vp9_block_error_sse2 can only handle 16 bytes at a time but
the function requires to handle a sequence of 32 bytes at a time
so each 16 bytes is handled in a different register.
With AVX2 optimization the 32 bytes can be handled in one register instead
of two in the SSE2
The vp9_block_error was optimized by 85%.
The user level was optimized by 1.2%
Change-Id: Ia8fffe60e61eff7432a5fbd538757894f6c319fd
The various motion search functions share a
common function prototype. In the case of
vp9_full_range_search() two of the parameters
are not needed.
Change-Id: I0e190af54a3b3f276409f20e8ec55912f9b0b798
Simplify the calculation of KF bitrate in similar way
to previous patch for GF/arf.
This has no impact on derf or std hd sets but gives a
small net gain of ~0.1% for yt and yt-hd sets.
Change-Id: Ida64ac1428d9c2a62adb67056fadbf0180eff030
The variation in boost calculation for gf and arf groups
is not significant enough to justify the extra complexity.
Also removed some other spurious code that no longer
has much material impact.
The handling of the rare case, where the boost bits
number is less than the number of bits a that would
be allocated if a frame was not boosted, will be dealt
with in a subsequent patch.
This change actually helps on all sets a little by
~0.1% - 0.2% with slightly bigger gains on SSIM.
Change-Id: Id42c1ac22a80a8c4993cfa0e51bc733eb9ed4f75
As a side-effect, the max_sad check is removed from the
C-implementation of VP8, for consistency with VP9, and to
ensure that the SAD tests common to VP8/VP9 pass.
That will make the VP8 C implementation of sad a little slower
but given that is rarely used in practice, the impact will be
minimal.
Change-Id: I7f43089fdea047fbf1862e40c21e4715c30f07ca
Allow slightly larger minq-maxq range for P frames. This improves
the compression performance of speed -5 for rtc set by 2.7% in psnr.
Change-Id: I438653d52d0fe51111509c6092e2334bac2de0cf
Inline loopfilter has been already handled in vp9_decode_frame().
Collecting all similar code in one place now.
Change-Id: I358a0280fc7c2b27cca520bc1e8c16c4eb6491dd
Re-factor duplicate code.
Add two pass check for use of section_intra_rating as
it is un-initialised in the 1 pass and rt case.
Change-Id: I93120796f07961b8a21fb26e1a9f0d3d13949994
One of a series of changes to clean up two pass
allocation as precursor to support for multiple arf
or boosted frames per GF/ARF group.
This change pulls out the calculation of the total bits
allocated to a GF/ARF group into a function, to aid
readability and reduce the line count for define_gf_group().
This change should have no material impact on output.
Change-Id: I716fba08e26f9ddde3257e7d9b188453791883a3
This commit enables a chessboard pattern for partition search. All
the black blocks run regular partition search ranging from 8x8 to
32x32. The rest white blocks take the nearby blocks' information
to adaptively decide the effective search range.
The compression performance for rtc set at speed -5 is down by 1.5%.
For pedestrian 1080p at speed -5, the runtime goes from 41594 ms to
39697 ms, i.e., about 5% faster.
Change-Id: Ia4b96e237abfaada487c743bca08fe1afd298685
tx_mode supercedes whatever mechanism is used to push for 16x16
allowing for the use of the 4x4 transform.
Change-Id: I6c3f05ab9fe52050e40cc6303de9334653763289
vp9_is_upper_layer_key_frame() definition does not match declaration--
it was missing the second const.
Change-Id: I71312579eb443be1924b8b06d8b3177c3dcb40f3
Merged minq tables for arf and gf cases.
These tables were almost the same and for
VBR the arf table was not used at all.
Change-Id: Ie3c87e91dab613cf06f6945ac1ace0e0e4213d34
Small adjustment to the active Q range calculations.
These changes should slightly extend the available Q range
for KF/GF/ARF and narrow it for other frames.
