]> granicus.if.org Git - libvpx/log
libvpx
8 years agoMake generic SIMD work with clang.
Steinar Midtskogen [Thu, 1 Sep 2016 17:45:29 +0000 (19:45 +0200)]
Make generic SIMD work with clang.

Change-Id: I2c504a078a7137bea6ba50c5768c1295878e9ea1

8 years agoFix clang-format warnings in aom_dsp/simd/v64_intrinsics_arm.h
Jingning Han [Thu, 1 Sep 2016 19:36:25 +0000 (12:36 -0700)]
Fix clang-format warnings in aom_dsp/simd/v64_intrinsics_arm.h

Change-Id: I221bf4520d7030133e3b2fea883a995b3d6f6282

8 years agoNon-normative quality improvements to CLPF.
Steinar Midtskogen [Fri, 2 Sep 2016 08:56:54 +0000 (10:56 +0200)]
Non-normative quality improvements to CLPF.

BDR improvements:
     PSNR  PSNRHVS SSIM  MSSSIM CIEDE2000 PSNR Cb  PSNR Cr
LL: -0.17% -0.13% -0.11% -0.12%   -0.18%   -0.19%   -0.21%
HL: -0.21% -0.14% -0.15% -0.11%   -0.37%   -0.39%   -0.52%

Change-Id: I58c00a1cc0ddfc3376644f66345e99472482a613

8 years agoAdded high bit-depth support in CLPF.
Steinar Midtskogen [Fri, 9 Sep 2016 13:23:35 +0000 (15:23 +0200)]
Added high bit-depth support in CLPF.

Change-Id: Ic5eadb323227a820ad876c32d4dc296e05db6ece

8 years agoFix a memleak in CLPF.
Steinar Midtskogen [Fri, 9 Sep 2016 15:36:22 +0000 (17:36 +0200)]
Fix a memleak in CLPF.

The memleak appeared in eb5794da1659f87597291d84c2fbdfd89280065d.

Change-Id: Ifdd6d64aafa0d0ce4dfaf1844f594d5f843bf2e0

8 years agoReduce memory footprint for CLPF decoding.
Steinar Midtskogen [Wed, 24 Aug 2016 11:00:04 +0000 (13:00 +0200)]
Reduce memory footprint for CLPF decoding.

Instead of having CLPF write to an entire new frame and
copy the result back into the original frame, make the
filter able to work in-place by keeping a buffer of size
frame_width*filter_block_size and delay the write-back
by one filter_block_size row.

This reduces the cycles spent in the filter to ~75%.

Change-Id: I78ca74380c45492daa8935d08d766851edb5fbc1

8 years agoMake CLPF handle frame widths and heights not divisible by 8.
Steinar Midtskogen [Wed, 7 Sep 2016 06:15:11 +0000 (08:15 +0200)]
Make CLPF handle frame widths and heights not divisible by 8.

Change-Id: If5eb33b6b090f43ba64c82468576b89eddd872c3

8 years agoCLPF: Don't assume sb size=64 and w&h multiple of 8 + valgrind fix.
Steinar Midtskogen [Thu, 25 Aug 2016 10:22:24 +0000 (12:22 +0200)]
CLPF: Don't assume sb size=64 and w&h multiple of 8 + valgrind fix.

Change-Id: I518ad9c58973910eb0bdcb377f2d90138208c570

8 years agoSilence some harmless compiler warnings in CLPF.
Steinar Midtskogen [Fri, 2 Sep 2016 08:02:30 +0000 (10:02 +0200)]
Silence some harmless compiler warnings in CLPF.

Change-Id: I4a6d84007bc17b89cfd8d8f2440bf2968505bd6a

8 years agoAdded generic SIMD support for CLPF.
Steinar Midtskogen [Fri, 5 Aug 2016 10:12:38 +0000 (12:12 +0200)]
Added generic SIMD support for CLPF.

