Rafael de Lucena Valle [Thu, 20 Oct 2016 00:21:09 +0000 (22:21 -0200)]
Add Hadamard for Power8
Change-Id: I3b4b043c1402b4100653ace4869847e030861b18
Signed-off-by: Rafael de Lucena Valle <rafaeldelucena@gmail.com>
Marco Paniconi [Thu, 16 Mar 2017 02:42:55 +0000 (02:42 +0000)]
Merge "vp9: Fix some issues with denoiser and SVC."
Marco [Wed, 15 Mar 2017 23:51:34 +0000 (16:51 -0700)]
vp9: Fix some issues with denoiser and SVC.
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
Marco [Wed, 15 Mar 2017 20:44:26 +0000 (13:44 -0700)]
vp9: Turn off ml_partition_search_early_termination.
Fails on nightly ubsan, valgrind tests.
Enabled on commit:6701014
Change-Id: Ied3f5cb38e39cba54ac134f4514107cdfdfce159
Yi Luo [Wed, 15 Mar 2017 02:32:52 +0000 (02:32 +0000)]
Merge "Improve idct32x32_1024_add SSSE3 intrinsics performance"
Linfeng Zhang [Wed, 15 Mar 2017 00:38:17 +0000 (00:38 +0000)]
Merge "Fix overflow issue in 32x32 idct NEON intrinsics"
Jerome Jiang [Wed, 15 Mar 2017 00:03:52 +0000 (00:03 +0000)]
Merge "vp9: Using source sad for speedup for dynamic resizing."
Linfeng Zhang [Tue, 14 Mar 2017 16:31:52 +0000 (09:31 -0700)]
Fix overflow issue in 32x32 idct NEON intrinsics
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
Jerome Jiang [Tue, 14 Mar 2017 23:29:46 +0000 (23:29 +0000)]
Merge "vp9: Enable row multithreading for SVC in real-time mode."
Jerome Jiang [Mon, 13 Mar 2017 22:27:02 +0000 (15:27 -0700)]
vp9: Using source sad for speedup for dynamic resizing.
Only for speed >= 7.
Change-Id: I3ac85fbb4023cf7e6f8333806b345b0174382a09
Yi Luo [Mon, 6 Mar 2017 23:11:49 +0000 (15:11 -0800)]
Improve idct32x32_1024_add SSSE3 intrinsics performance
- Function level speed improves ~12%.
Change-Id: I9b7dbddabf08c7d0f6b25264e6074d5ccbe39290
James Zern [Tue, 14 Mar 2017 19:21:42 +0000 (19:21 +0000)]
Merge "vp9/encoder: fix segfault on win32 using vs < 2015"
Yunqing Wang [Tue, 14 Mar 2017 18:07:05 +0000 (18:07 +0000)]
Merge "Apply machine learning-based early termination in VP9 partition search"
Marco Paniconi [Tue, 14 Mar 2017 17:50:17 +0000 (17:50 +0000)]
Merge "vp9: Speed >= 8: Enable simple_block_yrd speed feature."
Marco [Tue, 14 Mar 2017 16:17:06 +0000 (09:17 -0700)]
vp9: Adjust copy partition threshold, for speed 8.
Reduce it from 5 to 4, small/no change in metrics or speed.
Small reduction in dragging artifact near moving head.
Change-Id: Ic3bc5ca67c70bf0c89fc2ed14454840a28ae5b6a
Marco [Mon, 13 Mar 2017 05:38:52 +0000 (22:38 -0700)]
vp9: Speed >= 8: Enable simple_block_yrd speed feature.
Enable speed feature for resolutions > VGA.
avgPSNR on RTC down by ~1.7%.
Speedup on ARM: ~5%.
Change-Id: I7a3fe5f7425aa8df3f4a2eced1afa355bc0d4c95
Marco Paniconi [Mon, 13 Mar 2017 19:18:30 +0000 (19:18 +0000)]
Merge "vp9: Fix to source_sad feature for SVC."
Linfeng Zhang [Mon, 13 Mar 2017 18:49:01 +0000 (18:49 +0000)]
Merge "Add vpx_highbd_idct32x32_135_add_c()"
Marco [Wed, 8 Mar 2017 18:57:48 +0000 (10:57 -0800)]
vp9: Fix to source_sad feature for SVC.
Allow speed feature sf->use_source_sad to be used
on highest spatial layer for SVC.