The results for this change in isolation are broadly positive
for SSIM and average PSNR and slightly up but mixed for opsnr.
derf +0.293% opsnr, +1.286% SSIM
std-hd + 0.528% opsnr, + 1.746% SSIM
yt +0.056% opsnr, +0.457% SSIM
yt-hd -0.147% opsnr, + 0.226% SSIM
Change-Id: If065280342027ecc5d44b49fc1d440dfef041002
Includes changes that are not compatible with VS windows builds.
Amongst other things stdint.h is not supported in VS.
This reverts commit 89fbf3de50.
Change-Id: Ifa86d7df250578d1ada9b539c9ff12ed0c523cdd
When the variance is far less than sse, the block is considered to
be under light change. All the energy is compacted into DC coeff
and can be coded at low cost. In such situation, switch the rate-
distortion modeling from sse+var based back to variance based.
Note that this is a temporary solution to handle the rare situations
where the scene light changes.
Change-Id: I1ee0fe2b9eda6b5fac40152e1841bf23f4d229fd
The rounding of the ARNR filter output prior to
normalization by the filter strength was incorrect
when strength = 0.
In this case 1 << (strength - 1) would not create the
required rounding of 0, rather it would outrange. This
patch fixes this issue.
Change-Id: I771809ba34d6052b17d34c870ea11ff67b418dab
This commit enables SSSE3 version full inverse 8x8 2D-DCT and
reconstruction. It makes the runtime of vp9_idct8x8_64_add down
from 256 cycles (SSE2) to 246 cycles.
Change-Id: I0600feac894d6a443a3c9d18daf34156d4e225c3
When ARNR filtering is disabled, by setting
arnr_max_frames=0, mode_skip_mask was being set to
-1 for the ARF frame resulting in no mode being
selected for the block.
The intent is to restrict the reference frame to the
previous ARF frame and the mode to one of ZEROMV,
NEARMV or NEARESTMV.
Change-Id: Ifc3920b153142cd01d422910c94d2f20ffb6f129
On balance Deb's modified rate control for VBR seems
to be outperforming especially on some low motion YT
clips so I have switched this to be the default mode for
now.
Change-Id: I0713d430cad6425ac5c48fccdf332e12814ee44a
Assembly implementation of ssse3 8x8 forward 2D-DCT. The current
version is turned on only for x86_64. The average unit runtime
goes from 157 cycles down to 136 cycles, i.e., about 12.8% faster.
This translates into about 1.5% speed-up for pedestrian_area 1080p
at speed 2.
Change-Id: I0f12435857e9425ed7ce12541344dfa16837f4f4
This reverts commit 59e733ca81.
Hold off removing arnr_type to give users the opportunity
to change their script files to handle its deprecation. A
follow-up patch will mark the control for setting arnr_type
as deprecated and it will be removed completely in a later
revision of the code.
Change-Id: I8b817c744e144d3714234a4cd4309816d0c7e3e8
This member of VP9_COMP seemed unnecessary since it
only shadowed VP9EncoderConfig.key_freq that is
accessible through VP9_COMP.
Change-Id: Ib751bb1cf1b0b3c50a2a527d7c34f6829dd6fee3
The encoder was not handling requests to place keyframes at
fixed intervals, i.e. kf_min_dist == kf_max_dist, correctly.
In this case when looking to place the next keyframe it was
accumulating stats all the way up to the end of the firstpass
file. This patch corrects this behavior.
Change-Id: I948ad9f1d7faa0c05861df588136cce3bb61d7e7
This commit introduces a chessboard pattern search for the prediction
filter type search. It runs extensive search in alternate blocks and
allows the rest blocks to refer coding decisions of their nearby
neighbors.
For pedestrian 1080p at 4000 kbps, the runtime of speed -5 goes down
from 43990 ms to 42200 ms. The overall compression performance for
RTC set is changed by -1.37%.
Change-Id: Icfe220c49451cda796f0ca91d935c9ed01e56c9d
ARNR filtering is now forced to be centered on the ARF
frame and the other two options have been removed.
The other modes of constructing the ARNR frame were
not used and there does not seem to be any good
reason to maintain them.
This is purely an encoder-side change.
