Jingning Han [Sat, 25 Oct 2014 01:44:52 +0000 (18:44 -0700)]
Merge "Tile based adaptive mode search in RD loop"
James Zern [Fri, 24 Oct 2014 18:43:51 +0000 (11:43 -0700)]
Merge changes I8a9c9019,Ic7b2faa3,I44d42a50,I3f3a3924,I10747b32,I31b49c9e
* changes:
add vp9_loop_filter_data_reset
move LFWorkerData allocation to VP9LfSync
vp9_loop_filter_frame_mt: remove pbi dependency
vp9_loop_filter_frame_mt: pass planes directly
vp9_loop_filter_frame_mt: pass VP9LfSync directly
vp9: store TileWorkerData allocations separately
Jingning Han [Thu, 23 Oct 2014 23:54:45 +0000 (16:54 -0700)]
Tile based adaptive mode search in RD loop
Make the spatially adaptive mode search in rate-distortion
optimization loop inter tile independent. Experiments suggest that
this does not significantly change the coding staticstics.
Single tile, speed 3:
pedestrian_area 1080p 1500 kbps
59192 b/f, 40.611 dB, 101689 ms
blue_sky 1080p 1500 kbps
58505 b/f, 36.347 dB, 62458 ms
mobile_cal 720p 1000 kbps
13335 b/f, 35.646 dB, 45655 ms
as compared to 4 column tiles, speed 3:
pedestrian_area 1080p 1500 kbps
59329 b/f, 40.597 dB, 101917 ms
blue_sky 1080p 1500 kbps
58712 b/f, 36.320 dB, 62693 ms
mobile_cal 720p 1000 kbps
13191 b/f, 35.485 dB, 45319 ms
Change-Id: I35c6e1e0a859fece8f4145dec28623cbc6a12325
Marco [Fri, 24 Oct 2014 16:31:38 +0000 (09:31 -0700)]
vp8: Updates to noise level estimate.
-Use full bandwidth (when temporal layers is on) for checking switching.
-Normalize metric wrt num_blocks.
-Rounding fix to update of average noise level metric.
-Make default internal denoiser mode == kDenoiserOnYUV (in denoiser set_parameters()).
-Adjust some thresholds.
Change-Id: Ib827512b25a7bf1f66c76d3045f3a68ce56b1cd2
Paul Wilkins [Fri, 24 Oct 2014 12:51:25 +0000 (05:51 -0700)]
Merge "Enable dual arf with constant q."
Paul Wilkins [Fri, 24 Oct 2014 12:50:39 +0000 (05:50 -0700)]
Merge "Move frame re-sizing into the recode loop"
Adrian Grange [Fri, 17 Oct 2014 22:12:43 +0000 (15:12 -0700)]
Move frame re-sizing into the recode loop
The point at which frames are scaled to their
coded dimensions is moved into the re-code loop.
This is in preparation for a further patch that
will add logic into the re-code loop to reduce
the coded frame size if the encoder is struggling
to hit the target data rate at the native frame
size.
Change-Id: Ie4131f5ec6fb93148879f6ce96123296442bf2d1
Tom Finegan [Thu, 23 Oct 2014 21:20:06 +0000 (14:20 -0700)]
Merge "build/make/iosbuild.sh: Remove armv6 support."
Yaowu Xu [Thu, 23 Oct 2014 18:20:22 +0000 (11:20 -0700)]
Merge "Move filter_ref initialization"
James Zern [Thu, 26 Dec 2013 18:15:56 +0000 (13:15 -0500)]
add vp9_loop_filter_data_reset
Change-Id: I8a9c9019242ec10fa499a78db322221bf96a0275
Yaowu Xu [Thu, 23 Oct 2014 15:27:25 +0000 (08:27 -0700)]
Move filter_ref initialization
To outside the loop to avoid repeating the operations.
Change-Id: I66c1986e98ce0d7594caad3d3b45de655b299bff
Paul Wilkins [Wed, 22 Oct 2014 12:46:40 +0000 (13:46 +0100)]
Enable dual arf with constant q.
Add second level arf Q adjustment when using dual arfs
in constant Q mode.
Previously in constant Q mode enabling dual arf hurt by ~5%
but with this change the average benefit is ~1-1.5% with some
mid range data points up ~10%.
Note however that it still hurts on some clips including
some very low motion show content.