Change-Id: I260eb0478902764f49f83e43b17024fe86ff3b22
Yunqing Wang [Mon, 27 Feb 2017 22:26:15 +0000 (14:26 -0800)]
Apply machine learning-based early termination in VP9 partition search
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
Marco Paniconi [Mon, 13 Mar 2017 06:11:12 +0000 (06:11 +0000)]
Merge "vp9: Fix condition for intra search in non-rd pickmode."
Marco [Sat, 11 Mar 2017 06:50:43 +0000 (22:50 -0800)]
vp9: Fix condition for intra search in non-rd pickmode.
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
James Zern [Fri, 10 Mar 2017 07:29:54 +0000 (23:29 -0800)]
inv_txfm_ssse3,butterfly: fix win32 abi compatibility
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
James Zern [Fri, 10 Mar 2017 07:36:11 +0000 (23:36 -0800)]
vp9/encoder: fix segfault on win32 using vs < 2015
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
Marco Paniconi [Fri, 10 Mar 2017 18:26:06 +0000 (18:26 +0000)]
Merge "vp9: Sample encoder vpx_temporal_svc_encoder: enable row-mt"
Marco [Fri, 10 Mar 2017 16:46:23 +0000 (08:46 -0800)]
vp9: Sample encoder vpx_temporal_svc_encoder: enable row-mt
Enable row-mt in the sample encoder vpx_temporal_svc_encoder.c,
under certain condiitons.
Change-Id: Ic103ee81a9d80be5bf6e5778cc21fc3199db909d
Yi Luo [Fri, 10 Mar 2017 17:14:30 +0000 (17:14 +0000)]
Merge "Improve idct32x32_135_add SSSE3 intrinsics performance"
Marco [Tue, 7 Mar 2017 22:32:30 +0000 (14:32 -0800)]
vp9: Enable row multithreading for SVC in real-time mode.
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
Yi Luo [Fri, 3 Mar 2017 00:52:41 +0000 (16:52 -0800)]
Improve idct32x32_135_add SSSE3 intrinsics performance
- 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
Johann Koenig [Thu, 9 Mar 2017 23:12:36 +0000 (23:12 +0000)]
Merge "ppc: include ppc.h for ppc_simd_caps()"
James Zern [Thu, 9 Mar 2017 22:51:08 +0000 (22:51 +0000)]
Merge "move vp9_scale_and_extend_frame_c to vp9_frame_scale.c"
Johann [Thu, 9 Mar 2017 17:26:45 +0000 (09:26 -0800)]
ppc: include ppc.h for ppc_simd_caps()
Change-Id: Idc829eb066cf4e905d062cb9c08424e0f1b7e1a7
James Zern [Thu, 9 Mar 2017 04:42:35 +0000 (20:42 -0800)]
move vp9_scale_and_extend_frame_c to vp9_frame_scale.c
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
Marco Paniconi [Thu, 9 Mar 2017 03:58:14 +0000 (03:58 +0000)]
Merge "vp9: Enable two speed features for SVC real-time mode."
Marco [Thu, 9 Mar 2017 00:10:45 +0000 (16:10 -0800)]
vp9: Enable two speed features for SVC real-time mode.
Enable short_circuit_low_temp_var and limit_newmv_early_exit
for SVC, 1 pass CBR mode.
Change-Id: I77df2b2c6cc40657bb8ea76e19dfc2fdaad6389e
Marco [Thu, 9 Mar 2017 00:01:58 +0000 (16:01 -0800)]
vp9: Add control to vpx_temporal_svc_encoder for row-mt.
Keep it off as default for now.
Change-Id: Ia2518a8ce96c9735c3fe67215dde25a35e8620af
Jerome Jiang [Wed, 8 Mar 2017 23:14:27 +0000 (23:14 +0000)]
Merge "Shift speed 2 from non-large VP9 tests to large ones."
Johann Koenig [Wed, 8 Mar 2017 22:38:21 +0000 (22:38 +0000)]
Merge "Add support for POWER8/VSX"
Yunqing Wang [Wed, 8 Mar 2017 22:31:30 +0000 (22:31 +0000)]
Merge "Make the partition search early termination feature to be frame size dependent"
Yunqing Wang [Wed, 8 Mar 2017 20:24:15 +0000 (12:24 -0800)]
Make the partition search early termination feature to be frame size dependent
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
Linfeng Zhang [Tue, 7 Mar 2017 21:06:06 +0000 (13:06 -0800)]
Update vpx_idct32x32_1024_add_neon()
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
Rafael de Lucena Valle [Thu, 20 Oct 2016 00:21:09 +0000 (22:21 -0200)]
Add support for POWER8/VSX
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>
Linfeng Zhang [Wed, 8 Mar 2017 18:46:33 +0000 (10:46 -0800)]
Add vpx_highbd_idct32x32_135_add_c()
When eob is less than or equal to 135 for high-bitdepth 32x32 idct,
call this function.
