This reverts commit 5cc8df5bcf.
Reason for revert: <INSERT REASONING HERE>
We need to do this on all key frames in the stream (not just the first one). Will make another cleaner change for this.
Original change's description:
> vp9-svc: Fix to first superframe when inter_layer is off.
>
> When the application selects the setting INTER_LAYER_PRED_OFF
> each spatial stream should be decodeable separately.
> For this we need to force key frames on all spatial layers
> on the first superframe.
>
> In order to maintain the quality at the beginning of the stream
> the active_worst for spatial layer of the second superframe is set
> to the last_QP of the correspondng spatial layer of the first superframe.
> Also make sure nonrd_keyframe is set for non-base spatial layers.
>
> Change only affects SVC mode wit number_spatial_layers > 1 and
> svc->disable_inter_layer_pred == INTER_LAYER_PRED_OFF.
> And only affects first and second frame of sequence.
>
> Change-Id: I8ee9a0873ab1d3a02515774571f719617771ad41
TBR=marpan@google.com,builds@webmproject.org,jianj@google.com
Change-Id: If73d9f3932224fc6751e773763adf7e8ee67d17f
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
When the application selects the setting INTER_LAYER_PRED_OFF
each spatial stream should be decodeable separately.
For this we need to force key frames on all spatial layers
on the first superframe.
In order to maintain the quality at the beginning of the stream
the active_worst for spatial layer of the second superframe is set
to the last_QP of the correspondng spatial layer of the first superframe.
Also make sure nonrd_keyframe is set for non-base spatial layers.
Change only affects SVC mode wit number_spatial_layers > 1 and
svc->disable_inter_layer_pred == INTER_LAYER_PRED_OFF.
And only affects first and second frame of sequence.
Change-Id: I8ee9a0873ab1d3a02515774571f719617771ad41
Cyclic refresh is disabled on key frames, but we did not
disable it for for spatial layers whose base is a key frame
(i.e., on a key-superframe).
This fix means generally somewhat lower frame-level QP will be
used for those spatial layers whose base is a key frame,
which will generally mean little better quality for the
key-superframes.
Change-Id: Idf090651aa2f5856fb6696c89198a9f6d5d50280
In the SVC encoder LAST ref frame should be the last temporal
reference at the same resolution. This is the case for the default/fixed
patterns, but may not be the case for arbitrary pattern in flexible mode.
Add check that the LAST reference frame has same resolution as the current frame.
If the reference scale for LAST is different from current treat the current
frame as key frame just for the purpose of superblock partitioning.
This avoids potential segfault in vp9_int_pro_motion_estimation() for different
scaled reference.
Change-Id: I4276ff616de46cd4e12c73316f85ae313f170242
For the fixed/default SVC patterns, GOLDEN is the
spatial reference, except on key frames, where LAST
is labeled as the spatial reference.
The current code was assuming GOLDEN is always the
spatial reference for the purpose of selecting the
subpel motion (due to the downsampling filter).
Fix is make sure flag_svc_subpel is set and used
with spatial_ref, which is labeled as the proper
spatial reference before entering mode check.
Some quality improvement on key frames.
Change-Id: Id236bcd47055b035731cc910ed84449d7e29f50c
In the constrained framedrop mode for svc: modify the buffer check
condition relative to (non-zero) dropmark to include uppper spatial layers,
in addition to the current spatial layer.
But keep the single layer check if the buffer goes below zero, since
in this case (buffer underflow) we should force drop of that layer
regardless of upper layers.
Change-Id: Id277f0b4a3ae6275effdd5f5f0c80e3229c17424
To reduce the memcpy() cycles in vp9_rd_pick_inter_mode_sb().
The maximum value of mode_map is (MAX_MODES - 1) = 29.
Change-Id: I5704bd66838ea0b075f0afb001f5cbebfd3f1602
Add the logic for the constrained framdrop mode for SVC.
Add test case in datarate unittests.
Also lower target bitrates in the tests to better test
frame dropper.
Change-Id: I8ee1b8cb56d835c233ad1fbe0fc1456cb2e7291f
Add encoder control to set the frame drop thresholds per
spatial layer, and add a frame drop mode: 0 = per-layer drop,
and 1 = constrained drop mode (a drop on a given layer forces
drops to all upper layers).
Default is mode 0 (per-layer dropping).
Implementation for mode 1 will come in subsequent change.