Change-Id: Ie03f9a5b0a4c708a586532198d755a1e7509f149

8 years agoMerge "Added generic SIMD library supporting x86 SSE2+ and ARM NEON." into nextgenv2
Yaowu Xu [Mon, 10 Oct 2016 18:17:50 +0000 (18:17 +0000)]
Merge "Added generic SIMD library supporting x86 SSE2+ and ARM NEON." into nextgenv2

8 years agoMerge "New CLPF: New kernel and RDO for strength and block size" into nextgenv2
Yaowu Xu [Mon, 10 Oct 2016 18:17:41 +0000 (18:17 +0000)]
Merge "New CLPF: New kernel and RDO for strength and block size" into nextgenv2

8 years agoMerge "Hybrid forward transforms 16x16 AVX2 optimization" into nextgenv2
Yi Luo [Fri, 7 Oct 2016 01:52:10 +0000 (01:52 +0000)]
Merge "Hybrid forward transforms 16x16 AVX2 optimization" into nextgenv2

8 years agoMerge "Added sse2 inverse 8x16 and 16x8 transforms" into nextgenv2
Debargha Mukherjee [Fri, 7 Oct 2016 00:03:33 +0000 (00:03 +0000)]
Merge "Added sse2 inverse 8x16 and 16x8 transforms" into nextgenv2

8 years agoMerge "A bug fix for var-tx" into nextgenv2
Debargha Mukherjee [Fri, 7 Oct 2016 00:02:30 +0000 (00:02 +0000)]
Merge "A bug fix for var-tx" into nextgenv2

8 years agoHybrid forward transforms 16x16 AVX2 optimization
Yi Luo [Wed, 21 Sep 2016 17:45:01 +0000 (10:45 -0700)]
Hybrid forward transforms 16x16 AVX2 optimization

- Unit tests are added for AVX2 SIMD.
- Encoder speed improvement:
  AV1 baseline and EXT_TX, three 1080p sequences at bitrate:
  800 Kbps, 2 Mbps, 6 Mbps, on i7-6700 CPU, average
  user level time reduction: 3.86%.

Change-Id: Ibbd7837ee3a831c6b1e4e471bf6c8d3fa3a19ff4

8 years agoFix left shift of negative integer in hbd directional predictors
Alex Converse [Thu, 6 Oct 2016 18:36:08 +0000 (11:36 -0700)]
Fix left shift of negative integer in hbd directional predictors

Change-Id: Id78139ae2dfa2d521bd50618b4a81cf24e09e391

8 years agoAdded sse2 inverse 8x16 and 16x8 transforms
Peter de Rivaz [Thu, 29 Sep 2016 08:14:54 +0000 (09:14 +0100)]
Added sse2 inverse 8x16 and 16x8 transforms

Change-Id: I43628407b11e5c8e6af4df69f2acdc67ac827834

8 years agoMerge "Silence some warnings" into nextgenv2
Debargha Mukherjee [Thu, 6 Oct 2016 18:15:15 +0000 (18:15 +0000)]
Merge "Silence some warnings" into nextgenv2

8 years agoA bug fix for var-tx
Debargha Mukherjee [Wed, 5 Oct 2016 07:48:28 +0000 (00:48 -0700)]
A bug fix for var-tx

Fixes a crash with supertx, ext-tx and rect-tx

Change-Id: I6b5f4cfd6e209558541a791be685b55156aa0138

8 years agoAdded generic SIMD library supporting x86 SSE2+ and ARM NEON.
Steinar Midtskogen [Wed, 3 Aug 2016 11:17:33 +0000 (13:17 +0200)]
Added generic SIMD library supporting x86 SSE2+ and ARM NEON.

Change-Id: I037f4c44f621a7e909b82ccb6a299d41bcbf8607

8 years agoNew CLPF: New kernel and RDO for strength and block size
Steinar Midtskogen [Fri, 6 May 2016 11:48:20 +0000 (13:48 +0200)]
New CLPF: New kernel and RDO for strength and block size

This commit ports a CLPF change from aom/master by manually
cherry-picking:
7560123c066854aa40c4685625454aea03410b18

Change-Id: I61eb08862a101df74a6b65ece459833401e81117

8 years agoMerge "Make ref_mv_idx syntax context dependent on block distance only" into nextgenv2
Jingning Han [Thu, 6 Oct 2016 15:55:40 +0000 (15:55 +0000)]
Merge "Make ref_mv_idx syntax context dependent on block distance only" into nextgenv2

8 years agoMerge "Simplify file dependencies of SIMD implementation of interpolation filters...
Angie Chiang [Wed, 5 Oct 2016 16:26:25 +0000 (16:26 +0000)]
Merge "Simplify file dependencies of SIMD implementation of interpolation filters" into nextgenv2

8 years agoMake ref_mv_idx syntax context dependent on block distance only
Jingning Han [Wed, 5 Oct 2016 16:08:44 +0000 (09:08 -0700)]
Make ref_mv_idx syntax context dependent on block distance only

This allows the hardware decoder to start decoding ref_mv_idx
syntax prior to the sorting stage and hide the latency of entropy
decoding. The compression performance change is about 0.01% level.