Change-Id: Ic772636d23f280752973852b9740083532a49de2
For speed 3 and above, such search is only allowed at speed 3.
The change helped cif and stdhd set by 1.2% and .7% in compression,
but increased the encoding time by around 5%.
Change-Id: Ifa4832327f1c1bef3decb032ceb769cbf50e059f
The end_useage parameter is confusingly named since it
now actually defines the rate control method used.
Change-Id: I98912caabfe556b7af0b939a645d1336409e4d71
This commit enables a background detection approach for adaptive
quantizer control. It combines the cyclic refresh pattern and the
background information to determine the segment id for adaptive
quantizer selection, prior to the non-RD mode decision process.
It hence allows proper quantization information update for a more
precise rate-distortion modeling in the non-RD mode decision.
The compression performance of speed -5 for rtc set is improved
by 2.5%, at no speed change.
Change-Id: Ic3713e8ed9185b403b5b1679d19dabd57506d452
1. We didn't scale source image in lower layers so that
the stats are incorrect.
2. We didn't extend borders for re-constructed image.
Change-Id: Ia8d7bafbdb695ffa7f504e171f9449812e7bb0a3
To make direct side by side testing this patch combines two
VBR corrections schemes to allow more direct side by side testing.
(The other patch was by Debargha chg id I0cd1f7...)
Change-Id: I271c45e5c4ccf8de8305589000218b80d9dc3a25
The background detection only tracks luma component. This commits
removes the frame buffer pointer retrieval for chroma components.
Change-Id: I098bd2950f5e5829ed5dc2b48568167248da7fad
This patch sets up a quad_tree structure (pc_tree) for holding all of
pick_mode_context data we use at any square block size during encoding
or picking modes. That includes contexts for 2 horizontal and 2 vertical
splits, one none, and pointers to 4 sub pc_tree nodes corresponding
to split. It also includes a pointer to the current chosen partitioning.
This replaces code that held an index for every level in the pick
modes array including: sb_index, mb_index,
b_index, ab_index.
These were used as stateful indexes that pointed to the current pick mode
contexts you had at each level stored in the following arrays
array ab4x4_context[][][],
sb8x4_context[][][], sb4x8_context[][][], sb8x8_context[][][],
sb8x16_context[][][], sb16x8_context[][][], mb_context[][], sb32x16[][],
sb16x32[], sb32_context[], sb32x64_context[], sb64x32_context[],
sb64_context
and the partitioning that had been stored in the following:
b_partitioning, mb_partitioning, sb_partitioning, and sb64_partitioning.
Prior to this patch before doing an encode you had to set the appropriate
index for your block size ( switch statement), update it ( up to 3
lookups for the index array value) and then make your call into a recursive
function at which point you'd have to call get_context which then
had to do a switch statement based on the blocksize, and then up to 3
lookups based upon the block size to find the context to use.
With the new code the context for the block size is passed around directly
avoiding the extraneous switch statements and multi dimensional array
look ups that were listed above. At any level in the search all of the
contexts are local to the pc_tree you are working on (in?).
In addition in most places code that used to call sub functions and
then check if the block size was 4x4 and index was > 0 and return
now don't preferring instead to call the right none function on the inside.
Change-Id: I06e39318269d9af2ce37961b3f95e181b57f5ed9
There is no need to initialize source/dst frame buffers at frame
level. These will be done at block coding stage. This commit hence
removes the redundant operations.
Change-Id: I11d9f2556058c6205c8e58ed53e31f78622c41b7
Add code to monitor over and under spend and
apply limited correction to the data rate of subsequent
frames. To prevent the problem of starvation or overspend
on individual frames (especially near the end of a clip) the
maximum adjustment on a single frame is limited to a %
of its un-modified allocation.
Change-Id: I6e1ca035ab8afb0c98eac4392115d0752d9cbd7f
This commit compares the current original frame to the previous
original frame at 64x64 block level and decides if the entire
block belongs to background area. If it is in the background area,
skip non-RD partition search and copy the partition types of the
collocated block in the previous frame.