If the control is not used, then the spatial layer frame
drop thresholds (water mark) are all equal and set to the value
given by the encoder config (oxcf->drop_frames_water_mark).
Bump up the ABI version.
Change-Id: Id038d4181b86fa98b3d44d026f96d5f344d81629
adaptive_rd_threshold_mt is set to 1 when speed >= 7 for SVC.
QVGA in SVC uses speed 5 which set adaptive_rd_threshold_mt to 0.
If VGA or HD is dropped for the last super frame, the flag is still 0
when the encoder is destroyed. Thus memory won't be released.
Change the bitrate threshold in datarate test.
Change-Id: I55352cc0b030568d38eb735d99c2fa29058d3690
Compiler -- gcc (Debian 7.3.0-5) 7.3.0
./libvpx/vp9/encoder/vp9_denoiser.c:374:9: assuming signed overflow
does not occur when assuming that (X + c) < X is always false
[-Wstrict-overflow]
for (j = 0; j < xmis; j++) {
Change-Id: Ib7397e718ff717bdabc088fc4c6e1771381fb522
Add VP9E_SET_SVC_INTER_LAYER_PRED to disable inter layer (spatial)
prediction.
0: prediction on
1: prediction off for all frames
2: prediction off for non key frames
Bump up ABI version.
Change-Id: I5ab2a96b47e6bef202290fe726bed5f99bd4951f
SVC frame dropper: modify the logic to allow for individual
spatial layers to drop. This removes the constraint that all
upper spatial layers must drop when a given spatial layer drops.
Add a flag to the pkt to indicate whether a spatial layer is
encoded or dropped. This is needed for applications that enable
this feature (frame dropping for SVC).
For a current spatial layer, if its previous spatial layer is
dropped, then disable certain features for that layer:
inter-layer prediction, base_mv, partition_reuse, copy partition.
Also add the constraint to never drop a spatial layer if its
base layer is a key frame.
Updates to sample encoder (vp9_spatial_svc_encoder) and the
SVC datarate unittests to properly handle frame dropping.
Bump up ABI version.
Change-Id: I7d14ccf67b8d014a7abfce5ba3989fc623e94067
Adjustment to initial active based on image size.
Add extra breakout case for kf boost loop.
Small adjustment to q delta calculation for key frames.
Net % improvements for all standard tests sets (-ve values) measured
using c-bvr mode.
(Overall PSNR, SSIM, PSNR-HVS)
Low Res: -0.223 -0.229 -0.107
Mid Res: -0.175 0.008 -0.180
High Res: -0.131 0.106 -0.206
NFlix 2K: -0.390 -0.271 -0.489
NFlix 4K: -1.370 -0.825 -1.590
Change-Id: I06a39de43594e1a99bb0cb281af15cdb8058a8ed
If a given spatial layer decides to drop, due to the
buffer/overshoot conditions for that layer, then drop
that current spatial layer and all spatial layers above.
In the current implementation the svc frame counter
(and hence the pattern for the non-flexible SVC case)
are updated on frame drops.
Also add last spatial layer encoded to the pkt.
This is useful for RTC applications that enable
frame dropping for SVC.
Update to the SVC datarate tests:
enabled frame dropper on all SVC datarate tests, and
made a fix to properly set the temporal_layer_id, which
works now even on frame drops.
Change-Id: If828c193f3cb6b1839803fd52fe9fbbda5b5a039
When an enhancement spatial layer is skipped, we should check
for updating the layer frame counters.
Change-Id: Ib79d0955c62fb465f59ef2f9ac45240ae2614d7b
This causes assert to trigger in choose_partitioning().
This can happen in some cases when enhancement layers
are enabled midway during the stream.
Change-Id: I69c3c8b4b1e3f1c7d8d7294d633ca5ddca148e8b
For SVC, if any of the layer scale ratios are not
2x2, then disable the partiton_reuse, which assumes
2x2 scaling between layers.
Change-Id: I8b3163de0826052bbb1bfe03554a074c89510558
Set phase_shift = 0 if the scale factors are
above 3/4. Removes artifact for scale factors
close to 1.
phase_shift = 8 is to get an averaging filter
(decimated pixel aligns to 8/16, midway between source pixels),
and only makes sense for scale factors multiples of
2 (1/2, 1/4,...).
Removes artifact for high scaling ratios.
