02194c0f001444cf4021609e5049
Also set the flag to 1 when exit early choosing 64x64 block
such that skipping new mv for golden works in these scenerios.
Change the size of prev_segment_id to the number of superblocks
to save memory.
Borg test shows quality regression of 0.012% on average PSNR
and 0.035% on SSIM.
Change-Id: I5014224c8617d439d35c66ece3fed9ae30b31d23
Adds an optional output framestats.csv file that prints comparions
per-frame instead of averaged over the entire clip. It prints
per-channel and combined metrics for SSIM and PSNR.
Change-Id: Id28dfade27bc5775b59a9d83cfe8b37d1d52b686
The fix relaxes the max qindex based on the data from previous loop of
coding if output frame size is greater than maximum frame size allowed
Change-Id: Iac1f63ec67559d68766e090a7cbb80b812b2560f
before calling vp9_apply_encoding_flags() which may crash if the
resolution was invalid. this is the same change as:
c0523090b vp8e_encode: check validate_config return
BUG=https://bugzilla.mozilla.org/show_bug.cgi?id=1315288
Change-Id: Icd2aab322422e83d3a778fca6d7789e5000239d7
If enabled will compute source_sad for every superblock on every frame,
prior to encoding. Off by default, only on for speed=8 when
copy_partition is set.
Change-Id: Iab7903180a23dad369135e8234b7f896f20e1231
Used with --framestats=file.csv. Currently prints raw codec QP (not
internal 0-63 range) and bytes per frame.
Change-Id: Ifbb90129c218dda869eaf5b810bad12a32ebd82d
Avoid many visual artifacts. Compression quality is improved by more
than 1%. Encode speed is about 4% for QVGA and 6% for VGA faster on
android.
Change-Id: I4dd0a81429ddf7efdef1e80a191da5fb8de8e8af
Renames SSIM to VpxSSIM as an upscaled weighted SSIM metric, then prints
Y, U and V channels unweighted as well as a weighted but not scaled SSIM
score that's 8/1/1 parts Y/U/V (same as VpxSSIM).
Change-Id: Iff800cc8f145314eeb1a9b4af1e11a25bec095ca
For speed 8, it speeds up the encoding on android by 6% for QVGA and
7.4% for VGA with the new threshold. Overall PSNR is improved by 0.667
for rtc.
Change-Id: I4a644560b32c0b5b4e9f49ffb953d000413a3732
If enabled denoiser will only denoise the top spatial layer for now.
Added unittest for SVC with denoising.
Change-Id: Ifa373771c4ecfa208615eb163cc38f1c22c6664b
This commit reworks the SSSE3 implementation of the forward 8x8
2D-DCT. It uses a cyclic rotation approach to the temporary xmm
registers. It reduces the average cycles from 158 to 154. The SSE2
version uses 169 cycles.
Change-Id: I1b79b9642aae0ed3fb3cefb5b70246e6de5d5caa
When aq=3 mode is on and the gf_cbr_boost is set: make sure golden frame
is always refreshed, and don't incorporate segement cost in qp setting
on the boosted golden frame.
Better performance on RTC set with gf_cbr_boost on,
for example with gf_cbr_boost=50, gains from ~0.5-3%.
Change-Id: Ie811f5e4d444ff3320bd6e2c1745b2c4c09a8460
Quality improved by 1.866 and 0.386 for two noisy clips (dark720p and
marcooffice720p), respectively.
Change-Id: Ib33a7672ae9ca53da156208f7cd13f07b5543e44
Avoid the qp-clamping on gf/alt frame if gf_cbr_boost_pct is set.
Change only affect CBR mode when gf_cbr_boost_pct is set.
Change-Id: I0655ed4f2b047c8ed1ed33a070c17960ad776704
This was much more amenable to optimization than the across filter.
Speedup of almost 2.5x
BUG=webm:1320
Change-Id: I49acc0f9cb2e7642303df90132cbc938acade4c4
Speed test shows 25% gain on vpx_idct16x16_256_add_neon(),
and vpx_idct16x16_10_add_neon() got trippled.
Change-Id: If8518d9b6a3efab74031297b8d40cd83c4a49541
The speedup is pretty poor. I would be concerned except the SSE2 is
worse:
Existing SSE2 improvement: 22%
New neon improvement: 35%
BUG=webm:1320
Change-Id: Ied598a261134aa6cbe69f96f58589d2bae17bf62
Increase the boost threshold below which GOLDEN update will use same
rate correction factor as INTER_NORMAL.
Improves performance when gf_cbr_boost_pct is set (between 0 and 100)
in CBR mode.
Change-Id: I9f54cc18664786a100b13a416b7137ae03bd0cab
Set short_circuit_low_temp_var to 3 for speed 8 for all res.
No strong visual difference on all clips.
Change-Id: Ia6d9a314291ab1c14d5421bbdd769974083aeb2a
Constraints on encoder config:
-target_bandwidth is no larger than 80% of level bitrate limit
-target_bandwidth * (1 + max_over_shoot_pct) is no larger than
88% of level bitrate limit
-min_gf_interval is no smaller than level limit
-tile_columns is no larger than level limit
Constraints on rate control:
-current frame size plus previous three frames' size is no larger
than the CPB level limit
-current frame size is no larger than 50%/40%/20% of the CPB
level limit if it's a key/alt-ref/other frame.
Change-Id: I84d1a2d6d6e3c82bfd533b3309ce999cfaba2c8b
The source sad could be used to copy the partition without going into
choose_partitioning function to speed up vp9 encoding. Computing source
sad takes little time. Speed test on Android and Linux shows little
encoding time gain (less than 1.4%).
Turned off for now since partition copy is turned off.
Change-Id: I61c9d5b8f22329760cb29a4ee30a7f9c232ce8d3
vp9: Set short circuit to level 3 for VGA for speed 8. Also change the
threshold_32x32 to 5/8*thresholds[1] to improve quality regression
caused to VGA clips.
Change-Id: Ia1590e91e7cb22be78d5b85013387bb1be4272e3
Comment out check on buffer underrun, as it currently fails
on some of the svc tests.
Also cast the update of bits_in_buffer_model_, as this can
go negative now due to the buffer underrun.
This fixes the issue in #1352.
BUG=webm:1350
BUG=webm:1352
Change-Id: Ibd4ef23921daf09e5c15b000aca904aa4573599c
For out-of-range cases, returned UINT_MAX instead of INT_MAX in the
sub-pixel mv search to be consistent with the "uint32_t" return type.
Change-Id: I8e206d771228c13d89bafbbe9f14722c8ecc6a7a
The function 'vp9_find_best_sub_pixel_tree_pruned_more' is modified
to return INT_MAX for handling invalid MV cases from UINT32_MAX.
yunqingwang:
patch 3: rebased on top of the tree.
patch 4: The return type of vp9_find_best_sub_pixel_tree* was changed
to uint32_t to fix ubsan warnings. Changing UINT_MAX back to INT_MAX
was not quite right. Patch 4 modified vp9_temporal_filter.c to accept
uint32_t.
(Note: Inconsistency exists in vp9_find_best_sub_pixel_tree*, which
will be fixed in a separate CL.)
Change-Id: Ib1a79dc2aa41ea6335c21669c76883cdbb7e0535
This reverts commit f0b491a524.
This change results in unsigned integer overflows (as reported by
-fsanitize=integer) in datarate_test.cc,
for many of --gtest_filter=VP9/DatarateOnePassCbrSvc.OnePassCbrSvc*:
unsigned integer overflow: 167198 - 185560 cannot be represented in type
'unsigned long'
As the encoder didn't change, but the input with the change to
(correctly) use Y4mVideoSource, this revert is merely masking the issue.
BUG=webm:1352
Change-Id: Iecd9a6c83b3fca67c566732a5c92d36193cc2060
Fix compile warnings about implicit type conversion for
target=armv7-android-gcc in vpxenc.c.
BUG=webm:1348
Change-Id: I9fbabd843512f2a1a09f4bb934cd091e834eed9c
Comment out check on buffer underrun, as it currently fails
on some of the svc tests.
BUG=webm:1350
Change-Id: I73c88b800cdcc06bd2f900f7b7e2a5fd08248065
When source frame is altref, we only do zero-mv mode, so we can skip
the find_predictors(). No change in compression.
Small speed gain, ~1%.
Only affects 1 pass vbr with lookhead altref, for ytlive with
the macro flag USE_ALTREF_FOR_ONE_PASS on.
Change-Id: I9318c5da8521f017bf54919cd652438b3a6313d1
Remove superfluous test. Produces a small improvement in instruction scheduling.
Measured a 1% to 1.5% reduction in execution time for routine vp9_optimize_b
with different compilers.
No change in behavior.
TEST=Verified that encoded files match bit for bit, with and without this
change.
BUG=b/33678225
Change-Id: I2bf248d4c25fc0256147d7a8766ff9108ae9cba3
Add feature to copy partition from the last frame.
The copy is only done under certain conditions that SAD is below threshold.
Feature is currently disabled, until threshold is tuned.
Feature will be initially used for Speed 8 (ARM).
Under extreme case of always copying partition for speed 8:
Encode time is reduced by 5.4% on rtc_derf and 7.8% on rtc.
Overall PSNR reduced by 2.1 on rtc_derf and 0.968 on rtc.
Change-Id: I1bcab515af3088e4d60675758f72613c2d3dc7a5
Simplify address arithmetic on token_costs to reduce the number of generated
instructions that are used for address arithmetic inside routine
vp9_optimize_b. It also helps improve instruction scheduling depending on
compiler and optimization level.
Measured a 9.3% reduction in retired instructions and 5.3% reduction in
execution time for this routine with GCC v4.8.4 and optimization flags -O3,
and a reduction of up to 11.6% in execution time with other compilers.
No change in behavior.
TEST=Verified that encoded files match bit for bit, with and without this
change.
BUG=b/33678225
Change-Id: I6098650fb5cd2aa04e014fe6e68ca20761f3a21f
relocate the assignment to 'in' outside of the for loop. this quiets a
spurious warning in visual studio builds since:
86e340c enable vpx_idct32x32_1024_add_neon in hbd builds
+ give the variable a more descriptive name
BUG=webm:1294
Change-Id: I5c3da5c7939621477e0fc0ad3a1b2a3045c5bffd
Correctly set interp_filter to SWITCHABLE for INTRA mode.
Also reduce threshold on noise level for re-evaluating zeromv.
Change-Id: Id32c01e193209fb380aa07204f0be3babf29f70a
For when denoising enabled: change condition to enable
the recheck_zeromv_after_denoising for only very high noise level.
This is causing an issue, so enabling it for very high noise
to effectively shut it off.
Change-Id: Ic40d6025f3f398338cedd270d17c0ccd9a3daa84
The new test is causing valgrind failures:
[ RUN ] SSE2/VpxPostProcDownAndAcrossMbRowTest.CheckCvsAssembly/0
==28923== Invalid read of size 16
28923== at 0x724016: ??? (deblock_sse2.asm:146)
Disable during investigation. The test is new but the code is not.
Change-Id: I5521e5fd48a595e3798b833bf7e3cc97b81c1975
To avoid decode performance hit of 2% when running on hyperthreaded
cores.
This patch only uses the mutex's when we are running tsan.
This is safe because 32 bit operations like read and store are atomic
on all the platforms we care about. Tsan warns about race situations,
but in this case either situation ( read occurs before write or write
before read) the worst case is that we go around one extra time in the
loop. So the ordering doesn't really matter.
That said a few other things have been tried :
for instance as per here:
webrtc/base/atomicops.h#52
In this patch they use:
__atomic_load_n(i, __ATOMIC_ACQUIRE);
__atomic_store_n(i, value, __ATOMIC_RELEASE);
This code works on gcc, clang ( replacing protected write and read), and
avoids tsan errors. Incurring no penalty in performance. In C11 its
replaced by straight atomic operands.