For vidyo1 in the rtc set, this makes the speed -5 coding speed
about 8% faster. The overall compression performance is down by
1.37% for rtc set.
Change-Id: Iccf920562fcc88f21d377fb6a44c547c8689b7ea
Delete code relating to the old VP8_TUNE_SSIM flag
as this code does not currently work and is largely made
redundant in VP9 by the various AQ modes.
Change-Id: I71f28e1f680573d296422254489000678552b17b
Remove duplicate rd_thresh code introduced when vp9_rd_pick_inter_mode_sub8x8()
was forked from vp9_rd_pick_inter_mode_sb().
Change-Id: I3c9b7143d182e1f28b29c16518eaca81dc2ecfed
Fix rate control bug whereby the rate factor heuristics
were being updated on arf overlays causing a rate surge
for a few frames followed by a corrective drop.
This fix eliminates many of the overshoot problems that
we were seeing on hard clips (even without applying
stricter vbr rate control) and also helps quality on
almost all clips with some hard clips improving by >5%.
Overall quality results measured at speed 2.
Derf +1.78% opsnr , +2.44% SSIM
Stdhd +2.41% opsnr, +2.85% SSIM
Change-Id: I2369df6295c2705963fa6307877f6acb304bcc39
We don't use declarations from this file. The real declarations
(differently named) are in vp9_rtcd_defs.pl, e.g. vp9_full_search_sad.
Change-Id: I73cbf064305710ba20747233cfdbe67366f069a0
Added command line flags "resize-width" & "resize-height"
to allow the user to specify the frame size to encode at.
These two flags are ignored if the "resize-allowed" switch
is not set to 1.
All frames in the clip are then encoded at this size, which
must be smaller than the raw frame size.
Change-Id: I3d64bd9303d5c0bd678461a866a1ea621700d744
A previous path improved speed 2 quality a little but
more extensive testing showed that it slowed encode
by a few %.
The change will have a similar effect for speed 3 but
should not impact speeds 4+;
This experiment should reverse that and give a speed
up at the cost of a small quality loss.
Borg results pending.
Change-Id: I4493fc1541aaf44587f1a41ff219f7088da9252c
Both values are already checked as command line arguments:
RANGE_CHECK_HI(cfg, g_lag_in_frames, MAX_LAG_BUFFERS);
RANGE_CHECK_HI(extra_cfg, sharpness, 7);
Change-Id: I584798d587152d88dfd517c210054b466f4e5f8a
With a more approriate one vp9_setup_src_planes() as only src buffer
pointers need to be initialized here.
Change-Id: I40fac4d8b2d39eb7d0c65b9b6afab45138a13936
This increases the range of Q values available to
normal inter frames to allow encoder a better chance
to hit the target rate.
Change-Id: I33cd96469a46577fdcea631e26d3355710909e6d
The limits applied under the flag
"LIMIT_QRANGE_FOR_ALTREF_AND_KEY"
behaved in an undesirable way if the gap between
active_worst_quality and active_best_quality was
changed.
In this patch, the adjustment is made using the
vp9_compute_qdelta_by_rate() function and fixed
rate multiplier values. Hence it is not impacted by
the relative value of active_best_quality.
Change-Id: I93b3308e04ade1e4eb5af63edf64f91cd3700249
The cause is because VP9 encoding use vp8_vpxyv12_extendframeborders_neon
on arm which only extend boarder size 32. But VP9's border size is 160
Change-Id: I1ff7e945344a658af862beb1197925e677e8ff57
Problem has been introduced recently with the cleanup patch
I0816ec12ec0a6f21d0f25f10c214b5fd327afc6c
Change-Id: Iaacb956a6039eb5826b82618dc03be32053fb892
This commit changed the initialization of best_mode_index to -1 to make
sure it is not mistakenly used for mode masking.
Change-Id: I75b05db51466070dd23c4ee57a4d4b40764dc019
In mode selection loop, once mode_index pass mode_skip_start, all
modes with a different reference frame from current best mode are
masked out using mode_skip_mask.