Change-Id: I5b7789a2f42a6127d9e801cc010c20a7113bdd9b
Paul Wilkins [Thu, 23 Oct 2014 09:02:48 +0000 (02:02 -0700)]
Merge "Initialization bug for multi arf."
Jingning Han [Thu, 23 Oct 2014 01:46:20 +0000 (18:46 -0700)]
Merge "Allow checking zeromv mode in vp9_pick_inter_mode"
Yunqing Wang [Thu, 23 Oct 2014 00:10:39 +0000 (17:10 -0700)]
Merge "vp9_ethread: allocate frame contexts outside VP9_COMMON struct"
Tom Finegan [Wed, 22 Oct 2014 23:26:14 +0000 (16:26 -0700)]
build/make/iosbuild.sh: Remove armv6 support.
Apple no longer supports armv6 targets in Xcode or iOS.
Change-Id: Ia7aa496afe62162c43c004530324406486c9a052
Frank Galligan [Wed, 22 Oct 2014 22:50:36 +0000 (15:50 -0700)]
Merge "Fix Neon convolve profiling"
Yunqing Wang [Wed, 22 Oct 2014 21:37:38 +0000 (14:37 -0700)]
vp9_ethread: allocate frame contexts outside VP9_COMMON struct
This patch allocated frame contexts outside VP9_COMMON. This allows
multiple threads to share the same copy of frame contexts, and
reduces the overhead. It also guarantees the correct update of
these contexts during bitstream packing. This patch doesn't change
encoding result.
Change-Id: Ic181a2460b891d1d587278a6d02d8057b9dbd353
Yaowu Xu [Wed, 22 Oct 2014 21:53:06 +0000 (14:53 -0700)]
Merge "Fix a subtle issue in re-use inter_pred"
Jingning Han [Wed, 22 Oct 2014 21:47:15 +0000 (14:47 -0700)]
Allow checking zeromv mode in vp9_pick_inter_mode
This improves the compression performance of speed -5 by 0.6%. The
speed impact is less than 1%.
Change-Id: Ie77daa561976dfc8b479061e1221bdf428eb0c3b
JackyChen [Wed, 22 Oct 2014 20:52:03 +0000 (13:52 -0700)]
Merge "vp9_denoiser_sse2.c: improve code style."
Yaowu Xu [Wed, 22 Oct 2014 18:57:09 +0000 (11:57 -0700)]
Fix a subtle issue in re-use inter_pred
The initialization of this_mode_pred does not work when the ref_frame
loop ever goes beyond LAST_FRAME. This commit fixes the subtle issue
and allows potentially expanding the loop to test GOLDEN_FRAME.
Change-Id: Ibbd427a22160d1d9eacb8ed0c87f88d6cef9c0f3
JackyChen [Fri, 17 Oct 2014 19:16:13 +0000 (12:16 -0700)]
vp9_denoiser_sse2.c: improve code style.
denoiser_sse2.c: fix typos in comment.
Change-Id: Ic0fb102331b0e533c058da3cab1fbc30de9a0070
Frank Galligan [Wed, 22 Oct 2014 17:14:38 +0000 (10:14 -0700)]
Fix Neon convolve profiling
When profiling, gprof can't distinguish between matching labels in
different files.
Change-Id: I56770df212ed314a0d8568071fa8157624ef1e8f
Paul Wilkins [Wed, 22 Oct 2014 17:51:07 +0000 (18:51 +0100)]
Initialization bug for multi arf.
Moved erroneous reset of cpi->multi_arf_last_grp_enabled.
Change-Id: Ibb0b96f6ed1d5eeb575a3b1c798e0fe2ee651d06
Marco [Wed, 22 Oct 2014 17:36:03 +0000 (10:36 -0700)]
vp8: Update to computation of noise level.
Avoid unneeded computation of mse in process_denoise.
Change-Id: I3a538178f2d138750d0c38ea4dd05dbbc2b7df41
Jingning Han [Wed, 22 Oct 2014 15:41:36 +0000 (08:41 -0700)]
Merge "Refactor rate distortion cost structure in non-RD coding mode"
Yaowu Xu [Wed, 22 Oct 2014 15:40:15 +0000 (08:40 -0700)]
Merge "Change speed features for good quality(cpu-used=5)"
Jingning Han [Tue, 21 Oct 2014 23:31:37 +0000 (16:31 -0700)]
Refactor rate distortion cost structure in non-RD coding mode
This commit refactors the rate distortion structure used in the
non-RD coding mode and saves a few RDCOST calculations.
