paulwilkins [Thu, 5 Jan 2017 12:35:53 +0000 (12:35 +0000)]
Aggressive VBR method.
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
Johann Koenig [Tue, 7 Feb 2017 02:49:58 +0000 (02:49 +0000)]
Merge "highbd x86: consolidate tran_low_t conversions"
Jerome Jiang [Fri, 3 Feb 2017 01:51:01 +0000 (17:51 -0800)]
vp9: speed 8: Tune threshold of ac skip and partitioning.
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
Johann [Thu, 2 Feb 2017 22:17:26 +0000 (14:17 -0800)]
highbd x86: consolidate tran_low_t conversions
Create new helper files specifically for converting tran_low_t types.
Change-Id: I7c4c458ef910f3b3d10a3cfbf9df4de7682fd905
Yunqing Wang [Sat, 4 Feb 2017 01:02:29 +0000 (01:02 +0000)]
Merge "Changes to facilitate multi-threading of encoding stage"
Yunqing Wang [Fri, 3 Feb 2017 16:29:21 +0000 (08:29 -0800)]
Fix visual studio build failure
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
Jerome Jiang [Thu, 2 Feb 2017 22:20:42 +0000 (22:20 +0000)]
Merge "Add unit tests for vp9_block_error_fp."
Jingning Han [Thu, 2 Feb 2017 22:18:31 +0000 (22:18 +0000)]
Merge "Add SSSE3 intrinsic 8x8 inverse 2D-DCT"
Jerome Jiang [Wed, 1 Feb 2017 04:21:55 +0000 (20:21 -0800)]
Add unit tests for vp9_block_error_fp.
BUG=webm:1365
Change-Id: I004e5cd7ca331d14b31b7fc3edeee45fce064026
Johann Koenig [Thu, 2 Feb 2017 20:15:45 +0000 (20:15 +0000)]
Merge "Update third_party/googletest to 1.8.0"
Johann [Thu, 26 Jan 2017 22:34:45 +0000 (14:34 -0800)]
Update third_party/googletest to 1.8.0
Change-Id: If61137e28291f2a0911e9260eb58f234e0d8594c
Ranjit Kumar Tulabandu [Thu, 2 Feb 2017 08:49:29 +0000 (14:19 +0530)]
Changes to facilitate multi-threading of encoding stage
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
Kaustubh Raste [Thu, 2 Feb 2017 08:09:21 +0000 (08:09 +0000)]
Merge "Add mips msa sum_squares_2d_i16 function"
Johann Koenig [Thu, 2 Feb 2017 03:25:31 +0000 (03:25 +0000)]
Merge "Remove neon assembly for idct 16x16 and 8x8"
Johann Koenig [Thu, 2 Feb 2017 03:03:58 +0000 (03:03 +0000)]
Merge changes I43521ad3,I013659f6
* changes:
satd highbd neon: use tran_low_t for coeff
satd highbd sse2: use tran_low_t for coeff
Jingning Han [Wed, 1 Feb 2017 20:19:01 +0000 (12:19 -0800)]
Add SSSE3 intrinsic 8x8 inverse 2D-DCT
The intrinsic version reduces the average cycles from 183 to 175.
Change-Id: I7c1bcdb0a830266e93d8347aed38120fb3be0e03
Yunqing Wang [Wed, 1 Feb 2017 22:04:14 +0000 (22:04 +0000)]
Merge "Changes to facilitate row based multi-threading of ARNR filtering"
Johann Koenig [Wed, 1 Feb 2017 21:56:35 +0000 (21:56 +0000)]
Merge changes I374dfc08,I7e15192e,Ica414007
* 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
Ranjit Kumar Tulabandu [Wed, 1 Feb 2017 11:04:49 +0000 (16:34 +0530)]
Changes to facilitate row based multi-threading of ARNR filtering
Change-Id: I2fd72af00afbbeb903e4fe364611abcc148f2fbb
Johann Koenig [Wed, 1 Feb 2017 20:55:11 +0000 (20:55 +0000)]
Merge "vp9_rdopt: declare 'c' closer to use"
Johann [Tue, 31 Jan 2017 23:35:20 +0000 (15:35 -0800)]
vp9_rdopt: declare 'c' closer to use
Clears up static clang analysis warning regarding a dead store. Only
declare 'c' when it will be used.