However there is no equivalent in the visual studio's we support as
int32 on all windows platforms is already atomic. To avoid tsan like
warnings on windows we'd need to use interlocked exchange and the
end result doesn't gain us any thing.
Change-Id: I2066e3c7f42641ebb23d53feb1f16f23f85bcf59
Implement vpx_post_proc_down_and_across_mb_row in NEON.
Runs about 6-7x faster than C.
BUG=webm:1320
Change-Id: Ic5c7d3552a88cfcf999ec5bf2bd46fee460642c2
The flag USE_ALTREF_FOR_ONE_PASS allows for alt-ref lookahead
in 1 pass vbr (from https://chromium-review.googlesource.com/#/c/365498).
This change is to make sure this macro flag only has effect if
the config flag cpi->oxcf.enable_auto_altef is also on.
No change in ytlive encoding, as USE_ALTREF_FOR_ONE_PASS is not
yet enabled.
Change-Id: I1a69681e4a15c5244581a3dab4587fca08f02e0f
Reapply this patch:
ff0107f Amend and improve VP8 multithreading implementation
Amended the patch to add a unit test, and fix an asan error.
BUG=webm:851
Change-Id: I6572c03256169c64e80248bf5a5e99f59a2fc93c
use_base_mv assumes 2x2 scaling, so fix is to shutoff
this feature unless spatial scale factors are 2.
Added svc unittest for 2 spatial layers with 5x5 scaling,
which generates the issue without this fix.
Also fix some settings in svc unittest:
let the speed setting vary (from 5 to 8), and enable static threshold.
BUG=webm:1344
Change-Id: Idfd0a6c633c21b49a0479601506302cfe974e30e
Set #threads to default 1 for all streams, change bit allocaton
for 3 temporal layers, and enable denoiser on middle resolution layer.
Change-Id: I4a57adbfdb2c319002b8f3cf359613842dc00d75
after:
2d3d95f enable vpx_idct16x16_256_add_neon in hbd builds
reorder INCLUDEs and fix indent of IF/ENDIFs
remove vpx_config.asm to avoid multiple symbol definitions in windows
builds and shift idct_neon.asm.S to the top to allow use of
CONFIG_VP9_HIGHBITDEPTH in the export list.
Change-Id: I0dacfbae62a6ec8fe4a26940c1a52da2dfad2029
One of the first pass stats "new_mv_count" is no longer used in VP9,
and is removed. This also makes it easy to implement a multi-threaded
first pass. This change doesn't affect the coding performance, which
has been verified by borg tests.
Change-Id: I4c7c7bf9465fda838eb230814ef0c631c068c903
1. Use correct projections when copying real dct/quant outputs.
2. Remove local random number generator and combine loops.
3. Quantization with minimum allowed step sizes instead of maximum.
This may generate larger inputs.
Change-Id: I154afc26230c894d564671cff4b8fd5485b69598
vpx_idct16x16_256_add_neon_pass1, vpx_idct16x16_10_add_neon:
this was a constant 8 in all cases meaning the results are stored
contiguously, this allows the number of stores to be reduced.
Change-Id: I7858a0a15a284883ef45c13dfd97c308df9ea09e
Use the segment weight factor based on the target (cr->percent_refresh)
if it less than the current estimate (avergae of past usage and target).
Small improvement at low bitrates.
Change-Id: Iba8fd909e203f94458901366d3a991f7ea854d49
the expansion of findstring and rtcd_dep_template_CONFIG_ASM_ABIS needs
to be deferred until the block is parsed as makefile syntax rather than
eval time where rtcd_dep_template_CONFIG_ASM_ABIS will be unset. this
ensures vpx_config.asm is properly created.
Change-Id: I7c38c6c082da78397936467482789dd468adc316
* changes:
Update vpx_idct4x4_16_add_neon() to pass SingleExtremeCoeff test
Refine 8-bit 4x4 idct NEON intrinsics
Add idct speed test.
Update partial_idct_test.cc to support high bitdepth
force this to be created before any other .S files. this change
additionally removes the file from the source list as it doesn't need to
be compiled on its own.
Change-Id: I6b4cd56ef6059d08f75f06fb749cddf76e0e165e
Increase the motion threshold and qp-delta for segment#2 boost.
This can increase the frame-drop at low bitrates, but generally
better spatial quality.
Only affects real-time mode with aq-mode=3, at very low bitrates.
Change-Id: I5ccb784667f70d0c27d369806b93b1f93d5605d1
For some filter level, the C/MSA doesn't match SSE2. Part of unit tests
are disabled. They will be re-enabled when C/MSA funcs are fixed.
BUG=webm:1321
Change-Id: Ib16b98b5eecb15d2252aa4ea267b782ee2b27533
replace downloads.webmproject.org with the canonical
storage.googleapis.com/... form. this appears less likely to fail when
dealing with multiple concurrent connections.
Change-Id: I0dcbd04df9e4057fa851f458b3ef7e3589f1f2f1
best_sub8x8[1] won't be used meaningfully when is_compound is false, but
may trigger an msan warning as the value is copied around and later
clamped.
BUG=667044
Change-Id: Icc24c3b72cdb550bebea44d4aaa4ff8bf3fbab56
Use the same feature as https://chromium-review.googlesource.com/#/c/411327/,
but allow it to be used for speed = 6 and 7, where
short_circuit_low_temp_var = 1.
Speed up of ~2-3% for speed 7, with little/no loss in compression.
Change-Id: I263a0f261ad9929034392d68f0153dc6376fdb5f
Remove unnecessary "virtual" before some functions. Change *_btm_* in
variable names to *_bottom_*.
Change-Id: Ifd4ce667537617f451cdfed47dd8c48817fd983b
VpxEncoderThreadTest was taking a very long time for some runs and
timing out a lot. This is an attempt to split the test into runs
that can be run nightly ( speeds 2 through 9) and runs that can
be run weekly ( speeds 0-1 ).
Change-Id: Iee6f61a561006d3a30381dd3b52b9a4dce07a70c
This is a boolean value that is written into bitstream, any value other
than 0 or 1 could have led to unexpected behavior. This commit fix the
issue by adding validation of the value to make sure it is boolean.
BUG=webm:1339
Change-Id: I2d3e69e8dbefcab9a0db9cb39a91a40ce531c5a1
fixes reloc errors like:
R_X86_64_PC32
vpx_dsp/x86/deblock_sse2.o:
requires dynamic R_X86_64_PC32 reloc against 'vpx_rv' which may overflow
at runtime
Change-Id: I218fc0e7c8258197f890d395f335e5a4fe82dccb
tests with 'Large' in the name are reserved for slow running tests which
may not be run on all platforms
Change-Id: I2a7d6dd46b29b50469893e46433844132fb727c2
This runs multiple encodes and decodes of vp8 and vp9 in parallel,
with so many threads that problems with synchronization can show up.
Change-Id: I2b297e7f43d1e741323c7ad9f50a3931ae609f16
Add a new, more aggresive short circuit: short_circuit_low_temp_var = 3 to skip
golden of any mode when variance is lower than threshold for low res.
This change only affects speed = 8, low resolution.
Metrics for avgPSNR/SSIM on rtc_derf (low resolution) show loss of
0.27/0.31%.
On Nexus 6, the encoding time is reduced by ~2.3% on average across all
low-res clips.
Visually little change on rtc_derf clips.
Change-Id: Ia8f7366fc2d49181a96733a380b4dbd7390246ec
this removes the need for __STDC_LIMIT_MACROS which is defined in
vpx_integer.h, but may be preceded by earlier includes of stdint.h;
fixes build with the r13 ndk
Change-Id: I3950c8837cf90d5584a20ce370ae370581c2182c
avoids the definition of min/max macros in headers that may appear in
c++ unit tests. the codebase uses VPXMIN/MAX for this purpose in any
case
Change-Id: I2b679b045d64fb34fd8780f704e3caf10a758d82
use 'android/cpufeatures' rather than 'cpufeatures'; this matches the
documentation, fixes compilation with r12b/r13 and still works with
older ndks.
Change-Id: I2f34233c164e6d4d46428f8905d5502cea4288a2
Changes only affects speed = 8 for low resolutions.
Metrics for avgPSNR/SSIM on rtc_derf (low resolutions) show loss of
0.5/0.6%.
On Nexus 6, the encoding time is reduced by ~5.9% on average across all
low-res clips.
Visually little/no change on rtc_derf clips.
Change-Id: I68dd50e558d72dcc1af8317d224bfae5e3bd872d
This commit enables asymptotic closed-loop encoding decision for
the key frame and alternate reference frame. It follows the regular
rate control scheme, but leaves out additional iteration on the
updated frame level probability model. It is enabled for speed 0.
The compression performance is improved:
lowres 0.2%
midres 0.35%
hdres 0.4%
Change-Id: I905ffa057c9a1ef2e90ef87c9723a6cf7dbe67cb
this was enabled in:
3ae2597 idct,NEON: add a tran_low_t->s16 load adapter
+ enable it for all NEON configs, both intrisincs and assembly versions
exist
BUG=webm:1294
Change-Id: I339088b2a398200f95658d040034fb9b2a7c8ce0
usage of the vp8 versions was removed in:
3f72509 vp8: remove VP8_SET_DBG* control support
vp9 had the usage stripped even earlier.
Change-Id: I978142eb6492552cd29c9c6feb1e89acfc5f7b84
this was enabled in:
3ae2597 idct,NEON: add a tran_low_t->s16 load adapter
+ enable it for all NEON configs, both intrisincs and assembly versions
exist
BUG=webm:1294
Change-Id: Iaade219e9d1de7b69423670d3ea6271b0965e068
idct4x4 and idct8x8 were universally enabled for high-bitdepth builds
in:
3ae2597 idct,NEON: add a tran_low_t->s16 load adapter
BUG=webm:1294
Change-Id: If142afb169c48728cc4b222e7c41aa4a63f95f0f
replace load_and_transpose_s16_8x8() in idct32_6_neon() with a separate
load_tran_low_to_s16() and transpose_s16_8x8(). the combined function is
used in idct32_8_neon() where the input is the correctly sized output
from the earlier stage.
BUG=webm:1294
Change-Id: I4257c4b3a421b2cf5d13651f966eee0680ef98a9
For noisy content, be more aggressive in skippping some blocks for
delta-qp to reduce noise pulsing artifact. Also treat frame boundary
case when dimension is not multiple of superblock size/64.
Only affects non-screen content case, and when source noise
is measured to be high (at least level kMedium).
Change-Id: Ib13a2a20ed1ce37ff3c44d95c3ef2635fd695222
This uses the same sdx4df pointers as vp8_diamond_search_sadx4 and
should therefore target the same optimizations.
See e4ddf9db6a
Change-Id: Ic298e9b25c34bbe6b7a0799509355b0addb56675
The matching on ads2gas_apple.pl is too liberal and catches
CONFIG_EXTERNAL_BUILD and CONFIG_INTERNAL_STATS because they have RN in
the names.
The RN renaming feature is not used in any existing assembly files. It
was used in some armv6 files but they were removed.
Change-Id: Ib65abf1947d3e89f0d1584e2a5de399d24008f95
Two functions do not pass this test:
vpx_idct8x8_64_add_ssse3
vpx_idct8x8_12_add_ssse3
The test has been modified to avoid triggering an issue with those
functions but they still must be investigated.
BUG=webm:1332
Change-Id: I52569a81e8e6e0b33c4a4d060d0b69c3fc4f578e
Add condition that usable_ref_frame > LAST.