Change-Id: I62c3416c300d2c5372f21b96d93a6b633a34ab3a
Yaowu Xu [Mon, 20 Oct 2014 21:41:20 +0000 (14:41 -0700)]
Change speed features for good quality(cpu-used=5)
The existing speed features produce horrible encoding results, almost
30% worse than cpu-used=4, this commit adjust the speed features to
produce relatively resonable results to be within 3%-5% of cpu-used=4.
Change-Id: I0ca6ebafb33024d4a0cbcf04c78a4a00b8dd1ecf
Jingning Han [Tue, 21 Oct 2014 16:47:58 +0000 (09:47 -0700)]
Remove unused copy_partitioning
Change-Id: I75a2a3772ed17e73180eb4f263cc838cae4927b0
Deb Mukherjee [Tue, 21 Oct 2014 16:44:46 +0000 (09:44 -0700)]
Merge "Adding test vectors for profile 2 and 3"
Jingning Han [Tue, 21 Oct 2014 16:44:11 +0000 (09:44 -0700)]
Merge "Remove deprecated constrain_copy_partitioning function"
Jingning Han [Tue, 21 Oct 2014 16:43:55 +0000 (09:43 -0700)]
Merge "Remove unused sb_has_motion function in vp9_encodeframe.c"
Jingning Han [Tue, 21 Oct 2014 16:43:45 +0000 (09:43 -0700)]
Merge "Remove deprecated use_lastframe_partitioning feature"
Jingning Han [Tue, 21 Oct 2014 16:43:35 +0000 (09:43 -0700)]
Merge "Hybrid partition search for rtc coding mode"
Paul Wilkins [Tue, 21 Oct 2014 14:02:24 +0000 (07:02 -0700)]
Merge "Extend --auto-alt-ref so it can enable multi-alt ref."
Paul Wilkins [Tue, 21 Oct 2014 14:02:13 +0000 (07:02 -0700)]
Merge "Resolve compiler warning."
Deb Mukherjee [Tue, 21 Oct 2014 06:51:58 +0000 (23:51 -0700)]
Adding test vectors for profile 2 and 3
Also adds mising test vectors for profile 1.
Change-Id: I5269af8b9e6b708ce50a0d9c69b6cae4bb2870dc
Jingning Han [Mon, 20 Oct 2014 22:34:19 +0000 (15:34 -0700)]
Remove deprecated constrain_copy_partitioning function
Its functionality has been replaced with choose_partitioning and
threshold based control on split mode check.
Change-Id: Ic9bb321df06b524f5c38ea5874dc6f6a8f93c5e3
Jingning Han [Mon, 20 Oct 2014 22:29:58 +0000 (15:29 -0700)]
Remove unused sb_has_motion function in vp9_encodeframe.c
Change-Id: I035fb6aa5c10741b065e27befb097d8087e3c62f
Jingning Han [Mon, 20 Oct 2014 22:00:24 +0000 (15:00 -0700)]
Remove deprecated use_lastframe_partitioning feature
This speed feature has been deprecated in both yt and rtc coding
modes. This commit removes the related operations.
Change-Id: I079c79c6adafe45581af2ebf8b98faebcface1ce
Jingning Han [Fri, 17 Oct 2014 15:58:28 +0000 (08:58 -0700)]
Hybrid partition search for rtc coding mode
This commit re-designs the recursive partition search scheme in
rtc speed -5. It first checks if the current block is under cyclic
refresh mode. If so, apply recursive partition search. Otherwise,
perform sub-sampled pixel based partition selection. When the
pre-selection finds the partition size should be 32x32 or above,
use the partition size directly. Otherwise, apply partition search
at nearby levels around the preset partition size.