Change-Id: I86b34f31f6c99a36ae2780416175cc0bd90ff492

8 years agoSilence some warnings
Debargha Mukherjee [Tue, 4 Oct 2016 21:30:16 +0000 (14:30 -0700)]
Silence some warnings

Change-Id: I8efb64eac3438484e7a77a8a1db198223fc52bfa

8 years agoFix a compiler warning in ext-inter experiment
Debargha Mukherjee [Tue, 4 Oct 2016 20:10:23 +0000 (13:10 -0700)]
Fix a compiler warning in ext-inter experiment

Change-Id: If36417c1384646da57453344b208e7653a4d31e5

8 years agoFix an integer overflow issue in restoration
Debargha Mukherjee [Tue, 4 Oct 2016 18:50:00 +0000 (11:50 -0700)]
Fix an integer overflow issue in restoration

https://bugs.chromium.org/p/webm/issues/detail?id=1306

Change-Id: Icd11d373ff08954121c097728e4c7791791e223f

8 years agoMerge "ext_tx: fix a signed overflow" into nextgenv2
Alex Converse [Tue, 4 Oct 2016 17:24:06 +0000 (17:24 +0000)]
Merge "ext_tx: fix a signed overflow" into nextgenv2

8 years agoMerge "Fix high bitdepth variance overflow on uint32_t" into nextgenv2
Yi Luo [Tue, 4 Oct 2016 17:11:07 +0000 (17:11 +0000)]
Merge "Fix high bitdepth variance overflow on uint32_t" into nextgenv2

8 years agoSimplify file dependencies of SIMD implementation of interpolation filters
Angie Chiang [Mon, 3 Oct 2016 23:39:27 +0000 (16:39 -0700)]
Simplify file dependencies of SIMD implementation of interpolation filters

This is a similar change to following aom CL
https://aomedia-review.googlesource.com/#/c/1961/

Move SIMD related functions from filter.c/h to following files
av1_convolve_ssse3.c
av1_highbd_convolve_filters_sse4.c

Change following c files to header files.
av1_highbd_convolve_filters_sse4.c
av1_convolve_filters_ssse3.c

Change-Id: I41a3cc6b0789e632451aeda82f5eb97a4d78e370

8 years agoFix high bitdepth variance overflow on uint32_t
Yi Luo [Mon, 3 Oct 2016 23:37:00 +0000 (16:37 -0700)]
Fix high bitdepth variance overflow on uint32_t

BUG=webm:1305

Change-Id: I4c56631359e298b99e618c07bcbae9f793c5e2ac

8 years agoFix filter type mismatch warning on Visual Studio
Yi Luo [Wed, 28 Sep 2016 00:12:44 +0000 (17:12 -0700)]
Fix filter type mismatch warning on Visual Studio

- Move filter look-up functions to corresponding optimization modules.

BUG=webm:1296

Change-Id: I87f399609052db2dbc7e5a590afb08b82e3fa89f

8 years agoext_tx: fix a signed overflow
Alex Converse [Mon, 3 Oct 2016 23:16:24 +0000 (16:16 -0700)]
ext_tx: fix a signed overflow

Change-Id: I9a08bc5da1a84c3d4b8fe2d457bb80406c0bc028

8 years agoMerge "Further changes to new-quant tables" into nextgenv2
Debargha Mukherjee [Mon, 3 Oct 2016 21:10:30 +0000 (21:10 +0000)]
Merge "Further changes to new-quant tables" into nextgenv2

8 years agoMerge "decode_with_drops.sh : make sample test work for av1" into nextgenv2
Yaowu Xu [Mon, 3 Oct 2016 19:57:42 +0000 (19:57 +0000)]
Merge "decode_with_drops.sh : make sample test work for av1" into nextgenv2

8 years agoMerge "decode_with_drop.sh: vp8->aom" into nextgenv2
Yaowu Xu [Mon, 3 Oct 2016 19:57:29 +0000 (19:57 +0000)]
Merge "decode_with_drop.sh: vp8->aom" into nextgenv2