This is to avoid potentially skipping all last-nonzero mv modes,
if golden is used as a reference but skipped completely for the
current block.
This has no effect currenty, as we always consider testing golden
mode for each block.
Change-Id: I3182cf44664081935a90ed43aa7b32e710e60e22
Fixed formatting bug introduced by the fix to BUG=webm:1322
( Iedc4477aef1746aa0a4f84d88a1156296fd3ba87)
Change-Id: I715ee446c0e8584967ab87ba4e355759dd394187
vp9_init_macroblockd() resets the error_info to cm's global copy; this
needs to be set to the thread-level target to avoid jumping to the
incorrect stack, resulting in hang or crash.
broken since:
1f4a6c8 vp9/tile_worker_hook: add multiple tile decoding
includes v1.5.0, v1.6.0
BUG=629481
Change-Id: Icbf1696b25ba8c479e845fbf227b3c3ca73542f5
vpx_config.asm and idct_neon.asm.S are required since:
3ae2597 idct,NEON: add a tran_low_t->s16 load adapter
Change-Id: If5959a25edb370dd7dcdca71c96e9a5aad0840ce
enable idct4x4* and idct8x8* which are compatible for 8-bit decodes in
high-bitdepth mode. the adapter narrows 32-bit input to 16, whether the
expansion can be avoided at all in this case remains a TODO. roughly
matches sse2.
BUG=webm:1294
Change-Id: I3ea94e5a2070dfd509b5de0c555aab4e1f4da036
A past patch made it so that every frame that had a decode error
caused a corrupted frame to be counted. Unfortunately it was possible
to get both a decode error and a corrupt frame for the same frame
and thus double count an error. This code makes that impossible.
Change-Id: Iea973727422a3bf093ffda72fa358a285736048b
Permits skipping 0, 1/2 or 3/4 of the frames, corresponding to
temporal layers 2, 1 and 0 of a 3-temporal-layer encoding. 1/2
corresponds to TL0 in a 2-layer encoding.
Change-Id: I7f6d131f63707e5262fc67d111bfb3a751ede90d
Allow for passing in the layer bitrates at command line.
Fix to allow passing in bitrate for each spatial-temporal layer.
Change to some default values for 1 pass cbr mode:
spatial scale and qp-max/min.
Small fixes to some build warnings.
Change-Id: I3f9a776262712480a6570bb863a835b2fc49935a
This change is a step in a larger change to the way boost and interval are
determined for ARF and Key frames.
This patch contains some pluming for the general case but focuses on the
key frame boost calculation. This now relies more heavily on the rate at
which the error score increases between the primary and secondary reference
frame. This seems to be less fragile when dealing with different frame sizes.
For example larger image formats tend in the first pass to see a higher
% of intra coded blocks and the use of this number in calculating the frame
decay factor was leading to much lower boost numbers for 4K, for example,
than the same clip coded at 2K.
This change does give overall gains but they are MUCH larger for the 4K Netflix
set. For the 4K Netflix set the average gain is around 3% with some clips > 20%
whereas for the same set at 2K the average gain is 0.5-1%.
In general for small image formats the boost is most often reduced a little whereas
4K clips the boost is increased. There are some -ve cases such as Akiyo at 352x288
where the reduced boost hurts the metrics, especially for SSIM, even while
the set as a whole improves. This is most notable at very low Q and may be the
subject of a future patch.
Some common code for KF and ARF was separated in this patch for the purposes of
tuning but may later be re-merged if appropriate.
Change-Id: Iaa15ac5a58d2be89181100d95cef6a8dc4b12d0d
Fixes a case where recode is not triggered based on the value
of maxq passed into the recode loop test function.
BUG=b/32375284
Change-Id: I15ad985d0525c68e0443cfaf842440d2754b2266
The result of the transform is added to the destination buffers. In the
existing tests the destination buffer is always empty so that portion of
the code was never exercised.
Change-Id: I1858c4fed2274f1b9faf834d2ba4186a4510492a
Switch to using correctly sized inputs and outputs. This simplifies
adding tests with varying strides.
Change-Id: I716a0d8173dcf6a86d56656ac9d3101b7ec27642
Removed a couple of adjustments that no longer move the needle
much but complicate the process of tuning.
Change-Id: Ie320f5cf155e6aac14a4757ea9ada2cd59f27590
Modified the encoder multi-thread test so that it included cpu-used=0 and
frame-parallel=0.
frame_parallel_decoding_mode is 1 by default, which disables probability
updating and gives lower encoding quality. Current VP9 multi-threading
encoder and decoder support probability updating. To test this part, we
should turn on it in the unit test, namely, setting frame-parallel to 0.
Change-Id: Ia1f86e01f0de628f50d819ae31509de3e1b6c755
This patch modified the motion search counts used in:
https://chromium-review.googlesource.com/#/c/305640/
These 2 counts were originally added as thread data, and used to
make decisions in motion search. The tile encoding order can be
inconsistent while using different number of threads, which can
cause bitstream mismatch. Here moved them to tile data to solve
the issue.
BUG=webm:1322
Change-Id: Iedc4477aef1746aa0a4f84d88a1156296fd3ba87
dst += stride behaving better with gcc/clang
Expanding inline function dc_SIZExSIZE() save intructions for
vpx_dc_predictor_SIZExSIZE_neon().
Change-Id: Id0ccbd58b6a31df539141fd33bdf28633339150d
A failure to decode is most likely equivalent to a corrupt
frame for the purpose of returning a failure.
Change-Id: Ie53db2b8130b40b725841f5f7a299d63aa56913d
Re-use the tile worker threads to pack the bitstream in parallel
on a per-tile basis. Restricting this to real-time only for now
(further testing is needed to ensure this does not make 2-pass
worse in any case).
BUG=webm:1309
Change-Id: I8a80da7c5089b837d0df79a5c49d5e3022dfc8ec
In variance partition low resolutions may use varianace based on
4x4 average for better partitioning.
Increase the threshold for doing this at speed = 8.
Improves speed by ~5%, with little loss, < 1%, on RTC_derf set.
Change-Id: Ib5ec420832ccff887a06cb5e1d2c73199b093941
the intrinsics are neutral to ~20% faster on cros/android
devices when using gcc-4.9/clang-3.8.1 and gcc-4.9/clang-3.8.x from the
r13 ndk. neutral results typically came with gcc-4.9 while larger
positive gains were achieved with clang 3.8.x.
BUG=webm:1303
Change-Id: I4d31f9c017944681b881493525d4573a7a5b1e16
Control already exists for vp9, adding it to vp8.
Usage is only when error_resilient is off.
Added a datarate unittest for non-zero boost.
Change-Id: I4296055ebe2f4f048e8210f344531f6486ac9e35
This reverts commit 9e8efa5b18.
this change causes ubsan warnings, failures in
vpxenc_vp9_webm_rt_multithread_tiled
BUG=webm:1309
Change-Id: I020c7be985c771bfff4b3de1afe51cc8edb980da
git log --no-merges 32d5ac4..9732ae9
9732ae9 EbmlElementSize: quiet uint64->int32 conv warning
da04eba SetProjectionPrivate: quiet uint64->size_t conv warning
6db32d5 mkvparser,Projection::Parse: fix int->bool conv
3bb0dfa cosmetics: fix a couple lint warnings
0e179d6 update .clang-format
fc5f88d Fix temp files being left on system.
c04a134 Add support for overriding PixelWidth and PixelHeight.
c0160e0 Add support to explicitly set segment duration.
02bc809 Add support to estimate file duration.
c97e3e7 Add support to output sub-sample encryption information.
26f4344 MakeUID: quiet unused param warning in Android builds
d6af52a Change check to fix compile error.
1720020 webm_parser: Add Mesh value for ProjectionType
78f2c5a webm_parser: Use ./ prefix for includes
da62f65 webm_parser: Remove webm/ prefix from public includes
e15e8f2 webm_parser: Update README build instructions
5023f2b mkvmuxer: Fix Colour::Valid()
cf16204 mkvmuxer_tests: Actually test cue points in the cue point test.
93e9fb3 Validate Colour element values.
8036925 mkvparser_tests: Add Projection element test.
f52d38c mkvparser_tests: Add Colour element test.
826436a mkvparser: minor SeekHead::Entry clean up.
24fb44a mkvmuxer_tests: Add Projection element test.
1e0a8ea mkvmuxer_tests: Add Colour element test.
0278616 mkvmuxer: Colour accessors/mutators.
2346f8f Add mkvparser wrapper functions.
54d6b6b webm_info: Add Projection element support.
65fee06 mkvmuxer_sample: Add support for Projection element.
9a3f2b5 mkvparser_sample: Add support for Projection element.
41e814a mkvparser: Add Projection element support.
483a0ff mkvmuxer: Add Projection element support.
676a713 Add support for the Projection element
725f362 mkvmuxer: Fix memory leak when Colour is set multiple times.
fa182de mkvparser_sample: Add output of audio track codec private size.
8f521f2 mkvparser_tests: Add invalid BlockGroup test.
39137d7 Remove docs saying binary elements default to 0
80685d3 Do not skip over unknown elements at the root level
c147504 Fix legacy Makefile.
58711e8 mkvparser_sample: Fix version info string.
837746f mkvparser_tests: Add invalid block test.
207cd80 Disambiguate sample sources and targets.
a112d71 mkvparser_tests: Refactor invalid file loading code.
5dea33e Disambiguate test source and target names.
125049e parser_tests: Add another truncated chapter string test.
1de8d4c parser_tests: Add truncated chapter string test.
ff8c2b6 parser_tests: Move cue validation to test_util.
4b0690f parser_tests: Add invalid lacing test.
9828e39 mkvmuxer: Set default doc type version to 4.
5495a59 webm_parser: Reference more files in CMakeLists.txt.
0c0ecd0 vpxpes_parser: Add start code emulation prevention support.
639a4bc webm2pes: Remove debug printfs().
9a51102 webm2pes: fflush() in the correct conversion function.
dc7f155 webm2pes: Track total bytes written.
d518128 webm_parser: Enable usage of werror.
e1fe762 webm2pes: Add test for mux/demux of large input.
1b24a79 vpxpes_parser: Read and store PTS when present.
6cf0a0f vpxpes_parser: Store frame payloads.
25d2602 webm_parser: Convert style to match the rest of libwebm
24be76d webm2pes: Replace VpxFrame with VideoFrame.
b451c3b Add a basic video frame storage class.
05c90eb libwebm_util: Clarify error text in superframe parser.
e6415af webm2pes: Make WritePesPacket() a public method.
8f840dd webm2pes: Move frame read out of PES packet write method.
448af97 webm2pes: Restore frame fragmentation support.
f8bb714 cmake: Integrate new parsing API and tests.
cb8ce0b Add a new incremental parsing API
900d322 vpxpes_parser/webm2pes: BCMV and PTS fixes.
4b73545 webm2pes: Add start code emulation prevention.
82903f3 Add column tiles and frame parallel to webm_info
5d91edf style_clean_up: Remove unnecessary parentheses
a95aa4b vp9_level_stats: correct total_uncompressed_bits_ calculation
f46566f mkvreader: Fix shorten-64-to-32 warning in 32 bit builds.
76630ca mkvwriter: Fix shorten-64-to-32 warning in 32 bit builds.
a8ffbd4 webm2pes: Fix format specifier warnings.
faf89d4 Add MaxLumaSampleRate grace percent to stats.
d31e6c9 Fix profile 2 in vp9_header_parser.
bd3ab3a Add flag to estimate last frame's duration to stats.
c182ed9 Fix lint issue in hdr_util.h
cc62ecd Add test for Cluster memory leak
196708a Change MaxLumaSampleRate to be based on frame resolution.
cbd676b mkvmuxer: Fix leak when a Cluster isn't finalized
9a235e0 mkvmuxer: Set doctype to matroska when muxing non-WebM codecs.