It is enabled in speed -5. The compression performance of rtc
speed -5 is improved by 9.4%. Speed wise, the run-time goes slower
from 1% to 10%.
nik_720p, 1000 kbps
33220 b/f, 38.977 dB, 10109 ms -> 33200 b/f, 39.119 dB, 10210 ms
vidyo1_720p, 1000 kbps
16536 b/f, 40.495 dB, 10119 ms -> 16536 b/f, 40.827 dB, 11287 ms
Change-Id: I65adba352e3adc03bae50854ddaea1b421653c6c
Yunqing Wang [Mon, 20 Oct 2014 19:37:55 +0000 (12:37 -0700)]
Merge "SAD32xh and SAD64xh for AVX2"
Yunqing Wang [Mon, 20 Oct 2014 15:26:46 +0000 (08:26 -0700)]
Merge "Remove the dependency in token storing locations"
Paul Wilkins [Mon, 20 Oct 2014 11:05:17 +0000 (12:05 +0100)]
Extend --auto-alt-ref so it can enable multi-alt ref.
Extend --auto-alt-ref from parameter so we can use it to
turn multi-arf on and off from the command line.
For now the range is 0-off, 1-on, 2-multi-arf on.
Rename play_alternate to enable_auto_arf
Change-Id: Id7b64407cfbe76ba0090a83b588a03e22a240386
Paul Wilkins [Mon, 20 Oct 2014 11:08:33 +0000 (12:08 +0100)]
Resolve compiler warning.
conversion from 'const int64_t' to 'int', possible loss of data.
Change-Id: I471a73bba5d448d9be0ef9cbf1590fa73aa74be1
Paul Wilkins [Mon, 20 Oct 2014 10:12:06 +0000 (03:12 -0700)]
Merge "Alter adjustment of two pass GF/ARF boost with Q."
levytamar82 [Thu, 2 Oct 2014 06:47:31 +0000 (23:47 -0700)]
SAD32xh and SAD64xh for AVX2
All sad function that process above 32 consecutive elements are optimized
for AVX2:
vp9_sad64x64
vp9_sad64x32
vp9_sad32x64
vp9_sad32x32
vp9_sad32x16
vp9_sad64x64_avg
vp9_sad64x32_avg
vp9_sad32x64_avg
vp9_sad32x32_avg
vp9_sad32x16_avg
The functions that appeared as a hotspot is vp9_sad32x32 and vp9_sad64x64
vp9_sad32x32 was optimized by 68% and vp9_sad64x64 was optimized by 90%
both of them gave and overall ~2.3% user level gain
Change-Id: Iccf86b375a2b54c5fbbe685902ead0c9a561b9fd
Debargha Mukherjee [Sat, 18 Oct 2014 21:37:10 +0000 (14:37 -0700)]
Merge "Add highbitdepth function for vp9_avg_8x8"
Yaowu Xu [Sat, 18 Oct 2014 20:31:47 +0000 (13:31 -0700)]
Merge "Remove unused VAR_BASED_FIXED_PARTITION flag"
Yaowu Xu [Sat, 18 Oct 2014 20:31:41 +0000 (13:31 -0700)]
Merge "Use rate/distortion thresholds to control non-RD partition search"
Peter de Rivaz [Thu, 16 Oct 2014 13:36:07 +0000 (14:36 +0100)]
Add highbitdepth function for vp9_avg_8x8
Cherry-picked from https://gerrit.chromium.org/gerrit/#/c/71914/
(
a92f987a6b7819ae5c62a429e126e1c26bdb1b71) on highbitdepth branch.
Change-Id: I6903e4e4cb57d90590725c8a1c64c23da7ae65e8
Marco [Fri, 17 Oct 2014 22:55:30 +0000 (15:55 -0700)]
Updates to aggressive denoising mode.
Fix to the noise level and bitrate threshold settings per resolution.
Change-Id: I25524b7f78afa1aeb76edef7be13c17bc8976508
Yunqing Wang [Thu, 16 Oct 2014 19:29:48 +0000 (12:29 -0700)]
Remove the dependency in token storing locations
Currently, the tokens for a tile are stored immediately after its
preceding tile, which causes a dependency. This is unnecessary
since we always allocate enough memory for tokens. Removing
the dependency allows token writing done in parallel. This patch
doesn't change encoding result.
Change-Id: I7365a6e5e2c2833eb14377c37e1503c9d0f26543
hkuang [Fri, 17 Oct 2014 21:13:44 +0000 (14:13 -0700)]
Merge "Correct the logic of ready_for_new_data."
JackyChen [Fri, 17 Oct 2014 18:16:22 +0000 (11:16 -0700)]
Merge "vp9_denoiser_sse2.c: solve windows build error."