Change-Id: I1ac0fc7f94bc44da63938c63cd1efcd6b95e0eb3
Johann Koenig [Wed, 1 Feb 2017 19:57:29 +0000 (19:57 +0000)]
Merge "deblock: annotate postproc parameters"
Johann [Tue, 31 Jan 2017 00:03:05 +0000 (16:03 -0800)]
satd highbd neon: use tran_low_t for coeff
BUG=webm:1365
Change-Id: I43521ad32b6c96737a8ef2b8c327f901fd7eaf84
Johann [Mon, 30 Jan 2017 19:39:02 +0000 (11:39 -0800)]
satd highbd sse2: use tran_low_t for coeff
BUG=webm:1365
Change-Id: I013659f6b9fbf9cc52ab840eae520fe0b5f883fb
Johann [Tue, 31 Jan 2017 16:16:19 +0000 (08:16 -0800)]
hadamard highbd ssse3: use tran_low_t for coeff
BUG=webm:1365
Change-Id: I374dfc08732932382043905f128e928b08cb4f57
Johann [Tue, 31 Jan 2017 00:16:27 +0000 (16:16 -0800)]
hadamard highbd neon: use tran_low_t for coeff
BUG=webm:1365
Change-Id: I7e15192ead3a3631755b386f102c979f06e26279
Johann [Fri, 27 Jan 2017 21:37:36 +0000 (13:37 -0800)]
hadamard highbd sse2: use tran_low_t for coeff
BUG=webm:1365
Change-Id: Ica414007d8412ceebfffa9e58e8416226a3fe934
Johann Koenig [Wed, 1 Feb 2017 19:41:41 +0000 (19:41 +0000)]
Merge "quantize ssse3: remove unused pxor"
Jingning Han [Fri, 27 Jan 2017 18:46:53 +0000 (10:46 -0800)]
Make satd unit test support all bit-depth settings
Turn on satd unit test for c function in both regular and high
bit-depth settings.
Change-Id: I4b0c56addfb84964ede0da3ab760fe0ee640cfd0
Jingning Han [Fri, 27 Jan 2017 18:38:58 +0000 (10:38 -0800)]
Unify the hadamard transform unit test for bit-depth settings
Unify the 8x8 and 16x16 Hadamard unit test system for both 8-bit
and high bit-depth settings.
Change-Id: I53373c1d43f3ced514ad1e53e03f0fb9b25d9ead
Jingning Han [Thu, 26 Jan 2017 23:00:04 +0000 (15:00 -0800)]
Fix real-time compression regression in hbd mode
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
Johann [Tue, 31 Jan 2017 23:05:51 +0000 (15:05 -0800)]
deblock: annotate postproc parameters
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
Johann Koenig [Tue, 31 Jan 2017 15:28:28 +0000 (15:28 +0000)]
Merge "Use Buffer class for post proc tests"
Kaustubh Raste [Tue, 31 Jan 2017 12:07:10 +0000 (17:37 +0530)]
Add mips msa sum_squares_2d_i16 function
average improvement ~4x-5x
Change-Id: I8d91b71d0677009be52b412e4f52b40b98573a53
Kaustubh Raste [Tue, 31 Jan 2017 04:30:43 +0000 (10:00 +0530)]
Add mips msa vpx_minmax_8x8 function
average improvement ~4x-5x
Change-Id: I83aee9977534fddb8a9b80d31af646c0b6b1a8c3
Kaustubh Raste [Tue, 31 Jan 2017 02:34:51 +0000 (02:34 +0000)]
Merge "Add mips msa vpx_vector_var function"
Johann [Tue, 31 Jan 2017 01:02:57 +0000 (17:02 -0800)]
quantize ssse3: remove unused pxor
Change-Id: Ifa22d77fd530827de0b32ae71810dc2213ab2937
Marco [Mon, 30 Jan 2017 19:12:49 +0000 (11:12 -0800)]
vp9: Modify bsize condition for using model_rd_large for speed 7.
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
Yunqing Wang [Sat, 28 Jan 2017 19:29:01 +0000 (19:29 +0000)]
Merge "Disable multi-threading in first pass for SVC encoding"
Kaustubh Raste [Fri, 27 Jan 2017 09:22:51 +0000 (14:52 +0530)]
Add mips msa vpx_vector_var function
average improvement ~4x-5x
Change-Id: I2f63ef83d816052ca8dc42421e7e9d42f7a7af6b
Marco Paniconi [Fri, 27 Jan 2017 22:47:44 +0000 (22:47 +0000)]
Merge "vp9: Fix to pick_filter_level for highbitdepth build."
Jerome Jiang [Fri, 27 Jan 2017 21:15:52 +0000 (21:15 +0000)]
Merge "Add macOS Sierra support in configure"
Marco [Fri, 27 Jan 2017 18:42:29 +0000 (10:42 -0800)]
vp9: Fix to pick_filter_level for highbitdepth build.