47f2843 Add parsing support for new features in CodecPrivate.
e3c9576 Add VP9 level output to webm_info.
5cf549f cmake: Log compiler flag at check time.
bbaaf2d Add class to gather VP9 level stats.
8bb68c2 Add file to parse data from VP9 frames.
296429a Add support to parse VP9 profile.
df3412f Add support for setting VP9 profile and level to sample_muxer.
87832d4 mkvmuxer: Fix Segment::Finalize in kLive mode
6df3e56 mkvmuxerutil.hpp: Add using directives for overloaded size utils.
ec47928 mkvmuxerutil: Revert to using mkvmuxertypes.
a1dc4f2 Fix parsing of VP9 level.
4e3d037 Add support to output Colour elements to webm_info.
d3656fd muxer_tests: ignore iwyu re gtest-message.h
e76dd5e Fix file name in mkvmuxertypes shim.
1be5889 Add temporary include shims at old file locations.
039df94 Add TEST_TMPDIR environment variable
Change-Id: I84bc1401b0aad71ad6727b687f1bede9953a7a08
While we are at it:
- Rename some variables to more meaningful names
- Reuse some common consts from a header instead of redefining them.
Change-Id: I75c4248cb75aa54c52111686f139b096dc119328
(cherry picked from aomedia 09eea21)
Add stronger condition for splitting 64x64, for low noise content.
This reduces dragging artifact near moving head.
Little/no change in metrics on RTC set.
Change-Id: I39b38cfd20f2ece53ff49c2aaf76ba9f82761be1
fc5f88d Fix temp files being left on system.
c04a134 Add support for overriding PixelWidth and PixelHeight.
c0160e0 Add support to explicitly set segment duration.
02bc809 Add support to estimate file duration.
c97e3e7 Add support to output sub-sample encryption information.
26f4344 MakeUID: quiet unused param warning in Android builds
d6af52a Change check to fix compile error.
1720020 webm_parser: Add Mesh value for ProjectionType
78f2c5a webm_parser: Use ./ prefix for includes
da62f65 webm_parser: Remove webm/ prefix from public includes
e15e8f2 webm_parser: Update README build instructions
5023f2b mkvmuxer: Fix Colour::Valid()
cf16204 mkvmuxer_tests: Actually test cue points in the cue point test.
93e9fb3 Validate Colour element values.
8036925 mkvparser_tests: Add Projection element test.
f52d38c mkvparser_tests: Add Colour element test.
826436a mkvparser: minor SeekHead::Entry clean up.
24fb44a mkvmuxer_tests: Add Projection element test.
1e0a8ea mkvmuxer_tests: Add Colour element test.
0278616 mkvmuxer: Colour accessors/mutators.
2346f8f Add mkvparser wrapper functions.
54d6b6b webm_info: Add Projection element support.
65fee06 mkvmuxer_sample: Add support for Projection element.
9a3f2b5 mkvparser_sample: Add support for Projection element.
41e814a mkvparser: Add Projection element support.
483a0ff mkvmuxer: Add Projection element support.
676a713 Add support for the Projection element
725f362 mkvmuxer: Fix memory leak when Colour is set multiple times.
fa182de mkvparser_sample: Add output of audio track codec private size.
8f521f2 mkvparser_tests: Add invalid BlockGroup test.
39137d7 Remove docs saying binary elements default to 0
c147504 Fix legacy Makefile.
80685d3 Do not skip over unknown elements at the root level
58711e8 mkvparser_sample: Fix version info string.
837746f mkvparser_tests: Add invalid block test.
207cd80 Disambiguate sample sources and targets.
a112d71 mkvparser_tests: Refactor invalid file loading code.
5dea33e Disambiguate test source and target names.
125049e parser_tests: Add another truncated chapter string test.
1de8d4c parser_tests: Add truncated chapter string test.
ff8c2b6 parser_tests: Move cue validation to test_util.
4b0690f parser_tests: Add invalid lacing test.
9828e39 mkvmuxer: Set default doc type version to 4.
5495a59 webm_parser: Reference more files in CMakeLists.txt.
0c0ecd0 vpxpes_parser: Add start code emulation prevention support.
639a4bc webm2pes: Remove debug printfs().
9a51102 webm2pes: fflush() in the correct conversion function.
dc7f155 webm2pes: Track total bytes written.
d518128 webm_parser: Enable usage of werror.
e1fe762 webm2pes: Add test for mux/demux of large input.
1b24a79 vpxpes_parser: Read and store PTS when present.
6cf0a0f vpxpes_parser: Store frame payloads.
25d2602 webm_parser: Convert style to match the rest of libwebm
24be76d webm2pes: Replace VpxFrame with VideoFrame.
b451c3b Add a basic video frame storage class.
05c90eb libwebm_util: Clarify error text in superframe parser.
e6415af webm2pes: Make WritePesPacket() a public method.
8f840dd webm2pes: Move frame read out of PES packet write method.
448af97 webm2pes: Restore frame fragmentation support.
f8bb714 cmake: Integrate new parsing API and tests.
cb8ce0b Add a new incremental parsing API
900d322 vpxpes_parser/webm2pes: BCMV and PTS fixes.
4b73545 webm2pes: Add start code emulation prevention.
82903f3 Add column tiles and frame parallel to webm_info
5d91edf style_clean_up: Remove unnecessary parentheses
a95aa4b vp9_level_stats: correct total_uncompressed_bits_ calculation
f46566f mkvreader: Fix shorten-64-to-32 warning in 32 bit builds.
76630ca mkvwriter: Fix shorten-64-to-32 warning in 32 bit builds.
a8ffbd4 webm2pes: Fix format specifier warnings.
faf89d4 Add MaxLumaSampleRate grace percent to stats.
d31e6c9 Fix profile 2 in vp9_header_parser.
bd3ab3a Add flag to estimate last frame's duration to stats.
c182ed9 Fix lint issue in hdr_util.h
cc62ecd Add test for Cluster memory leak
196708a Change MaxLumaSampleRate to be based on frame resolution.
cbd676b mkvmuxer: Fix leak when a Cluster isn't finalized
47f2843 Add parsing support for new features in CodecPrivate.
9a235e0 mkvmuxer: Set doctype to matroska when muxing non-WebM codecs.
e3c9576 Add VP9 level output to webm_info.
bbaaf2d Add class to gather VP9 level stats.
5cf549f cmake: Log compiler flag at check time.
8bb68c2 Add file to parse data from VP9 frames.
df3412f Add support for setting VP9 profile and level to sample_muxer.
296429a Add support to parse VP9 profile.
87832d4 mkvmuxer: Fix Segment::Finalize in kLive mode
6df3e56 mkvmuxerutil.hpp: Add using directives for overloaded size utils.
ec47928 mkvmuxerutil: Revert to using mkvmuxertypes.
4e3d037 Add support to output Colour elements to webm_info.
a1dc4f2 Fix parsing of VP9 level.
039df94 Add TEST_TMPDIR environment variable
d3656fd muxer_tests: ignore iwyu re gtest-message.h
e76dd5e Fix file name in mkvmuxertypes shim.
1be5889 Add temporary include shims at old file locations.
Change-Id: I6a1026814560be80d604a5ecb9b66406a1186dd9
Re-use the tile worker threads to pack the bitstream in parallel
on a per-tile basis. Restricting this to real-time only for now
(further testing is needed to ensure this does not make 2-pass
worse in any case).
BUG=webm:1309
Change-Id: Ia2c982da56697756e12f02643f589189b3271d98
Fix unit_tests_ubsan failure for VP8/DatarateTestLarge.DenoiserOffOn.
Failure was triggered by commit: df66f8e8.
Change-Id: I7cc5bd309e85950cfc5755e01d0eb942d9ca6984
For 1 pass vbr real-time mode:
Allow for the usage of alt-ref frame when non-zero lag-in-frames is used.
Use non-filtered alt-ref, and select usage based on fast scene/content
analysis/detection within the lag of frames.
Positive gains on ytlive set: overall avgPSNR ~3-4%.
Several clips are up between 5-14%, a few clips are neutral/small change.
Current speed decrease is about ~5-10%.
Use the flag USE_ALTREF_FOR_ONE_PASS to enable this feature
(off by default for now).
Change-Id: I802d2bf3d44f9cf01f6d15c76be9c90192314769
Put limit on gf interval based on lag, and allow
for the adjustment on next gf group also on key frame.
Small/neutral change on ytlive metrics.
Change only affects 1 pass vbr real-time mode.
Change-Id: I339c8f4398848698b6e10fe9482c52ca661b94a5
Updated code to process in 8bit as saturation/clipping takes care of
overflow
Removed unused macro
Change-Id: I113df60286fb28b216df800d95b2d3695ef71440
In 1 pass CBR, with error_resilience off, allow for
special logic to change the default gf behaviour.
In this CL: boost is turned off and the gf period
is set to a multiple of cyclic refresh period.
Change only affect 1 pass CBR mode, i.e, when the flag
gf_update_onepass_cbr is set.
Including the previous change (3ec8e11: to allow cyclic refresh
for error_resilience off), comparing metrics on RTC set for
error_resilience off vs on: avgPSNR/SSIM up by ~6%.
Change-Id: Id5b3fb62a4f04de5a805bd1b418f2b349574e0bc
Due to change in command line to sample encoder from:
7eff8f3 Update to vpx_temporal_svc_encoder command line.
This caused the tests in vpx_temporal_svc_encoder.sh to fail.
Change-Id: Ic667da81955ad117d04610af21877fed1d4f188f
these are compatible as they only load one element of the input so the
larger size of tran_low_t makes no difference in little endian builds.
note the asm is incompatible with big-endian, but there are other points of
failure there so currently it's considered unsupported.
BUG=webm:1294
Change-Id: Icd2665a0699bccae92d1bea43a95b0a83fb17028
It only handles the realloc constraint (preserving low elements) by
serendipity, and we don't actually rely on that behavior anyway.
Meanwhile the calls may do extra copying that gets immediately clobbered
by the callers.
Change-Id: I8dfa89e4a81084b084889c27bd272fdf85184e8d
vp8_short_inv_walsh4x4_msa - Optimized to process in short vector type
Updated below functions to store exact number of bytes in output rather than complete vector
idct4x4_addblk_msa
idct4x4_addconst_msa
dequant_idct4x4_addblk_msa
dequant_idct4x4_addblk_2x_msa
dequant_idct_addconst_2x_msa
Change-Id: Ic1b3752e2421dc7d70a082dcdaab9d140d7e5d9c
cyclic_refresh was tied to error_resilience mode.
Allow it to be on also for 1 pass CBR mode even if
error_resilience is off.
Other option to use new control for this, but prefer to avoid
that for now.
Change-Id: I3625b292ee059a890e31338b514e211bf0ab5c3e
This change will make the highbd txfm input range check more comprehensive
The 25-bit highbd input range is composed by
12 signal input bits + 7 bits for 2D forward transform amplification + 5 bits for
1D inverse transform amplification + 1 bit for contingency in rounding and quantizing
BUG=https://bugs.chromium.org/p/webm/issues/detail?id=1286
BUG=https://bugs.chromium.org/p/chromium/issues/detail?id=651625
Change-Id: I04c0796edd7653f8d463fba5dc418132986131e7
These caused the following warning with GCC 5:
warning: logical not is only applied to the left hand side of
comparison [-Wlogical-not-parentheses]
assert(!is_compound == (cm->reference_mode == SINGLE_REFERENCE));
Change-Id: If296aabb2311ceb7d903b395c1549ef81c2cbf9b
(cherry picked from commit c6cf7a6111)
these functions are incompatible currently and unreferenced in rtcd,
exclude them from the build.