BUG=webm:1301
Change-Id: I8a5864f5c076e449c984e602946547a7b09c9fe6
Marco Paniconi [Wed, 8 Mar 2017 18:26:11 +0000 (18:26 +0000)]
Merge "vp9: Fix for denoising with SVC."
Marco [Wed, 8 Mar 2017 01:35:45 +0000 (17:35 -0800)]
vp9: Fix for denoising with SVC.
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
Linfeng Zhang [Tue, 7 Mar 2017 23:29:15 +0000 (15:29 -0800)]
cosmetics,dsp/arm/: vpx_idct32x32_{34,135}_add_neon()
No speed changes and disassembly is almost identical.
Change-Id: Id07996237d2607ca6004da5906b7d288b8307e1f
Linfeng Zhang [Wed, 1 Mar 2017 23:11:46 +0000 (15:11 -0800)]
cosmetics,dsp/arm/: rename a variable
Rename cospi_6_26_14_18N to cospi_6_26N_14_18N for consistency.
Change-Id: I00498b43bb612b368219a489b3adaa41729bf31a
Jerome Jiang [Tue, 7 Mar 2017 21:58:11 +0000 (13:58 -0800)]
Shift speed 2 from non-large VP9 tests to large ones.
This may fix the time out failure of valgrind tests in nightly
since more coverages were added on row-mt.
Change-Id: Id9414e66d1a266602c7495243d9f5cb69e17ccdc
James Bankoski [Tue, 7 Mar 2017 18:49:13 +0000 (18:49 +0000)]
Merge "tiny_ssim.c : adds y4m support to tiny_ssim."
Jim Bankoski [Thu, 9 Feb 2017 22:12:55 +0000 (14:12 -0800)]
tiny_ssim.c : adds y4m support to tiny_ssim.
Change-Id: I7a13b7e3a1e11ddbe4be3009edf03528e1bc7647
James Zern [Sat, 4 Mar 2017 00:47:17 +0000 (00:47 +0000)]
Merge "vp8_create_decoder_instances: correct pbi[] memset"
Alex Converse [Fri, 3 Mar 2017 23:45:39 +0000 (23:45 +0000)]
Merge "Narrow cat6_high_cost tables to uint16_t"
James Zern [Fri, 3 Mar 2017 23:23:32 +0000 (15:23 -0800)]
vp8_create_decoder_instances: correct pbi[] memset
clear the entire array on error. the size used previously was equal to
the number of elements.
BUG=webm:1364
Change-Id: I2f2e16ed6e867f41d4774a5a8ac9cedaee11ce46
Alex Converse [Fri, 3 Mar 2017 23:02:56 +0000 (15:02 -0800)]
Narrow cat6_high_cost tables to uint16_t
Saves 2688 bytes of rodata.
Change-Id: I46633b6e50c2845181c70fff6273a8e58fdd1e56
Vignesh Venkatasubramanian [Fri, 3 Mar 2017 19:05:52 +0000 (19:05 +0000)]
Merge "vp9,realtime: Enable row multithreading for non-rd"
Marco Paniconi [Thu, 2 Mar 2017 22:25:03 +0000 (22:25 +0000)]
Merge "vp9: Speed 8: reduce the adaptive_rd_thresh level."
Marco [Thu, 2 Mar 2017 21:01:53 +0000 (13:01 -0800)]
vp9: Speed 8: reduce the adaptive_rd_thresh level.
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
Vignesh Venkatasubramanian [Mon, 13 Feb 2017 19:36:02 +0000 (11:36 -0800)]
vp9,realtime: Enable row multithreading for non-rd
Enable row level multithreading for realtime encodes where non-rd
path is used (speed >= 5).
Change-Id: I5439cb49a02171166d8e1de06c7d5e6f8e819a41
Yi Luo [Wed, 1 Mar 2017 00:38:41 +0000 (16:38 -0800)]
Improve idct32x32_34_add SSSE3 intrinsics performance
- Split the transform into first half and second half.
- Reschedule the instructions to avoid stack spillover.
- Function level speed improves ~16%.