Change-Id: I53b3fa8bfc0a0717eb1b730c29f2b70060b1b1b7
Jerome Jiang [Fri, 27 Jan 2017 01:53:00 +0000 (17:53 -0800)]
Add macOS Sierra support in configure
BUG=webm:1367
Change-Id: I3000b6d9f93ec49ca86d08151348d33d86bf0034
Ranjit Kumar Tulabandu [Fri, 27 Jan 2017 13:37:45 +0000 (19:07 +0530)]
Disable multi-threading in first pass for SVC encoding
BUG=webm:1366
Change-Id: I204ef8496884ba7c4debe64f23f50d298b4090c3
Marco Paniconi [Fri, 27 Jan 2017 15:15:36 +0000 (15:15 +0000)]
Merge "vp9: Modify bsize condition for using model_rd_large."
Marco Paniconi [Fri, 27 Jan 2017 15:15:15 +0000 (15:15 +0000)]
Merge "vp9: Fixes for usage of skin_map for high bit depth."
Marco [Fri, 27 Jan 2017 04:01:41 +0000 (20:01 -0800)]
vp9: Modify bsize condition for using model_rd_large.
In non-rd pickmode: small change in behavior for speed 6 and 7.
Remove condition on HIGHBITDEPTH flag.
Change-Id: I360a13fcc313d72612fe9b918162ef4bb278cdea
Kaustubh Raste [Fri, 27 Jan 2017 05:41:42 +0000 (11:11 +0530)]
Add mips msa vpx Integer projection row/col functions
average improvement ~4x-5x
Change-Id: I17c41383250282b39f5ecae0197ef1df7de20801
Kaustubh Raste [Fri, 27 Jan 2017 04:08:51 +0000 (04:08 +0000)]
Merge "Add mips msa vpx satd function"
Marco [Thu, 26 Jan 2017 22:09:57 +0000 (14:09 -0800)]
vp9: Fixes for usage of skin_map for high bit depth.
Also avoid noise_estimation and source_sad if use_highbitdepth is set.
Change-Id: I5fea396b8f8380ea377045d99ba22a52b92daa46
Johann [Sat, 10 Dec 2016 00:51:01 +0000 (16:51 -0800)]
Use Buffer class for post proc tests
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
Jerome Jiang [Thu, 26 Jan 2017 17:46:10 +0000 (17:46 +0000)]
Merge "vp9: Refactor copy partitioning to reduce duplication."
Jerome Jiang [Wed, 25 Jan 2017 19:31:05 +0000 (11:31 -0800)]
vp9: Refactor copy partitioning to reduce duplication.
Change-Id: Ia1b3c118adec5eccbd2900c8e4b9ea6b1e3e9b7c
Yunqing Wang [Thu, 26 Jan 2017 00:32:54 +0000 (00:32 +0000)]
Merge "Remove marco MVC in mcomp.c"
Hui Su [Wed, 25 Jan 2017 22:49:29 +0000 (22:49 +0000)]
Merge "Fix an overflow warning in optimize_b()"
Marco [Wed, 25 Jan 2017 19:39:02 +0000 (11:39 -0800)]
vp9-denoiser: Modify skip denoising condition for small blocks.
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
hui su [Mon, 23 Jan 2017 19:25:31 +0000 (11:25 -0800)]
Fix an overflow warning in optimize_b()
BUG=webm:1361
Change-Id: Ib840bf3b39f7b3c8c017d3488a83434e9a0f45f5
Jerome Jiang [Wed, 25 Jan 2017 17:54:14 +0000 (17:54 +0000)]
Merge "vp9: Adjust threshold for y sad used in copying partition."
Yunqing Wang [Wed, 25 Jan 2017 16:43:02 +0000 (16:43 +0000)]
Merge "Initialize errorperbit and sabperbit in ARNR filtering"
Yunqing Wang [Wed, 25 Jan 2017 16:40:32 +0000 (16:40 +0000)]
Merge "Multi-threading of first pass stats collection"
Jerome Jiang [Wed, 25 Jan 2017 01:20:05 +0000 (17:20 -0800)]
vp9: Adjust threshold for y sad used in copying partition.
Visual quality improvement is observed for noisy clips. Little effects
on speed tests on Nexus 6.