8 years agoMerge "decode_to_md5_test: fixes and runs quick encode and checks decode" into nextgenv2
Yaowu Xu [Mon, 3 Oct 2016 19:57:16 +0000 (19:57 +0000)]
Merge "decode_to_md5_test: fixes and runs quick encode and checks decode" into nextgenv2

8 years agoMerge "decode_to_md5.sh: vp8->aom" into nextgenv2
Yaowu Xu [Mon, 3 Oct 2016 19:56:59 +0000 (19:56 +0000)]
Merge "decode_to_md5.sh: vp8->aom" into nextgenv2

8 years agodecode_with_drops.sh : make sample test work for av1
Jim Bankoski [Tue, 14 Jun 2016 13:58:30 +0000 (06:58 -0700)]
decode_with_drops.sh : make sample test work for av1

Change-Id: I4175070840a6561c1cec5f5a50b64e425f3e2926

8 years agodecode_with_drop.sh: vp8->aom
Yaowu Xu [Mon, 3 Oct 2016 17:55:21 +0000 (10:55 -0700)]
decode_with_drop.sh: vp8->aom

Change-Id: I22dacbc2e4933a60ce7151204af9ee253990ca1f

8 years agodecode_to_md5_test: fixes and runs quick encode and checks decode
Jim Bankoski [Mon, 13 Jun 2016 14:16:38 +0000 (07:16 -0700)]
decode_to_md5_test: fixes and runs quick encode and checks decode

This test checks if there's any basic change to the bitstream or default
encoder by running an encode and checking that the md5 from the decode
doesn't change.

Any change to the default encoder or bitstream should be accompanied by
a change to the md5 in this file.

Change-Id: Ibdd5a1442296fd3e946823ec1f43e8ac4e66dd34

8 years agodecode_to_md5.sh: vp8->aom
Yaowu Xu [Mon, 3 Oct 2016 17:54:02 +0000 (10:54 -0700)]
decode_to_md5.sh: vp8->aom

Change-Id: I0dcb0643cf83ee99b63336df851cbca749c11b68

8 years agoSync ref-mv experiment between aom and nextgenv2
Jingning Han [Mon, 3 Oct 2016 14:21:27 +0000 (07:21 -0700)]
Sync ref-mv experiment between aom and nextgenv2

Change-Id: I134d276234b3b8aa7df1ab647892b5d739647f4c

8 years agoFurther changes to new-quant tables
Debargha Mukherjee [Thu, 29 Sep 2016 16:17:36 +0000 (09:17 -0700)]
Further changes to new-quant tables

Refactor to streamline the number of profiles needed, in
preparation for the next steps.

NO change in performance.

Change-Id: I753b89299897857f3c250c316b4cdc4fedcb90e8

8 years agoMerge "Rename aom_write_nmv_probs as av1_write_nmv_probs" into nextgenv2
Jingning Han [Sat, 1 Oct 2016 01:06:09 +0000 (01:06 +0000)]
Merge "Rename aom_write_nmv_probs as av1_write_nmv_probs" into nextgenv2

8 years agoAdd notes for an option
Yaowu Xu [Fri, 30 Sep 2016 22:39:53 +0000 (15:39 -0700)]
Add notes for an option

cherry-picked from aom/master:
2b407394907253be68bc497aa978b0adc298bbf8

Change-Id: Ia7b3bfd68e2c31b21f49a429fecc4d0b701b045f

8 years agoRename AOM_ENC/DEC_BORDER_IN_PIXELS
Yaowu Xu [Fri, 30 Sep 2016 22:07:57 +0000 (15:07 -0700)]
Rename AOM_ENC/DEC_BORDER_IN_PIXELS

Cherry-picked from aom/master:
e2721a65cbfb5b560cd884d60eb17f53539df5f0

Change-Id: I4ade58be91e7bca0cc4f2bed98a43177d7f590a5

8 years agoClean up av1_adapt_mv_probs format
Jingning Han [Fri, 30 Sep 2016 16:42:29 +0000 (09:42 -0700)]
Clean up av1_adapt_mv_probs format