However, the setting of mode_skip_mask may use an invalid mode if
there is no mode tested yet. This commit fixes the issue by making
sure a mode has been tested and selected. Otherwise, no mode will be
masked out because of their reference frame.
Change-Id: Ib0009e8a96836a65cf5347440fff8a2e1a67f29f
Calculate the difference variance between last source frame and
current source frame. The variance is calculated at 16x16 block
level. The variances are compared to several thresholds to decide
final partition sizes.
An adaptive strategy is implemented to decide using
SOURCE_VAR_BASED_PARTITION or FIXED_PARTITION based on motions
in the video. The switching test is done once every
search_type_check_frequency frames.
The selection of source_var_thresh needs to be investigated
further later.
RTC set Borg test showed 0.424% overall psnr gain, and 0.357%
ssim gain. For clips with large enough static area, the
encoding speedup is around 2% to 15%.
Change-Id: Id7d268f1d8cbca7fb8026aa4a53b3c77459dc156
This commit allows the non-RD mode decision flow to select
prediction filter type in NEWMV mode. It provides 8.14% compression
performance gains in both settings of AQ=0 and 3. The current speed
impact is about 5% to 10% slower.
Change-Id: Id66ecebf77abd8f90fb3f6a066c0e8dfb4bf1c42
Adds some high-level hooks for profile 2 before further
progress on the implementation.
According to the definitiion in this patch:
1. Profile 2 only supports 10 or 12 bit color but not 8
2. Profile 2 supports all color sampling modes: 444, 422 and 420,
and alpha plane.
3. Profile 3 is currently undefined.
Please consider the definition carefully and suggest modifications
to the definition as needed.
Change-Id: I5b284fc679e54ac5aee171af72fa7994cfd28995
Copy up to a certain bsize, otherwise set to a fixed bsize.
This helsp to reduce artifact near moving boundary caused by full partition
copy without checking motion of super-block.
This artifact can occur at speeds 3,4 in real-time mode.
Issue: https://code.google.com/p/webm/issues/detail?id=738.
Change-Id: I05812521fd38816a467f72eb6a951cae4c227931
ARF overlays now use the same rate correction factor as regular inter
frames, further testing would be needed to see if it makes sense to use
a completely separate rate correction factor for ARF overlays.
$ vpxenc --cpu-used=5 --fps=50/1 --target-bitrate=2000
parkjoy.y4m -o out.webm
=> Before: 3356 kb/s
=> After: 2271 kb/s
Change-Id: I73e4defa615ba7a8a2bdb845864f4b1721cbbffe
This commit estimates the motion vector rate cost right after full
pixel motion search. It combines this and the mode cost and compares
the corresponding rate-distortion cost. If it is already above the
current best one, skip the rest sub-pixel motion search and modeling
process. For pedestrian_area 1080p at 4000 kpbs, the speed -5 runtime
goes down from 39425 ms -> 38399 ms.
Change-Id: If4cd7119fd6c266798d5cf1d19d19ab425e52a26
Reinstates this macro and truns it on in order to avoid issues
due to some frames at the end starving in harder videos.
A more acceptable solution is in the works.
Change-Id: I3c46148e86fa6114e3fed245246fb3686a9e6700
This commit slightly increases the bit allocation for key frame. This
improves speed -5 coding performance by 2.77% with aq-mode=0 and by
2.78% with aq-mode=3.
Change-Id: Iaa3e777f80b9706306606af06e89852bac146659
For real-time mode under cbr, this increases the gain (5-10%)
for speed 5 (none/little change for 6), on vc-clips.
Change-Id: I9b38beeb3c820de22c43a0ba53a9456168dd24ba
Turns off the DISABLE_RC_LONG_TERM_MEM macro and makes other changes
in the way the bits are updated, to make 2-pass rate control track
target bitrates closer.
Change-Id: I5f3be4b11c2908e6a9a9a1dd4fcf4e65531c44d8
This commit reduces the frequency of frames using finer quantizer
in non-RD coding flow, and slightly tune up the quantizer resolution
when used. It provides 1.7% compression gains in speed -5 at no speed
difference.