Change-Id: Ib38e04002220708c34102de7b5c36e9940775d89
Ranjit Kumar Tulabandu [Tue, 27 Dec 2016 13:15:43 +0000 (18:45 +0530)]
Multi-threading of first pass stats collection
(yunqingwang)
1. Rebased the patch. Incorporated recent first pass changes.
2. Turned on the first pass unit test.
Change-Id: Ia2f7ba8152d0b6dd6bf8efb9dfaf505ba7d8edee
Marco [Tue, 24 Jan 2017 17:58:05 +0000 (09:58 -0800)]
vp9: Adjust some parameters in aq-mode=3 mode.
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
Jerome Jiang [Tue, 24 Jan 2017 20:58:09 +0000 (20:58 +0000)]
Merge "vp9: Copy partition using avg_source_sad."
Jerome Jiang [Mon, 23 Jan 2017 20:10:01 +0000 (12:10 -0800)]
vp9: Copy partition using avg_source_sad.
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 Wang [Tue, 24 Jan 2017 17:14:07 +0000 (17:14 +0000)]
Merge "Add the multi-threaded first pass encoder unit test"
Ranjit Kumar Tulabandu [Wed, 11 Jan 2017 13:29:11 +0000 (18:59 +0530)]
Initialize errorperbit and sabperbit in ARNR filtering
(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
Kaustubh Raste [Tue, 24 Jan 2017 05:14:22 +0000 (10:44 +0530)]
Add mips msa vpx satd function
average improvement ~4x-5x
Change-Id: If8683d636fe2606d4ca1038e28185bca53bbe244
Jerome Jiang [Tue, 24 Jan 2017 03:43:11 +0000 (03:43 +0000)]
Merge "vp9: Adjust the threshold to set avg_source_sad_sb flag."
Yunqing Wang [Tue, 24 Jan 2017 00:57:34 +0000 (16:57 -0800)]
Remove marco MVC in mcomp.c
Removed MVC so that mv_err_cost() is always called while calculating
the mv cost.
Change-Id: I28123e05fbfc2352128e266c985d2ab093940071
Jerome Jiang [Mon, 23 Jan 2017 23:58:15 +0000 (15:58 -0800)]
vp9: Adjust the threshold to set avg_source_sad_sb flag.
Affect only speed 8. Small/Negligible regression on rtc set.
Change-Id: I67a6b6b4008a22ed798bd980336d95bb799f64b4
Johann [Mon, 23 Jan 2017 21:30:37 +0000 (13:30 -0800)]
PartialIDctTest: reduce number of RunQuantCheck iterations
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
Marco [Mon, 23 Jan 2017 18:53:34 +0000 (10:53 -0800)]
vp9: Non-rd pickmode: fix to add ARF mode entries to THR_MODES.
BUG=webm:1359
Change-Id: Ie0c66efa2e19d1ec9c744d14e3fa8f1e6214cdd6
Marco [Sun, 22 Jan 2017 02:12:45 +0000 (18:12 -0800)]
vp9: Small threshold adjustment to unittest BasicRateTargeting444
Due to recent change to speed >=7 from commit:219cdab.
Change-Id: I366e7750ec91119881050ff6c05849504c7959e8
Kaustubh Raste [Sat, 21 Jan 2017 03:16:38 +0000 (03:16 +0000)]
Merge "Add mips msa vpx hadamard functions"
Jerome Jiang [Fri, 20 Jan 2017 20:35:07 +0000 (20:35 +0000)]
Merge "vp9: Add feature to use block source_sad for realtime mode."
Marco [Tue, 17 Jan 2017 18:39:26 +0000 (10:39 -0800)]
vp9: Add feature to use block source_sad for realtime mode.
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
Yunqing Wang [Fri, 20 Jan 2017 17:15:15 +0000 (09:15 -0800)]
Add the multi-threaded first pass encoder unit test
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
James Zern [Thu, 19 Jan 2017 23:09:20 +0000 (15:09 -0800)]
tools_common.h: add missing ';' in generic branch
missed in:
380a26112 Fix compile warnings for target=armv7-android-gcc
Change-Id: I2820fff00858a19f7dcf6e0fff189d455b7d640f
Johann [Thu, 19 Jan 2017 18:07:12 +0000 (10:07 -0800)]
Remove neon assembly for idct 16x16 and 8x8
Tested using test/partial_idct_test.cc:DISABLED_Speed
Both gcc 4.9 and clang 3.8 from the r13 Android NDK offer improvements
using the intrinsics:
<function> <clang asm> <gcc asm> <clang intrin> <gcc intrin>
idct16x16_256 1720ms 1703ms 1546ms 1554ms
idct16x16_10 1320ms 1247ms 518ms 488ms
idct16x16_1 107ms 108ms 64ms 68ms
idct8x8_64 924ms 931ms 866ms 989ms
idct8x8_12 826ms 824ms 519ms 514ms
idct8x8_1 172ms 166ms 110ms 125ms
idct8x8_64 isn't quite perfect (slight regression with gcc intrinsics)
but as a counter example idct16x16_10 goes from ~1300ms to ~500ms
On a sample clip, clang improved from 48.5 to 49fps and gcc stayed roughly
stable.