Change-Id: Ib5226d4fe3dcf916fe8954c7240966e3a32eed31

8 years agoSync assign_mv format
Jingning Han [Fri, 30 Sep 2016 16:37:46 +0000 (09:37 -0700)]
Sync assign_mv format

Change-Id: I4fea280d72d7e428f2ab0820fd728997d5a903c9

8 years agoMerge "Set spatial neighbor search resolution 16x16 for block size 64x64" into nextgenv2
Jingning Han [Fri, 30 Sep 2016 17:57:52 +0000 (17:57 +0000)]
Merge "Set spatial neighbor search resolution 16x16 for block size 64x64" into nextgenv2

8 years agoMerge "Search collocated reference block in 16x16 unit" into nextgenv2
Jingning Han [Fri, 30 Sep 2016 17:45:08 +0000 (17:45 +0000)]
Merge "Search collocated reference block in 16x16 unit" into nextgenv2

8 years agoRename aom_write_nmv_probs as av1_write_nmv_probs
Jingning Han [Fri, 30 Sep 2016 17:33:50 +0000 (10:33 -0700)]
Rename aom_write_nmv_probs as av1_write_nmv_probs

Change-Id: Ia33ce4918d3d40eba331f81909f3f1f0f3ab7a58

8 years agoSet spatial neighbor search resolution 16x16 for block size 64x64
Jingning Han [Fri, 30 Sep 2016 16:10:26 +0000 (09:10 -0700)]
Set spatial neighbor search resolution 16x16 for block size 64x64

When the block has width/height above or equal to 64, use 16x16
block search step for reference motion vector search in the non-
immediate rows and columns.

Change-Id: If11ce97a9328b879f30ef87115086aa0cd985a2f

8 years agoSearch collocated reference block in 16x16 unit
Jingning Han [Fri, 30 Sep 2016 16:04:21 +0000 (09:04 -0700)]
Search collocated reference block in 16x16 unit

Use 16x16 block resolution for collocated reference motion vector
search.

Change-Id: I1091b5b178e255eb6cc0b994de360994f7661b79

8 years agoMerge changes I319cb856,Ib009b6b6 into nextgenv2
Alex Converse [Thu, 29 Sep 2016 21:54:27 +0000 (21:54 +0000)]
Merge changes I319cb856,Ib009b6b6 into nextgenv2

* changes:
  Remove multi-entropy coder hacks from the treewriter
  Rename rans_dec_lut to rans_lut

8 years agoMerge "more ref_mv changes from aom/master" into nextgenv2
Jingning Han [Thu, 29 Sep 2016 21:46:55 +0000 (21:46 +0000)]
Merge "more ref_mv changes from aom/master" into nextgenv2

8 years agoMerge "Fix unit test failure for RECT_TX + VAR_TX" into nextgenv2
Yue Chen [Thu, 29 Sep 2016 21:41:10 +0000 (21:41 +0000)]
Merge "Fix unit test failure for RECT_TX + VAR_TX" into nextgenv2

8 years agomore ref_mv changes from aom/master
Yaowu Xu [Tue, 27 Sep 2016 19:55:32 +0000 (12:55 -0700)]
more ref_mv changes from aom/master

Change-Id: I9152f898dfacdf3877ed719f193bb1e0dbee0a1a

8 years agoFix compiler error for GLOBAL_MOTION+WARPED_MOTION
Yue Chen [Thu, 29 Sep 2016 19:12:47 +0000 (12:12 -0700)]
Fix compiler error for GLOBAL_MOTION+WARPED_MOTION

Fix the logical OR computation in .mk file. Otherwise, when both
experiments are on, the output of $(filter... will be two 'yes',
which will cause missing library issue.

Change-Id: I53c44e925dc9ea77c7467217c20e4f1bc7e20fc3

8 years agoMerge "Move warping model estimation functions to COMMON folder" into nextgenv2
Yue Chen [Thu, 29 Sep 2016 18:24:32 +0000 (18:24 +0000)]
Merge "Move warping model estimation functions to COMMON folder" into nextgenv2

8 years agoMerge changes Ideda50a6,Id2bced5f,If423eeb3 into nextgenv2
Alex Converse [Thu, 29 Sep 2016 16:43:12 +0000 (16:43 +0000)]
Merge changes Ideda50a6,Id2bced5f,If423eeb3 into nextgenv2

* changes:
  Port ANS from aom/master 25aaf40
  Refactor bitreader and bitwriter wrapper.
  Migrate aom/master ANS test from d311d02.