Change-Id: Id0a85869d6c6156dda0032c697ded2de78fad6bd
This change is targeted mainly at higher resolutions where typically
the average error per MB is much smaller. hence this patch replaces
a fixed error per MB factor with a tiered value.
It also adds in a fixed offset value that acts as a minimum return score.
Note also minor fix to debug stats output.
The results are overall beneficial (-ve) on our test sets, most notably for
higher definition formats (see below - overall psnr, ssim, psnr hvs)
low res: 0.184 -0.262 -0.166
mid res: 0.094 0.075 0.049
hd res: -0.752 -0.300 -0.800
NF 2K: -0.353 1.095 -0.302
NF 4K: -1.245 -0.578 -1.205
The most notable negative case is pierseaside 2K which appears to be worse by
8-10% (which has a big impact on the overall gain for the NF 2K set). Closer
inspection reveals that the drop does not relate to the key frame boost
per se as in both cases the key frame substantially undershoots its target. Rather
this is a side effect relating to the initial Q range allowed for the key frame and
a poor initial complexity estimate. This will hopefully be improved in a later
patch.
Change-Id: I4773ebe554782f4024c047c3c392c763a3fe843b
Extended ROI struct suitable for VP9.
ROI input from user is passed into internal struct and applied on every frame
(except key frame).
Enabled usage of all 4 VP9 segment features (delta_qp, delta_lf, skip,
ref_frame) via the ROI map input.
Made changes to nonrd_pickmode for the ref_frame feature.
Only works for realtime speed >= 5.
AQ_MODE needs to be turned off for ROI to take effect.
Change example in the sample encoder: vpx_temporal_svc_encoder.c to be suitable
for VP9.
Add datarate test.
Bump up ABI version.
BUG=webm:1470
Change-Id: I663b8c89862328646f4cc6119752b66efc5dc9ac
This reverts commit 4e5b4b5848.
Reason for revert: Commit message inaccurate.
Original change's description:
> Add ROI support for VP9.
>
> Extended ROI struct suitable for VP9.
> ROI input from user is passed into internal struct and applied on every frame
> (except key frame).
>
> Enabled usage of all 4 VP9 segment features (delta_qp, delta_lf, skip,
> ref_frame) via the ROI map input.
> Made changes to nonrd_pickmode for the ref_frame feature.
>
> Only works for realtime speed >= 5.
> AQ_MODE needs to be turned off for ROI to take effect.
>
> Change example in the sample encoder: vpx_temporal_svc_encoder.c to be suitable
> for VP9.
> Add datarate test.
>
> Bump up ABI version.
>
> BUG=webm:1470
>
> Change-Id: I7e0cf6890649adb98a5fda2efb6ae1fa511c7fc9
TBR=yaowu@google.com,jzern@google.com,marpan@google.com,builds@webmproject.org,jianj@google.com
Change-Id: I000dbd81e0c67cb8a0dcde4013ee9bf7afb038f0
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: webm:1470
This patch adds in detection of slide show key frame groups.
The detection assumes extremely low or 0 motion for all frames
in the key frame group.
If this case is detected the boost level is set to a very high value
and the min Q to a lower value for the key frame itself.
Alt refs and golden frames are disabled to save bits (up to a limiting
maximum interval currently set to 240 frames).
In test samples that I created, this patch gave rise to a substantial
improvement in overall psnr and a drop in data rate. In some cases the
average psnr fell, however, with the boost and minQ values set as they are.
This is to be expected because previously a relatively poor key frame
could be followed by progressively better alt refs. For example a key
frame at q7.5 but subsequent alt refs improving it to lossless. Given that
average psnr tends to be dominated by the best frames, a ramp like this
from q7.5 to lossless may give a better average psnr than, for example,
coding the entire sequence at q2.5. Overall psnr, however, will be much
better in the latter case. The option exists to boost the key frame further
which would insure much better results for all metrics, but at the expense
of smaller bitrate savings. Given that these samples tend to have very
good quality anyway this seems like a bad trade off.
For slides displayed for several seconds, bitrate savings of >= 20% are likely
and much larger gains are possible in some cases.
Change-Id: Ib4b61e153c55d3f2f561153da13fdb56f397a52b
Extended ROI struct suitable for VP9.
ROI input from user is passed into internal struct and applied on every frame
(except key frame).
Enabled usage of all 4 VP9 segment features (delta_qp, delta_lf, skip,
ref_frame) via the ROI map input.