BUG=webm:1303
Change-Id: I9d4fd2b41b46ea6174a887b40a82c8e6e4769ed4
Marco [Fri, 13 Jan 2017 23:13:36 +0000 (15:13 -0800)]
vp9: Modify usage of force_skip under low temporal variance in non-rd pickmode.
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
Kaustubh Raste [Thu, 19 Jan 2017 09:14:03 +0000 (14:44 +0530)]
Add mips msa vpx hadamard functions
average improvement ~4x-5x
Change-Id: I167132d894c04fa85dda8dde7906ff9c61b3a65d
Jerome Jiang [Fri, 13 Jan 2017 22:20:59 +0000 (14:20 -0800)]
vp9: Stop copying partition every a fixed number of frames.
Avoid quality loss when copying partition of superblock with large motions.
Maximum consecutively copied frames can be set (currently 5).
Change-Id: I11c30575514f02194c0f001444cf4021609e5049
Peter Boström [Wed, 18 Jan 2017 16:32:34 +0000 (16:32 +0000)]
Merge "Add CSV per-frame stats to vpxdec."
James Zern [Wed, 18 Jan 2017 07:36:53 +0000 (07:36 +0000)]
Merge "vp9_cx_iface,encoder_encode: check validate_img return"
Jerome Jiang [Wed, 18 Jan 2017 02:18:39 +0000 (18:18 -0800)]
vp9: Disable partition copy when resizing is enabled.
Change-Id: I4fa3262e0f1c4018604c954b020ec5d1e3d1465c
Jerome Jiang [Tue, 17 Jan 2017 21:02:52 +0000 (21:02 +0000)]
Merge "vp9: Set low variance flag when partition is copied."
Jerome Jiang [Fri, 13 Jan 2017 19:28:27 +0000 (11:28 -0800)]
vp9: Set low variance flag when partition is copied.
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
Johann Koenig [Tue, 17 Jan 2017 17:45:54 +0000 (17:45 +0000)]
Merge "Cygwin x86_64 support."
Moriyoshi Koizumi [Mon, 16 Jan 2017 10:27:00 +0000 (19:27 +0900)]
Cygwin x86_64 support.
This should have been taken into account at
64347a10
Change-Id: Ie8e3ad7cbaab3e5799e04bd50f2639390b0a2428
Peter Boström [Tue, 17 Jan 2017 15:47:50 +0000 (10:47 -0500)]
Add per-frame SSIM/PSNR stats to tools/tiny_ssim.
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
James Zern [Sat, 14 Jan 2017 00:53:03 +0000 (16:53 -0800)]
vp9_cx_iface,encoder_encode: check validate_img return
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
Marco [Fri, 13 Jan 2017 19:21:33 +0000 (11:21 -0800)]
vp9: Add speed feature flag for computing average source sad.
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
Marco Paniconi [Fri, 13 Jan 2017 19:07:56 +0000 (19:07 +0000)]
Merge "vp9: Adjust threshold for copy partiton, for speed=8."
Marco [Fri, 13 Jan 2017 18:27:51 +0000 (10:27 -0800)]
vp9: Adjust threshold for copy partiton, for speed=8.
Change-Id: I4799cb2b67d911ee385e6d6992c61633ca77e69d
Jingning Han [Fri, 13 Jan 2017 18:25:16 +0000 (18:25 +0000)]
Merge "Rework 8x8 transpose SSSE3 for avg computation"
Jingning Han [Fri, 13 Jan 2017 18:25:09 +0000 (18:25 +0000)]
Merge "Rework 8x8 transpose SSSE3 for inverse 2D-DCT"
Peter Boström [Fri, 13 Jan 2017 12:56:43 +0000 (07:56 -0500)]
Add CSV per-frame stats to vpxdec.
Used with --framestats=file.csv. Currently prints raw codec QP (not
internal 0-63 range) and bytes per frame.
Change-Id: Ifbb90129c218dda869eaf5b810bad12a32ebd82d
Marco Paniconi [Fri, 13 Jan 2017 06:22:52 +0000 (06:22 +0000)]
Merge "vp9: Update threshold for partition copy."