Jingning Han [Fri, 17 Oct 2014 18:15:19 +0000 (11:15 -0700)]
Merge "Add init and reset functions for RD_COST struct"
Jingning Han [Fri, 17 Oct 2014 18:15:03 +0000 (11:15 -0700)]
Merge "Reset rate cost value in rd mode search"
hkuang [Fri, 17 Oct 2014 17:09:56 +0000 (10:09 -0700)]
Correct the logic of ready_for_new_data.
This should be set right after decoder really start to decode frame
instead setting at the end.
Even decoder does not have a displayable frame to show and return NULL
to application, this should be set too.
Change-Id: If0313a834bc64e3b0f05a84f4459d444d9eab0d8
Jingning Han [Fri, 17 Oct 2014 16:31:42 +0000 (09:31 -0700)]
Reset rate cost value in rd mode search
When early termination is triggered, properly reset the rate cost
to invalid value to avoid potential ioc issue.
Change-Id: I3444390be2e49a34bb02cf8a74c33d5dbd96d88d
JackyChen [Fri, 17 Oct 2014 16:28:22 +0000 (09:28 -0700)]
vp9_denoiser_sse2.c: solve windows build error.
Change-Id: Ib5df91c8580d5dbeb0b3554edc9c2ca906ba4c4d
Jingning Han [Wed, 15 Oct 2014 22:45:27 +0000 (15:45 -0700)]
Remove unused VAR_BASED_FIXED_PARTITION flag
Change-Id: I4ce19b7cb1c45fed86e81ee785e787630020fb4f
Paul Wilkins [Thu, 16 Oct 2014 15:24:25 +0000 (16:24 +0100)]
Alter adjustment of two pass GF/ARF boost with Q.
Delete gfboost_qadjust() and move Q based adjustment
into calc_frame_boost(). Also remove clamping. Making
the adjustment here means that it influences not just the
boost level but also the selection of the GF/ARF interval.
This change gives a small average gain in PSNR but
larger gains in SSIM, especially for harder std-hd set (1.5%)
Change-Id: I3aa81b8feccaeff93d915e19fb9cf5cd64c86327
James Zern [Fri, 17 Oct 2014 10:26:06 +0000 (03:26 -0700)]
Merge "vp9_denoiser_sse2.c: eliminate gcc warnings"
JackyChen [Mon, 13 Oct 2014 06:29:10 +0000 (23:29 -0700)]
vp9_denoiser_sse2.c: eliminate gcc warnings
Change-Id: I5f63f48e11e31ea9951223c5b18f42a2471e4560
Deb Mukherjee [Thu, 16 Oct 2014 21:27:26 +0000 (14:27 -0700)]
Merge "Adds a set of end-to-end encode tests"
Jingning Han [Thu, 16 Oct 2014 19:35:11 +0000 (12:35 -0700)]
Merge "Fix an ioc issue in super_block_uvrd"
Deb Mukherjee [Wed, 15 Oct 2014 23:40:12 +0000 (16:40 -0700)]
Adds a set of end-to-end encode tests
Covers all profiles and input formats. The tests check if the
encode succeeds and if the psnr is sane.
Change-Id: I195a5330debf92562846121819b6eaf961e27c01
Jingning Han [Thu, 16 Oct 2014 16:19:42 +0000 (09:19 -0700)]
Fix an ioc issue in super_block_uvrd
This commit fixes an ioc issue that will happen when the cumulative
variables are not in effective use. The fix discards these
redundant additions.
Change-Id: Idbac5bfb989c0cedc5f8a323effce938519b2457
Paul Wilkins [Thu, 16 Oct 2014 17:46:20 +0000 (10:46 -0700)]
Merge "Change initialization of static_scene_max_gf_interval."
Paul Wilkins [Thu, 16 Oct 2014 17:46:08 +0000 (10:46 -0700)]
Merge "[spatial svc]Another workaround to avoid using prev_mi"
Paul Wilkins [Thu, 16 Oct 2014 17:44:58 +0000 (10:44 -0700)]
Merge "Revert "Move input frame scaling into the recode loop""
Paul Wilkins [Thu, 16 Oct 2014 17:44:47 +0000 (10:44 -0700)]
Merge "Revert "[spatial svc]Another workaround to avoid using prev_mi""
James Zern [Sun, 31 Aug 2014 20:16:37 +0000 (13:16 -0700)]
move LFWorkerData allocation to VP9LfSync
this removes an assumption that worker->data1 would be pointing to a
TileWorkerData allocation.
additionally, within the multi-threaded loopfilter pass VP9LfSync as a
parameter to the worker hook, removing the need for a shadow pointer in
LFWorkerData.