Made changes to nonrd_pickmode for the ref_frame feature.
Only works for realtime speed >= 5.
AQ_MODE needs to be turned off for ROI to take effect.
Change example in the sample encoder: vpx_temporal_svc_encoder.c to be suitable
for VP9.
Add datarate test.
Bump up ABI version.
BUG=webm:1470
Change-Id: I7e0cf6890649adb98a5fda2efb6ae1fa511c7fc9
This value was originally set in response to requests from the hardware
team before levels were properly defined for VP9.
Even if a level is not specified for an encode, it imposes a maximum
frame size for videos of dimensions <= 1080P. For larger formats the
limit was set at 250 bits per MB.
This patch modifies the limit to be more in line with the requirements
specified for level 4 (max rate for a 4 frame group of 16 Mbits). If a lower
level is specified at encode time and this mandates a smaller maximum frame
size then the level requirement will still take precedence.
Increasing this value allows for some slide shows or very low motion clips
to code a better quality key frame.
Change-Id: Ic08e0e09c8a918077152190c59732b9a1c049787
The stats input pointer, when passed in, already points to the
frame after the golden frame so should not be advanced here.
This fix has a small mostly positive effect on results in our test sets
(tested using corpus vbr settings) and gives a gain of almost 0.5%
in overall psnr (plus slightly smaller gains on other metrics) for the
4K set.
The bug also caused a crash in calculate_group_score() in another
patch which allows coding of slides in a slide show as a single
long KF group without ARFs or GFs.
Change-Id: I57a3a24baf442ce55dbc91fba05e056697c63a6f
scaled_temp frame is used as an intermediate buffer for
2 stage down-sampling: two stages of 1/2 down-sampling
for a target of 1/4x1/4. This is used in 3 layer SVC
to avoid duplicate frame downsampling (on middle layer).
As this allocation is only needed/used when the
number_spatial_layers > 2, add this condition to avoid
unneeded allocation for 1 and 2 spatial SVC.
Change-Id: If342466644f685c1ea3ca5344b581793e5136c09
For 3 spatial layers with 1/2 downsampling, the
downsampling filter for the middle layer was not
set for the very first frame, so it was defaulting
to the subsample filter (no averaging/phase = 0).
Its not set due to the two stage scaling that is
done for 1/4 on base layer, during which the intermediate
1/2 result is saved for the middle layer.
Fix for now is to set the default downsampling filter
to Bilinear (averaging/non-zero phase) for all layers on
init (vp9_init_layer_context):.
Change-Id: Ic7407810b34c621e7e7420682508d45478bdffcf
Eliminate false positives in previous patch.
The previous patch did a good job of detecting slide transitions but
in discussions a number of situations were identified that might trigger
harmful false positives. This risk seems to be born out by some testing
on a wider YT set done by yclin@.
This patch adds an additional clause that requires that the best case
inter and intra error for the frame are very similar,meaning it is almost
as easy to code a key frame as an inter frame. This will certainly prevent
the false positive conditions that Jim and I discussed and even if one
does occur it should not be very damaging.
The down side is that this clause may mean that we still miss some
real slide transitions, especially if the images are small and similar. If this
proves to be the case then some further adjustment of the threshold may be
required. However, in the specific problem sample provided we do trap every
transition correctly.
Change-Id: I7e5e79e52dc09bc47917565bf00cc44e5cddd44c
Only affects 2 temporal layer case.
Modified the flags for 2 temporal layers to make
top layer (top spatial, top temporal) a non-reference
frame, conistent with the 3 TL case.
Add mismatch check to the datarate test of changing
svc pattern on the fly, which is test for 2 temporal
layers.
This re-applies the change: 254e2f5501,
that was reverted in: 658eb1d675.
Change-Id: Ib5fd4a7a0312c0c05329ae75baac480af34b4694
This reverts commit 254e2f5501.
Reason for revert: <INSERT REASONING HERE>
Original change's description:
> vp9 svc: fix to make top layer frame non-ref
>
> Add mismatch check to the datarate test of changing svc pattern on the
> fly.
>
> Change-Id: I6a878736de44e6a40c077ed6430aabd7fadabdd9
TBR=marpan@google.com,builds@webmproject.org,jianj@google.com
Change-Id: Ibcb600438098f8dc380fe7e1de90cb81fc367468
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