BUG=webm:1294
Change-Id: I7790c195a91e1b142f56c04d2a5e305d9133b896
Rename vpx_lpf_horizontal_edge_8() to vpx_lpf_horizontal_16().
Rename vpx_lpf_horizontal_edge_16() to vpx_lpf_horizontal_16_dual().
Change-Id: I798ca8fbbd657d06d3db2bfb0fb3321168f49e52
This patch sets the 16x16 src_diff to zero and ensures correct calculation
of this_error for block sizes smaller than 16x16.
Change-Id: I7b7c02d267433c9f22c8ac9b8d5df2f499175172
change_config() may be called often in real-time application,
to update bitrate/framerate or qp-max/min.
No need to do update_frame_size() unless frame size has changed.
Change-Id: I23a51deade1e03adc91c468f9ffde3235298770c
For real-time mode at speed 8: turn off MINIMAL_LF at speed 8,
for non-screen content mode.
Visually better, avgPSNR/SSIM on rtc set go up by ~4-5%.
Speed decrease of about ~3%.
Change-Id: I8eb69330f02e0ceece1507d43cfc8a049a1d8291
to get_binary_prob(). the only other caller mode_mv_merge_probs() does
its own test on 0.
BUG=chromium:639712
Change-Id: I1178688706baeca2883f7aadbc254abb219a44ce
+ inline the function directly as there was only one consumer
(get_prob())
this is an attempt to reduce the amount of branches to workaround an amd
bug. this change is mildly faster or neutral across x86-64, arm.
http://support.amd.com/TechDocs/44739_12h_Rev_Gd.pdf
665 Integer Divide Instruction May Cause Unpredictable Behavior
BUG=chromium:639712
Suggested-by: Pascal Massimino <pascal.massimino@gmail.com>
Change-Id: Ia91823aded79aab469dd68095d44300e8df04ed2
Inline function called from test/dct16x16_test.cc wouldn't build due to:
invalid operands of types ‘__gnu_cxx::__enable_if<true, double>::__type
{aka double}’ and ‘int’ to binary ‘operator>>’
return (abs(ref->row) >> 3) < COMPANDED_MVREF_THRESH &&
this converts the test to abs() < COMPANDED_MVREF_THRESH << 3 which
hides the promotion issue.
Regression from commit de993a847f
BUG=webm:1291
Change-Id: I73b5943d07d5b61b709d299114216a2371a8fd62
provides msvc-like warnings for implicit conversions from 64-bit to
32-bit types
--enable-vp9-highbitdepth still requires some work
this also skips CXXFLAGS for now as some work would be needed to cleanup
third_party/*.cc or split it from test/*.cc where it comes to flags.
Change-Id: Ic9a095b73286eba5ed39bfc27ff69593748cbbf4
The original commit never set any 'specialize' line:
61311e6103
It appears the sadx4 version of function uses sdx4df calls to speed up
the search. There are no sse3 versions of the sdx4df functions, but
there are sse2 and msa versions.
There is a neon version of vpx_sad16x16x4d but not any of the smaller
versions. Perhaps if they existed this function could be expanded to use
them.
Change-Id: I936d7d6b1a3ff6dcd5a4d2322272708c47cdec13
* changes:
Expand -Wextra to more of the library
mips: clean up wextra warnings
Add compiler flag -Wsign-compare
Add compiler warning flag -Wextra and fix related warnings.
Remove unused zbin variable:
warning: unused parameter ‘zbin’
Use int for loop variables to avoid unsigned conversion:
warning: comparison between signed and unsigned integer expressions
Change-Id: Icea74b870c0ee68a8bf687e796a69392af25a8ad
Also, fix the warnings generated by this flag.
(cherry picked from commit ebeb1155d4fa6d28e2f40c92265245f8df097fcb)
From AOM. Don't actually add -Wsign-compare. It will be covered by
-Wextra.
Switch to vpx_integer.h from df9c9d6d4c43f02c58d4e776c53323788e013cbc
BUG=webm:1069
Change-Id: I1dc6e61caa5d56af4a55b6692ab620bb3144652a
Note: some of these warnings are enabled by a combination of -Wunused
(added earlier) and -Wextra.
Cherry-picked from AOM 4790a69faaec8f03d65f64ff070f6ab4307dbb16
Expands use of (void)x; on unused variables. AOM only supports one codec
in codec_factory.h
Does not include changes to HandleDecodeResult. AOM removed
invalid_file_test.cc which does use the video parameter.
Does not enable -Wextra yet. There are more issues to fix.
BUG=webm:1069
Change-Id: I322a1366bd4fd6c0dec9e758c2d5e88e003b1cbf
Current version does not build with options:
--enable-vp9-highbitdepth --enable-coefficient-range-checking
Change-Id: Ic3285f1a3e0d6be88da7f2cd8fa5a631368dd03b
Reduce the filt_guess for 1 pass cbr on inter-frames.
This reduces visual artifact seen in rtc clip (jimred.vga),
and improves metrics on rtc set.
Metrics on rtc set for cbr mode overall positive, most clips are up:
Speed 7 rtc: avgPSNR/SSIM up by: ~2.6/3.9%
Speed 8 rtc: avgPSNR/SSIM up by: ~1.3/2.5%
Change-Id: Ia4eccea1c19d65b583516df28823cd756c49464f
Added a cap on the maximum boost for an arf based on interval length.
Fixed bug where by the image size was not accounted for in determining
two of the motion breakout thresholds.
Overall small gains of 0.2-0.4% psnr but on large image format clips with
slow zooms the gain may be as much as 20% or more (e.g. in_to_tree
at 1080P)
Change-Id: Id0a47391203026742daa9c97afac5705fd8c4dfb
The value 35468 changes sign when stored in int16_t:
implicit conversion from 'int' to 'int16_t' (aka 'short')
changes value from 35468 to -30068
This negation requires adding back the original value to compensate.
Shifting the value keeps the value positive and saves a post-vqdmulh
shift.
This technique is used in webp and idct_dequant_full_2x_neon
BUG=b/28027557
Change-Id: I0c5ce09bea170fe08061856c2af6f841a557e0c3
This restores d9dce2f48e
Switched to using signed shift-and-narrow. Instead of saturating
negative results to 0, it was saturating them to 255.
BUG=webm:817
BUG=webm:1273
Change-Id: I571095336aa4182e3288b17924fcaaece42b0a49
removes some unnecessary casts and adds a few explicit uint32 ones for
larger sizes to quiet -Wshorten-64-to-32 warnings
Change-Id: I63c5fce8e62c426d5cf5c10a66a113c119a43518
Do nothing in vp9_highbd_iht#x#_##_add_c when input magnitude is beyond
20 bits. Note that, sign bit is not included here.
In the 20 bits, we use 12 bits for input signal, 7 bits for forward
transform amplification, and 1 bit for contingency in rounding and
quantizing
BUG=https://bugs.chromium.org/p/webm/issues/detail?id=1286
Change-Id: I332c6f68df4614fc2e7d2dc4c5bb0d0cff8a245c
When filtering it needs 6 pixels: 2 prior to the source, the source, and
3 after the source.
When filtering 16 wide, that means 21. To accomplish this the SSE2 reads
[-2] to [5], [6] to [13], and [14] to [21], a total of 24 bytes (reading
in groups of 8 is easy)
The filter then shifts this last set to the top half of the register and
uses 'or' to combine it with the previous set.
Valgrind detected an issue reading pixels [19], [20] and [21]:
Address 0x7f581c2 is 434 bytes inside a block of size 441 alloc'd
Note: we only need pixels [16], [17], and [18] as context for [15].
To fix this, it now reads 8 bytes starting at [11], which re-loads [11]
through [13], but stops at [18] and does not over-read any values.
This is shifted by 5 and 'or'd with xmm1. Although the lower bits are
not cleared, they overlap directly with [11] through [13], so 'or'
produces the correct results.
Change-Id: I0c89c03afa660fc9b0108ac055d7bd403e493320
On 32 bit machines 'new' does not always appear to allocate sufficiently
aligned buffers, causing intermittent test failures.
Change-Id: I0db4fc73782012e4eef71dc0fb540e74fdbfcebe
the --enable-postproc-visualizer configure option remains as a no-op as
do the control names and values for compatibility
+ remove the corresponding debug flags from vpxdec: --pp-*
Change-Id: I4a001cd9962b59560d7d6bda6272d4ff32b8d37c
similar to changes that were done in vp9 for encoded frame size
reporting. has the side-effect of quieting a -Wshorten-64-to-32 warning.
Change-Id: I89f74cb617fc29334ee351dc8dfaa3b8cfd4e5af
The referenced bug was fixed by saving neon registers. That this had any
effect was coincidental.
Both chromium and Android build with clang and neither uses this flag.
Change-Id: I470247d6fd9226fc207b42a187105581a94badc3
The vp9_mv_class0_tree is a balanced tree with two leafs and can
simply be coded as a boolean with probability class0[0].
Change-Id: If294dac825a5f945371092c74aa8e3f84cd962b6
(cherry picked from commit be8a8ab62ebdd111c6f2e9a33b15630570671eba)
assume __clang_major__==0 has the latest version of
_mm256_broadcastsi128_si256. fixes builds with custom clang toolchains.
BUG=b/30970831
Change-Id: I90becd56278e4716bd46e2ba9d910af977e8dfa6
The code only has issues when xoffset == 0 and yoffset == 0 which
represents a simple copy. Presumably this case does not need to be
handled because the issue has existed since 2010.
BUG=webm:1287
Change-Id: Ic47e2653f3b729e99b40e53d8d2d8d1501edaaa9
Build out the sixtap_predict test because the filters are
interchangeable. Add verbose failures and border checking.
Change-Id: I962f50041750dca6f8d0cd35a943424cf82ddcb1
This reverts commit d9dce2f48e.
Appears to be failing the SixtapPredict tests in some configurations and possibly test vectors as well.
Change-Id: Ica6aa83ebac47d0a76e451846e7da67b1c17a7d7
This function was removed when clang started introducing alignment hints
which caused the 32 bit vld1_lane_u32/vst1_lane_u32 to fail:
https://llvm.org/bugs/show_bug.cgi?id=24421
The load has been rendered safe with an implementation ~indiscernible
performance-wise that uses _u8 and over-reads just a touch.
It is still ~5x faster than C in the unaligned case and doing both
filters.
BUG=webm:892
BUG=webm:1273
Change-Id: Icf7167189391b46202f47233bb585c24c42bcc36
postproc.c is overloaded and used for both postproc and internal stats.
If only --enable-internal-stats is specified there are issues with
non-existent struct members and unused functions.
Change-Id: I82367f1ffce659c3918c9f964dbce94a716fbb89
All the other test which do not use 'pass' (which appears to be almost
all of them) do this.
Cleans -Wextra/-Wunused-parameter:
unused parameter ‘pass’
Change-Id: I1ff3acf3f3d1e831f94dcb00ea36337afe0aefe0
Remove the experiment LIMIT_QP_ONEPASS_VBR_LAG, as its
not currently used and no plan to use in near future.
Change-Id: Ib069f8d7225195be04b765d0ab477510dfba6a3b
This function was removed when clang started introducing alignment hints
which caused the 32 bit vld1_lane_u32/vst1_lane_u32 to fail:
https://llvm.org/bugs/show_bug.cgi?id=24421
The load has been rendered safe with an implementation ~indiscernible
performance-wise that uses _u8 and over-reads just a touch.
The store, when unaligned, has a version that is ~25% slower but safe
when xoffset = 0 (second pass filter only). When the first pass filter
(or both) are in play, the new version is almost identical in speed.