Change-Id: Ic7b2faa34e3eb59dbcb8a7c67f333448fa047c88
James Zern [Sun, 31 Aug 2014 19:46:19 +0000 (12:46 -0700)]
vp9_loop_filter_frame_mt: remove pbi dependency
Change-Id: I44d42a5098305a2d050ce8ff3c76baf7798c48af
James Zern [Sun, 31 Aug 2014 19:36:34 +0000 (12:36 -0700)]
vp9_loop_filter_frame_mt: pass planes directly
one less dependency on pbi
Change-Id: I3f3a392416d3523f4aea6682c3965885baf85197
Deb Mukherjee [Thu, 16 Oct 2014 16:23:57 +0000 (09:23 -0700)]
Merge "Fix in bit-shift operation for highbitdepth decode"
James Zern [Sun, 31 Aug 2014 19:21:42 +0000 (12:21 -0700)]
vp9_loop_filter_frame_mt: pass VP9LfSync directly
a step towards removing the pbi dependency
Change-Id: I10747b325e81c172f5e67031ea5159159fc26e91
James Zern [Sun, 31 Aug 2014 03:15:37 +0000 (20:15 -0700)]
vp9: store TileWorkerData allocations separately
move them from VP9Worker::data[12] to allow the structure to be reused a
bit more naturally by the multi-threaded loopfilter.
Change-Id: I31b49c9e93ca744fd7f6d6ed8696671188fb2c1d
Paul Wilkins [Thu, 16 Oct 2014 11:39:14 +0000 (12:39 +0100)]
Change initialization of static_scene_max_gf_interval.
This removes an unnecessary restriction that causes
a problem (noticed by AWG) when the forced key frame
interval is set to a very small value, such as 10. In this case
we were being forced to code minimal length GF groups.
Change-Id: I76ef5861a09638ff51f61fea02359554184ada53
Minghai Shang [Tue, 14 Oct 2014 23:25:03 +0000 (16:25 -0700)]
[spatial svc]Another workaround to avoid using prev_mi
We encode a empty invisible frame in front of the base layer frame to
avoid using prev_mi. Since there's a restriction for reference frame
scaling factor, we have to make it smaller and smaller gradually until
its size is 16x16.
Change remerged.
Change-Id: I9efab38bba7da86e056fbe8f663e711c5df38449
Paul Wilkins [Thu, 16 Oct 2014 14:53:21 +0000 (15:53 +0100)]
Revert "Move input frame scaling into the recode loop"
This reverts commit
452dc21500a2339ee685cb28efbd2af1b856ea12.
This change has introduced a significant quality regression on content
with forced key frames. (e.g. the YT and yt-hd set). It is most
noticeable in static content where the kf bits dominate. Here, despite
key frames being apparently coded at the same Q, there is a drop in all
metrics of ~20% (e.g clXR and BFa0).
Change-Id: Iba14cc61778c0846fa0a59c33c55a9fc49512cb4
Paul Wilkins [Thu, 16 Oct 2014 14:49:13 +0000 (15:49 +0100)]
Revert "[spatial svc]Another workaround to avoid using prev_mi"
This reverts commit
c113457af9880b8e15a36cdaabfd414d1c245693.
Temporary revert to allow clean revert of another commit.
Change-Id: Ia9b7b755e6c48e1b6e383329f121fef175a24b27
James Zern [Thu, 16 Oct 2014 08:09:15 +0000 (01:09 -0700)]
Merge "fix CONFIG_SPATIAL_SVC warning"
Marco [Wed, 15 Oct 2014 22:57:21 +0000 (15:57 -0700)]
Merge "Some updates for Speed 6/VAR_BASED_PARTITION."