Change-Id: I166889840d23aa8a273eca00f6fbdae8b4566f35
Chrome Cunningham [Wed, 1 Mar 2017 18:01:13 +0000 (18:01 +0000)]
Merge "VPX_CODEC_CAP_HIGHBITDEPTH for decoder interface"
Chris Cunningham [Thu, 16 Feb 2017 23:02:30 +0000 (15:02 -0800)]
VPX_CODEC_CAP_HIGHBITDEPTH for decoder interface
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
James Zern [Wed, 1 Mar 2017 00:17:49 +0000 (16:17 -0800)]
Revert "Fix for max qindex calculation of a gf interval"
This reverts commit
d3db846cc50b1b0a9f6efcbe2b36c9c1943bc528.
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
James Zern [Tue, 28 Feb 2017 23:13:11 +0000 (15:13 -0800)]
vp9_ethread_test,cosmetics: s/new-mt/row-mt/
Change-Id: I8c145337adf49d30b88a17ff31501b8751ed1fa0
James Zern [Fri, 24 Feb 2017 08:55:01 +0000 (00:55 -0800)]
stress.sh: add vp9_stress_test_row_mt
vp9_stress_test now forces --row-mt=0 to cover both versions
Change-Id: I8d134879435bf1d8e76ab3fd89e698efba0e86b2
James Zern [Fri, 24 Feb 2017 08:54:02 +0000 (00:54 -0800)]
stress.sh: parameterize thread count
Change-Id: Iae45266cea86585f0935af4012335198cf93719f
James Zern [Fri, 24 Feb 2017 08:30:08 +0000 (00:30 -0800)]
stress.sh: add one pass encodes
Change-Id: I38e6c988f17c56fbfacd95378b27ef8d77c75f90
Yunqing Wang [Tue, 28 Feb 2017 19:13:09 +0000 (11:13 -0800)]
Add a comment in encoder thread test
Added a comment.
Change-Id: I82f71c72598ad6f1eaa0b57b0b8ec56ab9658e81
Yunqing Wang [Tue, 28 Feb 2017 19:00:56 +0000 (11:00 -0800)]
Set row_mt to 0 by default
Set row_mt to 0 for now.
Change-Id: I922536a6d71a765e435daeaf4d932ef14363d19a
Marco [Mon, 27 Feb 2017 20:03:12 +0000 (12:03 -0800)]
vp9: Fix an issue with setting variance thresholds.
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
Vignesh Venkatasubramanian [Mon, 27 Feb 2017 18:50:02 +0000 (10:50 -0800)]
vp9_ethread_test: Rename new_mt to row_mt
Rename left over occurences of new_mt.
Change-Id: Ib884e84c801fcd366ca4b57ec912ac5972023375
Vignesh Venkatasubramanian [Fri, 24 Feb 2017 19:40:22 +0000 (11:40 -0800)]
vp9: Rename new_mt to row_mt
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
Yunqing Wang [Sat, 25 Feb 2017 02:31:21 +0000 (18:31 -0800)]
Remove an old leftover comment
Removed an old comment that wasn't true anymore.
Change-Id: I286ad8d7cb2843070a55e45a599d26bc226d6bd7
James Zern [Fri, 24 Feb 2017 23:36:52 +0000 (15:36 -0800)]
get_prob(): rationalize int types
promote the unsigned int calculation to uint64_t rather than int64_t for
type consistency
Change-Id: Ic34dee1dc707d9faf6a3ae250bfe39b60bef3438
Yunqing Wang [Fri, 24 Feb 2017 23:26:22 +0000 (23:26 +0000)]
Merge "Improve VP9 encoder threading test for better coverage"
Yunqing Wang [Wed, 22 Feb 2017 20:24:16 +0000 (12:24 -0800)]
Improve VP9 encoder threading test for better coverage
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
Jerome Jiang [Fri, 24 Feb 2017 16:56:33 +0000 (16:56 +0000)]
Merge "Make vp9_scale_and_extend_frame_ssse3 work for hbd when bitdepth = 8."
Johann [Fri, 17 Feb 2017 01:57:44 +0000 (17:57 -0800)]
consolidate block_error functions
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
Johann Koenig [Fri, 24 Feb 2017 05:24:34 +0000 (05:24 +0000)]
Merge "block error sse2: use tran_low_t"
Jerome Jiang [Wed, 22 Feb 2017 22:24:02 +0000 (14:24 -0800)]
Make vp9_scale_and_extend_frame_ssse3 work for hbd when bitdepth = 8.