Worst case performance (both filters, unaligned stores) is roughly 3-4x
faster than C.
BUG=webm:817
BUG=webm:1273
Change-Id: I1e490e94453e0872151fe0dafb05557463f6247d
Use the canonical 'vpx_codec_dec_cfg_t()' as opposed to 'vp9_zero()'
which just hammered everything to 0.
Change-Id: Id820efef700ad92a625797f8fd58e465b15eeca4
For some reason allocated_decoding_thread_count is signed, but decoding_thread_count is not.
Cleans -Wextra/-Wsign-compare:
comparison between signed and unsigned integer expressions
Change-Id: Id0ada78100acff27c1c4ed7493c563d13c55cdcd
Use vp9_zero() to set every element.
Cleans -Wextra/-Wmissing-field-initializers:
missing initializer for member ‘vpx_codec_dec_cfg::w’
missing initializer for member ‘vpx_codec_dec_cfg::h’
Change-Id: I5b41ce7d55a912e29b1d4c3e840cea80e8510fbe
This change was reverted before due to a hangouts encode-time
regression investigation. But since then this change has been
cleared of causing any noticeable regression.
This mode reduces some false detection, and uses the
same model as in vp9.
Change-Id: I9c82a748c5f601d0aca9f61ee218abfbd58c62bd
* changes:
Enable -Wundef by default
Define VP8_TEMPORAL_ALT_REF to !CONFIG_REALTIME_ONLY
Remove CONFIG_DEBUG guards from assert()
Remove unused function vpx_de_mblock
Fix -Wundef warning for OUTPUT_FPF
Fix -Wundef warning for __SANITIZE_ADDRESS__
Added casts to remove warnings:
BUG=webm:1274
In regards to the safety of these casts they are of two types:-
- Normalized bits per (16x16) MB stored in a 32 bit int (This is safe as bits
per MB even with << 9 normalization cant overflow 32 bits. Even raw 12
bits hdr source even would only be 29 bits :- (4+4+12+9) and the encoder
imposes much stricter limits than this on max bit rate.
- Cast as part of variance calculations. There is an internal cast up to 64 bit
for the Sum X Sum calculation, but after normalization dividing by the number
of points the result will always be <= the SSE value.
Change-Id: I4e700236ed83d6b2b1955e92e84c3b1978b9eaa0
Previously VP8_TEMPORAL_ALT_REF was only defined for non-realtime-only
builds. However, its value was checked with #if, not #ifdef.
Fixes -Wundef warnings.
BUG=webm:1069
Change-Id: If78d8731298f3f0d3662ffa25f973e7adaf67152
Using a tighter resize constraint on undershoot seems to help
results (especially SSIM) as significant undershoot on a frame
seems to have more of a damaging impact than overshoot.
This patch has been tuned so that in local testing using the
derf set it is encode speed neutral for speed setting 2.
Average quality result for speed 2 (psnr,ssim) were as follows:-
lowres 0.039, 0.453
midres 0.249, 0.853
hdres 0.159, 0.659
NetFlix -0.241, 0.360
Change-Id: Ie8d3a0d7d6f7ea89d9965d1821be17f8bda85062
Fixes windows build issue:
==> tests::VS10_x64 is broken
LINK : warning C4742: 'kYvuI601Constants' has different alignment in 'third_party\libyuv\source\row_common.cc' and 'third_party\libyuv\source\planar_functions.cc': 32 and 2 [.build-x86_64-win64-vs10\vpxdec.vcxproj]
LINK : warning C4744: 'kYvuI601Constants' has different type in 'third_party\libyuv\source\row_common.cc' and 'third_party\libyuv\source\planar_functions.cc': '__declspec(align(32)) struct (224 bytes)' and 'struct (224 bytes)' [.build-x86_64-win64-vs10\vpxdec.vcxproj]
LINK : warning C4742: 'kYuvI601Constants' has different alignment in 'third_party\libyuv\source\row_common.cc' and 'third_party\libyuv\source\planar_functions.cc': 32 and 2 [.build-x86_64-win64-vs10\vpxdec.vcxproj]
LINK : warning C4744: 'kYuvI601Constants' has different type in 'third_party\libyuv\source\row_common.cc' and 'third_party\libyuv\source\planar_functions.cc': '__declspec(align(32)) struct (224 bytes)' and 'struct (224 bytes)' [.build-x86_64-win64-vs10\vpxdec.vcxproj]
LINK : warning C4742: 'kYvuI601Constants' has different alignment in 'third_party\libyuv\source\row_common.cc' and 'third_party\libyuv\source\planar_functions.cc': 32 and 2 [.build-x86_64-win64-vs10\vpxenc.vcxproj]
LINK : warning C4744: 'kYvuI601Constants' has different type in 'third_party\libyuv\source\row_common.cc' and 'third_party\libyuv\source\planar_functions.cc': '__declspec(align(32)) struct (224 bytes)' and 'struct (224 bytes)' [.build-x86_64-win64-vs10\vpxenc.vcxproj]
LINK : warning C4742: 'kYuvI601Constants' has different alignment in 'third_party\libyuv\source\row_common.cc' and 'third_party\libyuv\source\planar_functions.cc': 32 and 2 [.build-x86_64-win64-vs10\vpxenc.vcxproj]
LINK : warning C4744: 'kYuvI601Constants' has different type in 'third_party\libyuv\source\row_common.cc' and 'third_party\libyuv\source\planar_functions.cc': '__declspec(align(32)) struct (224 bytes)' and 'struct (224 bytes)' [.build-x86_64-win64-vs10\vpxenc.vcxproj]
LINK : error C2220: warning treated as error - no 'executable' file generated [.build-x86_64-win64-vs10\vpxdec.vcxproj]
LINK : error C2220: warning treated as error - no 'executable' file generated [.build-x86_64-win64-vs10\vpxenc.vcxproj]
Change-Id: Ic3c4fff9209f5a52ff8f8ff321548d49ba09ec06
removes the need for an intermediate cast to int, which was missing in
the call added in:
69c5ba1 vpx_mem: Refactor code
quiets a visual studio warning:
C4146: unary minus operator applied to unsigned type, result still
unsigned
Change-Id: I76c4003416759c6c76b78f74de7c0d2ba5071216
This patch is to address concerns that changes to allow
recodes on the first frame in each ARF group do not give a
good enough speed quality trade off for speed 2. Though the
average impact on encode speed is 1-2%, for some hard clips
it is > 5% rise. For speed 1 this is less an issue and for Speed 0
the previous patch actually improves speed.
Change-Id: Ie1bcefdbfdf846d3f4428590173f621465dffe3a
This corrects a formatting error introduced in:
I1e9d548ce445d29002f0c59ebfd3957a6f15e702
where spaces were used as delimiters instead of tabs.
The corresponding fix for vp10 is in
Ica3d625d6672b3c47e0e208b45eede29b9004030.
Change-Id: Ibc4eb8fd82e6b926ba259a679dc98557cadba9b1
Current commit is just an API template for the rest of the code, and
I will add inner logic later.
Altref frames generate a lot of bitrate and at the same time
other frames refer to them a lot, so it makes sense to apply
special compensation-based adaptive quantization scheme for altref
frames. E.g., for blocks that are good predictors for the future
apply rate-control chosen quantizer while for bad predictors apply
worse one.
Change-Id: Iba3f8ec349470673b7249f6a125f6859336a47c8
Previously Tx domain rd was used in all cases above speed 0.
Coefficient optimization was only enabled for best and speed 0.
This patch selectively sets these features at other speed settings
based on block complexity.
For the Netflix and HD sets in particular the quality gains are
large compared to the speed hit. At speed 1 the average psnr
gain in the NF set is > 2.5% with one clip coming in at 18%
and some points almost 30%. Average gains for the lower
resolution test sets are around 1%.
The gains are biggest at low Q so some further optimization
may be possible.
Change-Id: I340376c7b2a78e5389a34b7ebdc41072808d0576
fixes SSE2/AddNoiseTest.CheckCvsAssembly/0 with -funsigned-char.
visibly broken since:
0dc69c7 postproc : fix function parameters for noise functions.
where the types diverged (char vs. int8)
but likely the return changed in:
2ca24b0 postproc - move filling of noise buffer to vpx_dsp.
when multiple implementations were merged.
Change-Id: I176ca1f170217f05ba7872b0c4de63e41949e999
set a max allocable size to prevent overflows in 32-bit and extremely
large allocation attempts in 64-bit. this could be amended to allow size
or num parameters to be 64-bits with the correct size being used at each
call site.
BUG=webm:819
Change-Id: Ia81004d6c4279680714c4488b4f6cf287ab396a5
vpx_realloc was allocating 1 byte more than needed every time.
Fixed this, and took this opportunity to do a small refactoring.
Change-Id: I38fcb62b698894acbbab43466c1decd12f906789
(cherry picked from aom: 2a876b4 aom_realloc correction.)
In the future this option will activate adaptive quantization special
for altref frames. Encoder will create the adaptive quantization map
on the basis of lookahead buffers similarity which is the estimate of
the future motion compensation performance.
Change-Id: Ia0088b3babb0f9a4899c79d8d819947ba5a03df2
This function only exists as a shortcut to subpixel variance with
predefined offsets. xoffset = 4 for horizontal, yoffset = 4 for vertical
and both for "hv"
Removing this allows the existing optimizations for the variance
functions to be called. Instead of having only sse2 optimizations, this
gives sse2, ssse3, msa and neon.
BUG=webm:1273
Change-Id: Ieb407b423b91b87d33c4263c6a1ad5e673b0efd6
decoding the same invalid keyframe twice would result in a crash as the
second time through the decoder would be assumed to have been
initialized as there was no resolution change. in this case the
resolution was itself invalid (0x6), but vp8_peek_si() was only failing
in the case of 0x0.
invalid-vp80-00-comprehensive-018.ivf.2kf_0x6.ivf tests this case by
duplicating the first keyframe and additionally adds a valid one to
ensure decoding can resume without error.
BUG=b/30593765
Change-Id: If0859035908b7870d67a7f3f646b5a080252eb6d
Disabled the split mode while encoding 4k video to speed
up the encoder.
Borg test result on 4k set:
Overall PSNR: +0.029%; SSIM: +0.009%.
Average encoder speedup at speed 2 is 2.5%.
Change-Id: I1519c658f07c3ac838affbe5aff0ed9b94f3f8f4
Adjusted speed 2 features to speed up 4k video encoding.
BDBR results from borg test:
PSNR: +0.313%; SSIM: +0.268%.
Average speedup: 8.5%
Change-Id: I1e2695a01fb3f3817c1df4480e184c2aed8f2eba
this fixes a crash in vp9_dec_setup_mi() via
vp9_init_context_buffers() should decoding continue and the decoder
resyncs on a smaller frame
BUG=b/30593752
Change-Id: I9ce8d94abe89bcd058697e8bd8599690e61bd380
Bias towards base_mv and skip 1/4 pixel motion search when using base mv.
2~3% speed up for 2 spatial layers, 3~5% speed up for 3 spatial layers.
PSNR loss:
(2 layers) 0.07dB for gips_stationary, 0.04dB for gips_motion;
(3 layers) 0.07dB for gips_stationary, 0.06dB for gips_motion.
Change-Id: I773acbda080c301cabe8cd259f842bcc5b8bc999
Add option, for newmv-last, to limit the rd-threshold update for early exit,
under a source varianace condition.
This can improve visual quality in low texture moving areas,
like forehead/faces.
Also add bias against golden to improve the speed/fps,
will little/negligible loss in quality.
Only affects CBR mode, non-svc, non-screen-content.