Change-Id: I430249a51260a841a0402666e5ec1566e4f7d5a6
The new tolerance is a little higher than before (especially
for kf/gf/arf) so this change gives an encode speed up
for some clips up for speeds 0-2.
Change-Id: I63f7d6c9cc11c7f58742f41e250dcd3eab1741eb
This code/setting was actually not used (since speed features were not set on first frame,
until a recent change) and should be removed.
In CBR mode, the q value for the first frame can be controlled by setting
the target size via the parameters rc_buf_initial_sz (and max_intra_size_pct).
Change-Id: I65afc64972b36c449bd5a8c25800e65da5389066
While encoding a frame, its last frame source can be used to give
acurate motion information. This patch prevents last frame to be
overwritten so that it is available during current frame encoding.
The last source is scaled when it is necessary. cpi->Last_Source
points to the last source frame.
Change-Id: I0e1ef5e9e1d2badf9d0c7a1a44a7ed5b24c09425
Use a crude correction factor to correct for
lower compression efficiency at higher encode
speeds when estimating the max Q for the
clip.
Change-Id: I5ae377647f4adf5e91d700a8791fb3b8f70efc73
This commit optimizes the bit allocation for the non-RD coding flow.
It applies slightly better quantizer to the frames, where all blocks
run a non-RD partition search. Such frames typically have better
rate-distortion trade off, thus improving the reconstruction quality
for next few frames reference at reasonably low increment in rate
cost.
The coding performance for rtc set at speed -5 with error-resilient
tuned on and rate control set as cbr is improved by 19.58%. It improved
the coding speed by about 10% for a portion of local test clips.
Change-Id: I9d56696cd4359dc8136ca10aff10fff05aaa2686
This commit adjusted the speed steps in rt mode to make the steps
more evenly spaced on speed and quality, specifically:
1. Merged 3 and 4 into one single step 3 and removed confilicting
features.
2. Move 8, 7, 6, 5 to be 7, 6, 5, 4 repsectively.
Change-Id: I38d56d61531f3561d772aef953c411c8fb38c063
Root cause is the different default register length between x86
and x64 platform. Change spatial_layer_id to long long.
Change-Id: If1a5972365c7a59f7e76cb4fd714610f3d48a8ff
Small speed gain for speed 1.
Quality is generally a little up for speed 2.
Speed 3 was similar to speed 4 but now positioned more
evenly between 2 and 4 speed and quality wise.
(opsnr +5.6% ssim +8.25% across all sets)
Speed 4 is a little slower than before but sizable quality gains.
(opsnr +3.7% ssim +6.8% across all sets)
The code has been cleaned up a bit so that for each incremental
speed step changes over the previous speed step are applied.
This makes it easier to see what is changing from one setting to
the next.
Change-Id: I2d98d0d6230af23486adaec01908f58942a7cdeb
Allow tx search for ARF and GF helps quality but a little slower.
Setting subpel_iters_per_step to 1 improves encode speed.
Setting sf->mode_skip_start = 10 improves speed.
Initial local results suggest overall impact on quality is neutral
but encode is up to 15% faster.
Change-Id: Ibde02cae6626a44c10a1da0cefe888afbb51f037
Removes a TODO. Changed meaning of some parameters
(target-max-percent refresh and starting index) to be
defined relative to superblock. Also, modify turn-off condition.
Change-Id: I5e55f372b7079c24f9cdac0b06fa34620dbf456b
This commit enables the non-RD mode decision coding flow to
adaptively apply partition search in non-refresh frame, when the
collocated block in previous frame suggests there might be a motion
activity. It refactors the update_state_rt() function to support
buffer swap of mode_info struct, thereby unifying the encoding
stage across various non-RD coding modes.
It provides 5% compression performance gains in speed -6 for rtc
test set, at about 12% speed slow down.
Change-Id: Iefa374aed5a11c4b7ff9a3ed36a98ea8bd184edb
Fix so that vp9_update_segment_aq() will use the correct (i..e, chosen)
encoding mode (from ctx struct) in update_state.