8 years agoFix unit test failure for RECT_TX + VAR_TX
Yue Chen [Thu, 29 Sep 2016 00:09:47 +0000 (17:09 -0700)]
Fix unit test failure for RECT_TX + VAR_TX

Disable rect_tx because we only support 4x4 Walsh-Hadamard transform
in lossless mode.

Fixes failure in ./test_libaom --gtest_filter=*Large*ScreencastQ0/1
Configuration: --enable-experimental --enable-var-tx --enable-rect-tx
 --enable-ref-mv --enable-ext_intra --enable-ext_tx --enable-debug
 --disable-optimizations

Change-Id: Ib6b3494c7dcf7182f1cab9b138388d054851a23d

8 years agoMerge "Change non-uniform-quant parameters" into nextgenv2
Debargha Mukherjee [Thu, 29 Sep 2016 16:04:58 +0000 (16:04 +0000)]
Merge "Change non-uniform-quant parameters" into nextgenv2

8 years agoMerge "Update codec name in test enviroment to match decoder" into nextgenv2
Debargha Mukherjee [Thu, 29 Sep 2016 16:04:37 +0000 (16:04 +0000)]
Merge "Update codec name in test enviroment to match decoder" into nextgenv2

8 years agoMerge "mvref_common.c: port refactoring from aom/master" into nextgenv2
Jingning Han [Thu, 29 Sep 2016 15:45:19 +0000 (15:45 +0000)]
Merge "mvref_common.c: port refactoring from aom/master" into nextgenv2

8 years agoMerge "Remove an intermediate variable" into nextgenv2
Jingning Han [Thu, 29 Sep 2016 15:45:10 +0000 (15:45 +0000)]
Merge "Remove an intermediate variable" into nextgenv2

8 years agoRemove multi-entropy coder hacks from the treewriter
Alex Converse [Wed, 28 Sep 2016 22:34:28 +0000 (15:34 -0700)]
Remove multi-entropy coder hacks from the treewriter

Change-Id: I319cb856a16ace343359c2aebc449c1d73bdedee

8 years agoRename rans_dec_lut to rans_lut
Alex Converse [Wed, 28 Sep 2016 22:29:17 +0000 (15:29 -0700)]
Rename rans_dec_lut to rans_lut

It's used in both encoding and decoding. Matches (historical)
implementation in aom/master.

Change-Id: Ib009b6b6023cfe69e99a0a92f3c70f4416fcdb47

8 years agoPort ANS from aom/master 25aaf40
Alex Converse [Wed, 28 Sep 2016 18:33:20 +0000 (11:33 -0700)]
Port ANS from aom/master 25aaf40

Reconciles the following commits from aom/master to nextgenv2:
25aaf40bbc24beeb52de9af7d7624b7d7c6ce9de
87073de5693df70eba1c9b9be2b2732ed3b08fb3

Change-Id: Ideda50a6ec75485cb4fa7437c69f4e58d6a2ca73

8 years agoClang-format ransac.c
Alex Converse [Wed, 28 Sep 2016 19:04:27 +0000 (12:04 -0700)]
Clang-format ransac.c

Change-Id: I1679da4fb8832133ab1bcb396f4bed4e5448e504

8 years agoRefactor bitreader and bitwriter wrapper.
Nathan E. Egge [Thu, 16 Jun 2016 13:00:39 +0000 (09:00 -0400)]
Refactor bitreader and bitwriter wrapper.

Move code for reading and writing literals and reading trees to use
just the aom_read_bit() and aom_write_bit() function calls.

Change-Id: Id2bced5f0125a5558030a813c51c3d79e5701873
(cherry picked from aom/master commit bc1ac15846a200272551699d45457039535e56b2)

8 years agoMigrate aom/master ANS test from d311d02.
Alex Converse [Wed, 28 Sep 2016 17:58:19 +0000 (10:58 -0700)]
Migrate aom/master ANS test from d311d02.

This helps in porting entropy coder changes that happened in aom/master.