Change-Id: I3a5229eee860c71499a6fd464c450b167b07534d
The flag was added because Apple clang and Chromium clang disagreed
for certain versions of instructions.
qsubaddx, qaddsubx, ldrneb and ldrneh were used in armv6 assembly
which was removed in d55724fae9
vqshrun was used in some neon assembly but superseded by
dcbfacbb98
.include was used for obj_int_extract/asm_offsets and removed in
6eec73a747
Change-Id: I32f4c9b536d0318482101c0b8e91e42b8f545f18
Changes the default recode rule for Speed 0 and best quality
from ALLOW_RECODE to ALLOW_RECODE_KFARFGF.
Tested on the NF, hdres, midres and lowres test sets, this setting
when combined with patch I40cb559... now performs "as well" in
metrics terms (in fact it came out a tiny amount better overall)
but encode time is 9.6% faster (measured as the average
from 27 mid rate local encodes on clips in the derf/lowres set.
Change-Id: I8c781c0cdfa3a9929cd9406d15582fce47d6ae3b
Allow recodes for the first inter frame in each arf group
even when the recode rule is set to ALLOW_RECODE_KFARFGF.
Small gains of 0.05%.
Change-Id: I40cb559d36a2bf0ebf5cf758c3f92e452b480577
disable clang-format for bilinear_filters_avx2
restores the row layout prior to:
099bd7f vpx_dsp: apply clang-format
but keeps the justification used by clang-format
Change-Id: Icf1733a37edb807e74c26b23a93963c03bd08fd7
This patch fixed a motion vector out of range bug:
vpxenc: ../libvpx/vp9/encoder/vp9_mcomp.c:69:
mv_cost: Assertion `mv->col >= -((1 << (11 + 1 + 2)) - 1) &&
mv->col < ((1 << (11 + 1 + 2)) - 1)' failed.
For blocks that returned without having full-pixel search, the original
MV limits were not restored, which caused the failure. Moved the set
MV limit function down to fix the bug.
Change-Id: Id7d798fc7214e95c6e4846c588f0233fcf1a4223
_beginthreadex does not align the stack on 16-byte boundary as expected
by gcc.
On x86 targets, the force_align_arg_pointer attribute may be applied to
individual function definitions, generating an alternate prologue and
epilogue that realigns the run-time stack if necessary. This supports
mixing legacy codes that run with a 4-byte aligned stack with modern
codes that keep a 16-byte stack for SSE compatibility.
https://gcc.gnu.org/onlinedocs/gcc/x86-Function-Attributes.html
Change-Id: Ie4e4ab32948c238fa87054d5664189972ca6708e
Signed-off-by: Aleksey Vasenev <margtu-fivt@ya.ru>
prevents use of an uninitialized value in the deconstructor should the
test fail before tmpfile_ is set.
Change-Id: I8b49fd05f0d05e055fdf653bd46983d30f466a68
applied against a x86_64 configure with and without
--enable-vp9-highbitdepth
clang-tidy-3.7.1 \
-checks='-*,google-readability-braces-around-statements' \
-header-filter='.*' -fix
+ clang-format afterward
Change-Id: Ia2993ec64cf1eb3505d3bfb39068d9e44cfbce8d
Extract the duplicated data generation code in OperationCheck() of
Loop8Test6Param and Loop8Test9Param, and put in function InitInput().
Change-Id: Ied39ba4ee86b50501cc5d10ebf54f5333c4708f0
This patch fixed a motion vector(MV) out of range bug, which was caused
by not restoring the original values of the MV min/max thresholds after
the sub8x8 full pixel motion search. It occurred rarely and only was seen
while encoding a 4k clip for 200 frames.
BUG=webm:1271
Change-Id: Ibc4e0de80846f297431923cef8a0c80fe8dcc6a5
* changes:
Use common transpose for vpx_idct32x32_1024_add_neon
Use common transpose for vpx_idct8x8_[12|64]_add_neon
Use common transpose for vp9_iht8x8_add_neon
Use common transpose for vpx_idct16x16_[10|256]_add_neon
The code was expanding to Q registers so that vqrshn could be used, for
vector quad round shift and narrow. If 4 values are added together,
there is a shift by 2. If 8 values, a shift by 3. Since this accounts
for any possibility of overflow, we can skip the narrowing shift.
This allows keeping the values in D registers and casting the 16 bit
value to 8 bits.
Change-Id: I8d9cfa07176271f492c116ffa6a7b351af0b8751
The neon intrinsics are not able to load just the 4 values that are
used. In vpx_dsp/arm/intrapred_neon.c:dc_4x4 it loads 8 values for both
the 'above' and 'left' computations, but only uses the sum of the first
4 values.
BUG=webm:1268
Change-Id: I937113d7e3a21e25bebde3593de0446bf6b0115a
Increase the minimum distance.
Reduces the overshoot somewhat on some clips,
small gain in avgPSNR (~0.1%) on ytlive set.
Change-Id: Id5ddde20c2907dbdb536e79542eff775019c142b
This error was introduced by the patch:
8ce67d7 vp9 svc: Enable different speed setting for each spatial layer.
To use svc, svc_param should be cleared to 0 at the beginning.
Change-Id: I222f03ddae8a50e84b4690b78263abb742fae91e
Move best index into the token state. Shrink it down to one byte. This
is more cache friendly (access are group together) and uses less total
memory.
Results in 4% fewer cycles in optimize_b().
Change-Id: I75db484fb3dc82f59928d54b659d79c80ee40452
everything outside of third_party should follow 'PointerAlignment:
right' i.e., associate the '*' with the variable
+ add a note about the clang-format that generated this file
Change-Id: I13e3f4f5fb6e22a8fa7fc3d06879c995b7c41a39
- make Check() void as the EXPECT's are sufficient to document failure
cumulatively this has the effect of avoiding reporting incorrect Check()
failures due to earlier test failures.
Change-Id: I2cf775449f18c90c1506b8eadd7067adbc3ea046
The LLVM trunk has reached 4.0 and now __clang_major__ is not enough
to distinguish between old XCode Clang and the new 'real' Clang.
Using __apple_build_version__ allows to make this distinction.
BUG=chromium:631144
Change-Id: I0b6e46fddfe4f409c7b7e558bda34872e60ee2d9
Shutdown all threads before reclaiming any memory. The frame-level
parallel decoder may access data from another worker.
BUG=webm:1259
Change-Id: I26856ebd1f77cc4a4545331baa19bbf3e01c4ea4
1 - stops de allocating before threads are closed.
2 - limits threads to mb_rows when mb_rows < partitions
BUG=webm:851
Change-Id: I7ead53e80cc0f8c2e4c1c53506eff8431de2a37e
This commit changes the call in vp9 encoder from vp9_deblock() to
vp9_post_proc_frame() to ensure the data structures used in the call
are properly allocated. This fixes an encoder crash when configured
with --enable-internal-stats.
Change-Id: I2393b336c0f566665336df4f1ba91c405eb56764
bitstream.c: asserts are disabled when CONFIG_DEBUG is unset
vp8_dx_iface.c: split |s into 2 statements across #if bounds
Change-Id: I307d1e969134db5c9c0edd7690589b6b29116cbd
allows 'make test_libvpx', etc. some reworking of the makefiles would be
needed to avoid hard coding targets here.
Change-Id: I18982dbf691e7d36ab8bcf5934bab9340687b061
remove some (but not all yet!) tuple mis-use, and revamp the code a lot.
Factorize some common chores into MainTestClass.
Change-Id: Id37b7330eebe80d19b9d12a454f24ff9be6b1116
Allow usage of lookahead for VBR in real-time mode, for 1 pass vbr.
Current usage is for fast checking of future scene cuts/changes,
and adjusting rate control (gf interval and active_worst/target size).
Added unittests (datarate) for 1 pass vbr mode, with non-zero lag.
Added an experimental option to limit QP based on lookahead.
Overall positive gain in metrics on ytlive set:
avgPNSR/SSIM up on average ~1-3%; several clips up by 5, 7%.
Change-Id: I960d57dfc89de121c4824b9a9bf88d2814e74b56
Chromium changed the upstream default to --squash but this conflicts
with libvpx historical defaults.
Change-Id: I80f2f2b48e2ba08e02184b50e6d5f8f5e76fec24
remove some (but not all yet!) tuple mis-use, and revamp the code a lot.
Factorize some common chores into MainTestClass.
Change-Id: Ia14f3924140e8545e4f10d0504475681baae8336
* changes:
vpx_thread: use CreateThread for windows phone
vpx_thread: use WaitForSingleObjectEx if available
vpx_thread: use InitializeCriticalSectionEx if available
vpx_thread: use native windows cond var if available
vpx_thread.[hc]: update webp source reference
This change eliminates redundant computation in the two stage
downscaling, which saves ~1% encoding time in 3-layer svc encoding.
Change-Id: Ib4b218811b68499a740af1f9b7b5a5445e28d671
Modify the gfu_boost and af_ratio setting based on the
average frame motion level.
Change only affects 1 pass vbr.
Metrics overall positive on ytlive set.
On average up by ~1%, several clips up by 2-4%.
Change-Id: Ic18c49eb2df74cb4986b63cdb11be36d86ab5e8d
Use the precise context to estimate the zero token cost in trellis
optimization process. This improves the speed 0 coding performance
by 0.15% for lowres and 0.1% for midres. It improves the speed 1
coding performance by 0.2% for midres and hdres.
Change-Id: I59c7c08702fc79dc4f8534b64ca594da909e2c91
This commit allows the inter prediction residual to use uniform
quantization followed by trellis coefficient optimization in
speed 0. It improves the coding performance by
lowres 0.79%
midres 1.07%
hdres 1.44%
Change-Id: I46ef8cfe042a4ccc7a0055515012cd6cbf5c9619
Move the operations that update the context buffers outside this
function. The coeff_cost() takes all input as const value and returns
the coefficient cost.
This makes preparation for the next coefficient optimization CLs.
Change-Id: I850eec6e5470b91ea84646ff26b9231b09f70a0c
Replace the existing mv bias with a bias only for
NEWMV, and based on the motion vector difference of
its top/left neighbors.
For cbr non-screen-content mode.
Change-Id: I8a8cf56347cfa23e9ffd8ead69eec8746c8f9e09
Use a measure of noise energy to adjust Q estimate and
arf filter strength.
Gains 0.3-0.5% on Lowres and |Netflix sets.
Hdres and Midres neutral.
Change-Id: Ic0de552e7b6763e70eeeaa3651619831b423e151
Use pixel domain distortion metric in speed 0. This improves the
compression performance by 0.3% for both low and high resolution
test sets.
Change-Id: I5b5b7115960de73f0b5e5d0c69db305e490e6f1d
Safer to have the decoder operate normally and have
better-hw-compatibility only implement encoding changes.
Fixes some test failures.
Change-Id: I0dd70d002e4e893992f0cd59774b9363e6f7fe76
For real time CBR mode, use model_rd_for_sb_y for 32x32 if the sb is
a skin sb to avoid visual regression on the slowly moving face.
Refer to the cl: https://chromium-review.googlesource.com/#/c/356020/
Change-Id: I42c36666b2b474ce5ee274239d52ae8ab400fd46
The transform block row and column positions are always available
outside the callees. There is no need to re-compute these values
again. This approach has been used by the decoder. This commit
removes txfrm_block_to_raster_xy() function.