Change-Id: Icc4b66f3935fad5ec4516a4d57e843d12c365e64
This commit allows the recursive non-RD partition search to early
terminate sub search tree when the cumulative rate-distortion is
already above the best available.
Change-Id: Ifdbcbb4bee229f47fde3033200829577c9f1fc1d
This commit added a speed feature to make the logic of calculating
skip_recode on a block level more explicit. This also enable the
feature to be enabled at speed 5 where the previous logic is too
conservative, help gain back the lost speed for --rt(-5).
Change-Id: Ieb37ca3e85c2e7bda343486edf13d5f5395f2233
There were a few conflicts between the new non-RD partition search
and recent clean-up patches, which were not caught by git control.
This commit fixed these issues.
Change-Id: Ieebefbd6c19d81d0d13e3c568877d5cce2ab7797
This commit takes out the if statements on using adaptive motion
search flag. This feature is automatically enabled in non-RD mode
decision flow for variable partition types search.
Change-Id: I5a25cf9109d84d07aa61b3e02c8d32dda1e90cb0
This commit enables a recursive partition type search for non-RD
mode decisions. It allows the encoder to choose variable block
sizes in a 64x64 block based on rate-distortion modeling.
It improves speed -6 coding efficiency for rtc set by 2.4%. Most
of the gains come from 32-40dB range, where many sequences gain
about 5% to 20%. Local tests suggest there is no speed change.
Change-Id: I06300016e500a21652812b7b3b081db39a783d66
This commit reformats non-RD coding flow layout to allow mode
decision with fixed and variable block sizes.
Change-Id: I2cdd3bb9f26c499ee4a9849004fd925cdd195d09
2 functions were optimized for avx2 by using full 256 bit register
In order to handle 32 elements in parallel instead of only 16 in parallel:
1. vp9_sad32x32x4d
2. vp9_sad64x64x4d
The function level gain is 66% and the user level gain is ~1%.
Change-Id: I4efbb3bc7d8bc03b64b6c98f5cd5c4a9dd3212cb
The third pred_mv is stored in x->pred_mv[ref_frame]. This commit make
sure the correct mv is read.
Change-Id: Ibed24daf36703a63f0394c87b2381ee1d2eb7910
In non-rd pick_mode code, added mode skipping according to
thresholds. Used rd thresholds now, but we can modified them
later for real-time case.
RTC set borg test showed a 0.095% PSNR gain. For different rtc
clips, the real-time(speed 7) encoder speedup is 2% - 10%.
Change-Id: Ic72535c96b891092c662453be32d3168f7e34dcc
The flag x->skip_recode interacts badly with
the cpi->sf.use_nonrd_pick_mode and
cpi->sf.skip_encode_sb speed settings.
Restricting the use of the skip_decode flag when
these other speed choices are in use helps quality
for speeds 3 and 4 by a large amount with only a
small impact on speed.
Average improvmentes for 2 pass speed 4:
Derf +8.8%
Yt + 10.53%
Std-Hd +6.95%
yt-hd + 22.95%
Change-Id: I8010876d8012042a11077c92e69d813c3dfa58eb
This commit changed how q is validated in lossless mode. With this
commit, when --lossless=1 is specificed at commandline, --min-q and
--max-q are now ignored. This is to make the option non-ambiguious.
Change-Id: I33e85690460537509d33be75d6a3597be4affc09
This is an initial attempt to allow variable block size partition
in non-RD coding flow. It tests 8x8, 16x16 and 32x32 block size per
64x64 block, all using non-RD mode decision and the associated rate
distortion costs from modeling, then selects the best block size to
encode the entire 64x64 block. Such operations are triggered every
other 3 frames. The blocks of intermediate frames will reuse the
collocated block's partition type.
It improves the compression performance by 13.2%. Note that the gains
are not evenly distributed. For many hard clips, the compression
performance is improved by 20% to 28%. Local speed test shows that
it will also increase runtime by 50%, as compared to speed -7. It is
now enabled in speed -6 setting.
Change-Id: Ib4fb8659d21621c9075b3c369ddaa9ecb0a4b204