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
Johann [Thu, 16 Feb 2017 20:44:49 +0000 (12:44 -0800)]
block error sse2: use tran_low_t
Change-Id: Ib04990e4a7bda9fbf501f294da2057a2b2595deb
Johann Koenig [Thu, 23 Feb 2017 07:41:20 +0000 (07:41 +0000)]
Merge "vp8_fdct4x4 test: fix segfault again"
Marco Paniconi [Thu, 23 Feb 2017 03:24:26 +0000 (03:24 +0000)]
Merge "vp9: 1pass CBR: modify condition for reducing loop filter."
Jerome Jiang [Wed, 22 Feb 2017 23:19:29 +0000 (23:19 +0000)]
Merge "vp9: Non-rd pickmode: use simple block_yrd under some conditons."
Marco [Wed, 22 Feb 2017 23:06:28 +0000 (15:06 -0800)]
vp9: 1pass CBR: modify condition for reducing loop filter.
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
Marco [Fri, 17 Feb 2017 16:44:50 +0000 (08:44 -0800)]
vp9: Non-rd pickmode: use simple block_yrd under some conditons.
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
Marco Paniconi [Wed, 22 Feb 2017 19:52:24 +0000 (19:52 +0000)]
Merge "vp9: aq-mode=3: On key frame reset cr->reduce_refresh to 0."
Marco [Wed, 22 Feb 2017 18:45:21 +0000 (10:45 -0800)]
vp9: aq-mode=3: On key frame reset cr->reduce_refresh to 0.
This prevent possible reduction of cyclic refresh after key frame.
Change-Id: Idd4e49b69cd95476e7eccfa31b2bd8669569e9e8
Johann [Tue, 21 Feb 2017 19:12:45 +0000 (11:12 -0800)]
vp8_fdct4x4 test: fix segfault again
The output needs to be aligned. Input is read with 'movq' not 'movqda'
so it is not expected to be aligned.
Change-Id: Ibd48a84c1785917a6a97c3689a05322abba486b4
Jerome Jiang [Wed, 22 Feb 2017 17:49:17 +0000 (09:49 -0800)]
vp9: Only compute y_sad for golden in variance partition for speed < 8.
Only affects speed 8. No obvious quality regression. Systematic speed
ups by ~1% on Nexus 6.
Change-Id: Ia904ca28ea041c3281c532911ec38fb7d7f46a17
Yunqing Wang [Wed, 22 Feb 2017 16:55:03 +0000 (16:55 +0000)]
Merge "Refactored the row based multi-threading code"
Jerome Jiang [Wed, 22 Feb 2017 04:44:55 +0000 (04:44 +0000)]
Merge "Fix segmentation fault caused by denoiser working with spatial SVC."
Marco [Mon, 13 Feb 2017 18:16:42 +0000 (10:16 -0800)]
vp9: Incorporate source sum_diff into non-rd partition thresholds.
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
Yi Luo [Tue, 21 Feb 2017 20:07:47 +0000 (12:07 -0800)]
Following SSSE3 intrinsics functions also work for HBD
- vpx_idct8x8_12_add_ssse3
vpx_idct8x8_64_add_ssse3
vpx_idct32x32_34_add_ssse3
vpx_idct32x32_135_add_ssse3
vpx_idct32x32_1024_add_ssse3
- turn on unit tests.
Change-Id: I788b2b3b2074a6f3ab6a0e6f469c1327a123eff7
Johann Koenig [Tue, 21 Feb 2017 18:16:38 +0000 (18:16 +0000)]
Merge "Drop zbin_ptr and quant_shift_ptr"
Jerome Jiang [Sat, 18 Feb 2017 01:56:08 +0000 (17:56 -0800)]
Fix segmentation fault caused by denoiser working with spatial SVC.
Re-enable the affected test.
BUG=webm:1374
Change-Id: I98cd49403927123546d1d0056660b98c9cb8babb
Yi Luo [Tue, 21 Feb 2017 16:36:05 +0000 (16:36 +0000)]
Merge "Fix idct8x8 SSSE3 SingleExtremeCoeff unit tests"
Paul Wilkins [Tue, 21 Feb 2017 09:42:37 +0000 (09:42 +0000)]
Merge "Change to prediction decay calculation."
Marco Paniconi [Tue, 21 Feb 2017 05:37:22 +0000 (05:37 +0000)]
Merge "vp9: Fix for non-rd pickmode for high-bitdepth build."
Marco [Tue, 21 Feb 2017 04:15:40 +0000 (20:15 -0800)]
vp9: Fix for non-rd pickmode for high-bitdepth build.
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
Ranjit Kumar Tulabandu [Thu, 16 Feb 2017 13:37:41 +0000 (19:07 +0530)]
Refactored the row based multi-threading code
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