Jingning Han [Tue, 14 Oct 2014 17:26:28 +0000 (10:26 -0700)]
Add init and reset functions for RD_COST struct
Change-Id: I2902de7051a883fd22e27a655209233733969cfd
Jingning Han [Wed, 15 Oct 2014 21:54:52 +0000 (14:54 -0700)]
Merge "Replace copy_partitioning use case with choose_partitioning"
Jingning Han [Wed, 15 Oct 2014 19:18:48 +0000 (12:18 -0700)]
Use rate/distortion thresholds to control non-RD partition search
Compare the estimated rate and distortion to the thresholds scaled
according to the operating block size and determine if further
split partition search will be run. The compression performance of
speed -5 is changed by -0.074%. The encoding speed is 10% - 15%
faster.
vidyo1 720p
16545 b/f, 40.492 dB, 11475 ms -> 16535 b/f, 40.486 dB, 10100 ms
nik720p
16624 b/f, 36.310 dB, 10071 ms -> 16617 b/f, 36.313 dB, 8346 ms
Change-Id: Ic9197ab5761279ae55d2fb7813b2af0e0db497b8
Marco [Thu, 9 Oct 2014 23:01:52 +0000 (16:01 -0700)]
Some updates for Speed 6/VAR_BASED_PARTITION.
Reduce the intra_cost_penalty for non-rd mode,
and some updates to VAR_BASED_PARTITION.
Visual tests show some improvement at Speed 6, for RTC clips.
Change-Id: If9090daf7aed14906a32d931a538ab544bbca606
Jingning Han [Wed, 15 Oct 2014 18:37:20 +0000 (11:37 -0700)]
Replace copy_partitioning use case with choose_partitioning
This commit replaces the use of copy_partitioning with
choose_partitioning based on the sse of subsamped pixels, which
provides significantly better coding performance and runs at
similar speed, as compared to copy_partitioning. It improves rtc
speed 5 coding performance by 3%.
Change-Id: I52d3682a12dce0147f5e52383a594fc242ca3228
Deb Mukherjee [Wed, 15 Oct 2014 17:01:34 +0000 (10:01 -0700)]
Fix in bit-shift operation for highbitdepth decode
Fixes a bug introduced in a previous refactoring patch.
Change-Id: I243e74637cfd7a997c7a1fef03b06c290dd0dee6
James Zern [Wed, 15 Oct 2014 14:37:35 +0000 (16:37 +0200)]
fix CONFIG_SPATIAL_SVC warning
this change checks that CONFIG_SPATIAL_SVC is defined and adds a TODO to
ensure this is changed in the future as the release headers can't
depend on vpx_config.h.
vpx/vpx_encoder.h:164:5: warning: "CONFIG_SPATIAL_SVC" is not defined
[-Wundef]
Change-Id: I797a0150e5f56caf048e7ee00b282fbc9c5ede19
Minghai Shang [Tue, 14 Oct 2014 23:25:03 +0000 (16:25 -0700)]
[spatial svc]Another workaround to avoid using prev_mi
We encode a empty invisible frame in front of the base layer frame to
avoid using prev_mi. Since there's a restriction for reference frame
scaling factor, we have to make it smaller and smaller gradually until
its size is 16x16.
Change-Id: I60b680314e33a60b4093cafc296465ee18169c19
Adrian Grange [Tue, 14 Oct 2014 22:30:42 +0000 (15:30 -0700)]
Merge "Move input frame scaling into the recode loop"
Alex Converse [Tue, 14 Oct 2014 21:35:02 +0000 (14:35 -0700)]
Merge "Add a 32-bit friendly sse2 quantizer."
Yunqing Wang [Tue, 14 Oct 2014 21:06:23 +0000 (14:06 -0700)]
Merge "Remove an unneeded function call"
hkuang [Tue, 14 Oct 2014 20:45:11 +0000 (13:45 -0700)]
Merge "Correct the format."
Yunqing Wang [Tue, 14 Oct 2014 18:41:37 +0000 (11:41 -0700)]
Remove an unneeded function call
set_tile_limits() is called in vp9_change_config() already.
Change-Id: I91c3a0df2c1c7fd7e71546d8f51fd5b65838a7da
Alex Converse [Mon, 13 Oct 2014 22:17:05 +0000 (15:17 -0700)]
Add a 32-bit friendly sse2 quantizer.
This is based on the 64-bit ssse3 quantizer.
1.1x speedup for screen content at speed 7.
Change-Id: I57d15415ef97c49165954bbe3daaaf9318e37448