Change-Id: I5b90f91a0d8b7c35cfa7d171da9edf8202630108
BUG=b/29583578
original webp change:
commit d2afe974f9d751de144ef09d31255aea13b442c0
Author: James Zern <jzern@google.com>
Date: Mon Nov 23 20:41:26 2015 -0800
thread: use CreateThread for windows phone
_beginthreadex is unavailable for winrt/uwp
Change-Id: Ie7412a568278ac67f0047f1764e2521193d74d4d
100644 blob 93f7622797f05f6acc1126e8296c481d276e4047 src/utils/thread.c
100644 blob 840831185502d42a3246e4b7ff870121c8064791 src/utils/thread.h
Change-Id: Iade8fff6367b45534986c77ebe61abeb45bce0f8
BUG=b/29583578
original webp change:
commit 0fd0e12bfe83f16ce4f1c038b251ccbc13c62ac2
Author: James Zern <jzern@google.com>
Date: Mon Nov 23 20:40:26 2015 -0800
thread: use WaitForSingleObjectEx if available
Windows XP and up
Change-Id: Ie1a46a82722b8624437c8aba0aa4566a4b0b3f57
100644 blob d58f74e5523dbc985fc531cf5f0833f1e9157cf0 src/utils/thread.c
100644 blob 840831185502d42a3246e4b7ff870121c8064791 src/utils/thread.h
Change-Id: If165c38b378c6e0c55e17a1b071efd3ec3e7dcdd
BUG=b/29583578
original webp change:
commit 63fadc9ffacc77d4617526a50c696d21d558a70b
Author: James Zern <jzern@google.com>
Date: Mon Nov 23 20:38:46 2015 -0800
thread: use InitializeCriticalSectionEx if available
Windows Vista / Server 2008 and up
Change-Id: I32c5b4e5384d614c5a821ef511293ff014c67966
100644 blob f84207d89b3a6bb98bfe8f3fa55cad72dfd061ff src/utils/thread.c
100644 blob 840831185502d42a3246e4b7ff870121c8064791 src/utils/thread.h
Change-Id: I9ce49b3a86857267e504cd8ceab503b7b441d614
BUG=b/29583578
original webp change:
commit 110ad5835ecd66995d0e7f66dca1b90dea595f5a
Author: James Zern <jzern@google.com>
Date: Mon Nov 23 19:49:58 2015 -0800
thread: use native windows cond var if available
Vista / Server 2008 and up. no speed difference observed.
Change-Id: Ice19704777cb679b290dc107a751a0f36dd0c0a9
100644 blob 4fc372b7bc6980a9ed3618c8cce5b67ed7b0f412 src/utils/thread.c
100644 blob 840831185502d42a3246e4b7ff870121c8064791 src/utils/thread.h
Change-Id: Iede7ae8a7184e4b17a4050b33956918fc84e15b5
these are debug-only modules that can be added in manually when needed.
leave a reference in vp8_common.mk / vp9_common.mk for easy addition.
quiets -Wmissing-prototypes warning
BUG=b/29584271
Change-Id: Ifc8637d877edfbd562b34dc5c540428bba7951fc
Moved the API patch from NextGenv2. An example was included.
To try it, for example, run the following command:
$ examples/vpx_cx_set_ref vp9 352 288 in.yuv out.ivf 4 30
Change-Id: I4cf8f23b86d7ebd85ffd2630dcfbd799c0b88101
The scaling of the threshold for 10 and 12 bit here appears
to be in the wrong direction. For 10 and 12 bit we expect sse
values to be higher and hence the threshold used should be
scaled up not down.
Change-Id: I2678116652b539aef48100e0f22873edd4f5a786
This function seems to scale the threshold for testing an
SSE value in the wrong direction for 10 and 12 bit inputs.
Also for a true SSE the scalings should probably be << 4 and 8
Change-Id: Iba8047b3f70d04aa46d9688a824f3d49c1c58e90
CONVERT_TO_BYTEPTR(x) was corrected in:
003a9d2 Port metric computation changes from nextgenv2
to use the more common (x) within the expansion. offsets should occur
after converting the pointer to the desired type.
+ factorized some common expressions
Change-Id: I171c3faaa5606d098e984baa9aa74bb36042f57f
Force enable x86inc.asm when building for x86. Previously there were
compatibility issues so a flag was added to simplify disabling this
code.
The known issues have been resolved and x86inc.asm is the preferred
abstraction layer (over x86_abi_support.asm).
BUG=b:29583530
Change-Id: Ib935e97b37ffb22d7af72ba0f04564ae6280f1fd
inadvertently lost in the final patchset of:
078dff7 configure: remove old visual studio support (<2010)
this prevents an empty CONFIG_VS_VERSION and avoids make failure
Change-Id: I529d52eca59329e2715309efd63d80f0e1fed462
For real time CBR mode, use model_rd_for_sb_y for 32x32 if the mode is
newmv last, which is less aggressive in skipping transform and
quantization, to avoid quality regression in some conditions.
Change-Id: Ifa30be587f2a8a4a7f182a172de6ce277c0f8556
Speed test shows the new vertical filters have degradation on Celeron
Chromebook. Added "X86_SUBPIX_VFILTER_PREFER_SLOW_CELERON" to control
the vertical filters activated code. Now just simply active the code
without degradation on Celeron. Later there should be 2 set of vertical
filters ssse3 functions, and let jump table to choose based on CPU type.
Change-Id: Iba2f1f2fe059a9d142c396d03a6b8d2d3b981e87
Due to rounding, hbd variance may become negative. This commit put in
check and clamp of negative values to 0.
Change-Id: I610d9c8aa2d4eebe7bc5f2c5624a9e3cadad4c94
For forced key frames in particular this helps to make them
blend better with the surrounding frames where noise tends
to be suppressed by a combination of quantization and alt
ref filtering.
Currently disabled by default under and IFDEF flag pending
wider testing.
Change-Id: I971b5cc2b2a4b9e1f11fe06c67ef073f01b25056
Bug fix: The crash is caused by not allocating buffer for prev_mip in
postproc_state and prev_mip in postproc_state is only used for MFQE,
ohter postproc modules, deblocking and etc., should not use it.
BUG=webm:1251
Change-Id: I3120d2f50603b4a2d400e92d583960a513953a28
add a trailing ':', though it's optional with the tools we support, it's
more common to use it to mark a label. this also quiets the
orphan-labels warning with nasm/yasm.
BUG=b/29583530
Change-Id: I46e95255e12026dd542d9838e2dd3fbddf7b56e2
For real-time mode, increase variance threshold for 32x32 blocks in
var-based partitioning for resolution >= 720p, so that it is more
likely to stay at 32x32 for high resolution which accelerates the
encoding speed with little/no PSNR drop.
PSNR effect on different speed settings:
speed 8 rtc: 0.02 overall PSNR drop, 0.285% SSIM drop
speed 7 rtc: 0.196% overall PSNR increase, 0.066% SSIM increase
speed 5 rtc_derf: no effect.
Speed up:
gips_motion_WHD, 1mbps: 2.5% faster on speed 7, 2.6% faster on speed8
gips_stat_WHD, 1mbps: 4.6% faster on speed 7, 5.6% faster on speed8
Change-Id: Ie7c33c4d2dd7d09294917e031357fc5476c3a4bb
Avoids a segfault in high-bitdepth builds.
This restores the condition to its state prior to:
7991241 vp9: Change the scheme for modeling rd for bsize 32x32.
BUG=webm:1250
Change-Id: I6183d5b34cb89dfbf27b7bb589812148a72cd7de
For real-time CBR mode, use model_rd_for_sb_y_large instead of
model_rd_for_sb_y for 32x32 block. In the former model, transform
might be skipped more aggressively in some condtions, which speeds
up encoding time with only a little PSNR/SSIM drop on rtc test set.
No obvious visual quality regression.
PSNR effect on different speed settings:
speed 8 rtc: 0.129% overall PSNR drop, 0.137% SSIM drop
speed 7 rtc: 0.135% overall PSNR drop, 0.062% SSIM drop
speed 5 rtc_derf: 0.105% overall PSNR drop, 0.095% SSIM drop
Speed up:
gips_motion_WHD, 1mbps: 3.29% faster on speed 7, 2.56% faster on speed8
gips_stat_WHD, 1mbps: 2.17% faster on speed 7, 1.62% faster on speed8
BUG=webm:1250
Change-Id: I818babce5b8549b4b1a7c3978df8591bffde7173
Use quotes whenever possible and {} always for variables.
Replace multiple set_all calls with *able_feature().
Change-Id: If579d3f718bd4133cf1592b4554a8ed00cf9f2d3
decoder_peek_si_internal could potentially read more bytes than
what actually exists in the input buffer. We check for the buffer
size to be at least 8, but we try to read up to 10 bytes in the
worst case. A well crafted file could thus cause a segfault.
Likely change that introduced this bug was:
https://chromium-review.googlesource.com/#/c/70439 (git hash:
7c43fb6)
BUG=chromium:621095
Change-Id: Id74880cfdded44caaa45bbdbaac859c09d3db752
When building without multithreading and for a non-arm, non-x86 system,
ctx is unused.
Cleans up -Wextra warning:
unused parameter ‘ctx’ [-Werror=unused-parameter]
Change-Id: Ifddff89d2ebd45f7d71e3d415a8f2415dd818957
'duration' is not used in realtime-only mode:
Cleans up -Wextra warning:
unused parameter 'duration' [-Wunused-parameter]
Change-Id: I827dfe59ebcdc72c5a93fdf7e5aca063433914b1
In vp9_pick_inter_mode(), instead of using
vp9_get_pred_context_switchable_interp(xd) to assign filter_ref,
we use a less strict condition on assigning filter_ref.
This is to reduce the probabily of entering the flow of not
assigning filter_ref and then skipping filter search.
Overall PSNR gain 0.074% for rtc dataset
Details:
Low Mid High
0.185% -0.008% -0.082%
Change-Id: Id5c5ab38d3766c213d5681e17b4d1afd1529e676
Allows building simple targets with sane default flags.
For example, using the Android arm64 toolchain from the NDK:
https://developer.android.com/ndk/guides/standalone_toolchain.html
./build/tools/make-standalone-toolchain.sh --arch=arm64 \
--platform=android-24 --install-dir=/tmp/arm64
CROSS=/tmp/arm64/bin/aarch64-linux-android- \
~/libvpx/configure --target=arm64-linux-gcc --disable-multithread
BUG=webm:1143
Change-Id: I06f5a7564f5382cf1a4bad41aef4308566c53adf
Speed test shows the new vertical filters have degradation on Celeron
Chromebook. Added "X86_SUBPIX_VFILTER_PREFER_SLOW_CELERON" to control
the vertical filters activated code. Now just simply active the code
without degradation on Celeron. Later there should be 2 set of vertical
filters ssse3 functions, and let jump table to choose based on CPU type.
Change-Id: I37e3e9c5694737d9134a6bce6698d3e43f8fc962
For real-time CBR mode, use model_rd_for_sb_y_large instead of
model_rd_for_sb_y for 32x32 block. In the former model, transform
might be skipped more aggressively in some condtions, which speeds
up encoding time with only a little PSNR/SSIM drop on rtc test set.
No obvious visual quality regression.
PSNR effect on different speed setting:
speed 8 rtc: 0.129% overall PSNR drop, 0.137% SSIM drop
speed 7 rtc: 0.135% overall PSNR drop, 0.062% SSIM drop
speed 5 rtc_derf: 0.105% overall PSNR drop, 0.095% SSIM drop
Speed up:
gips_motion_WHD, 1mbps: 3.29% faster on speed 7, 2.56% faster on speed8
gips_stat_WHD, 1mbps: 2.17% faster on speed 7, 1.62% faster on speed8
Change-Id: I902f62def225ea01c145d7e5a93497398b8f5edf
development has moved to the nextgenv2 branch and a snapshot from here
was used to seed aomedia
BUG=b/29457125
Change-Id: Iedaca11ec7870fb3a4e50b2c9ea0c2b056a0d3c0
2016-06-17 18:26:08 -07:00
1200 changed files with 194033 additions and 248920 deletions
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