Change-Id: If423eeb3da552066cceb88227138ea61d6a20f07
(cherry picked from aom/master commit d311d02da55433d20aad6dd88e0bbb992919988d)

8 years agoUpdate codec name in test enviroment to match decoder
Peter de Rivaz [Tue, 27 Sep 2016 14:24:03 +0000 (15:24 +0100)]
Update codec name in test enviroment to match decoder

The codec name is defined in av1_dx_iface.c
This name needs to match kAV1Name in decode_test_driver.cc.
Otherwise the EndtoEndPSNRTest fails when built with --enable-ext-tile,
(because we need the IsAV1 function to return true.)

Change-Id: I05d5ea5b6fd4bbd49e8bcacd047fb81c27efb3b3

8 years agoChange non-uniform-quant parameters
Debargha Mukherjee [Fri, 23 Sep 2016 17:52:13 +0000 (10:52 -0700)]
Change non-uniform-quant parameters

Also adds hooks to choose different profiles for UV and intra.

Results
lowres: -0.15%
midres: -0.24%

Change-Id: I4af8bc3e9b82b6f8a061dce9f52c89afa6239ae1

8 years agoMove warping model estimation functions to COMMON folder
Yue Chen [Wed, 28 Sep 2016 00:48:38 +0000 (17:48 -0700)]
Move warping model estimation functions to COMMON folder

These functions will be called by both enc and dec in WARPED_MOTION
experiment.

Change-Id: I4b4a20af111b30822760aee8c9451e9ccbb2dd05

8 years agoMerge "Add a TODO for aom_highbd_fdct16x16_1_sse2 tests" into nextgenv2
Yi Luo [Tue, 27 Sep 2016 22:21:53 +0000 (22:21 +0000)]
Merge "Add a TODO for aom_highbd_fdct16x16_1_sse2 tests" into nextgenv2

8 years agomvref_common.c: port refactoring from aom/master
Yaowu Xu [Tue, 27 Sep 2016 17:29:34 +0000 (10:29 -0700)]
mvref_common.c: port refactoring from aom/master

Change-Id: I53cf072f33de957eed6bf6be270218db8ff33af9

8 years agoRemove an intermediate variable
Yaowu Xu [Tue, 27 Sep 2016 17:13:33 +0000 (10:13 -0700)]
Remove an intermediate variable

This commit changes to use function parameter "len" directly.

Change-Id: I072d165aeca59cfbbcf52c9be3c2a91e3191b980

8 years agoMerge "Fix for compile error with RECT_TX without EXT_TX" into nextgenv2
Yue Chen [Tue, 27 Sep 2016 06:46:19 +0000 (06:46 +0000)]
Merge "Fix for compile error with RECT_TX without EXT_TX" into nextgenv2

8 years agoMerge changes I13eed9cb,I3b213790,I7232f9ae into nextgenv2
Alex Converse [Mon, 26 Sep 2016 22:34:57 +0000 (22:34 +0000)]
Merge changes I13eed9cb,I3b213790,I7232f9ae into nextgenv2

* changes:
  Remove VP10 style bitreader and bitwriter wrappers
  Rename av1_ans_test to match aom/master.
  Migrate bitreader to the interface from aom/master

8 years agoMerge "rename pred_mv_s8 to pred_mv" into nextgenv2
Yaowu Xu [Mon, 26 Sep 2016 21:12:04 +0000 (21:12 +0000)]
Merge "rename pred_mv_s8 to pred_mv" into nextgenv2

8 years agoRemove VP10 style bitreader and bitwriter wrappers
Alex Converse [Mon, 26 Sep 2016 20:09:09 +0000 (13:09 -0700)]
Remove VP10 style bitreader and bitwriter wrappers

Change-Id: I13eed9cb6950ea4fbdd586d43b73ac0cc2d78d33

8 years agoRename av1_ans_test to match aom/master.
Alex Converse [Mon, 26 Sep 2016 19:49:00 +0000 (12:49 -0700)]
Rename av1_ans_test to match aom/master.

Change-Id: I3b2137903a87a1f8169ff45e940575b917c26a6a

8 years agoMigrate bitreader to the interface from aom/master
Alex Converse [Fri, 23 Sep 2016 21:21:02 +0000 (14:21 -0700)]
Migrate bitreader to the interface from aom/master

Change-Id: I7232f9ae3d97e730f66e4b80f550192e3ef7230b

8 years agoMerge "Add double precision warping for ransac" into nextgenv2
Sarah Parker [Mon, 26 Sep 2016 19:03:52 +0000 (19:03 +0000)]
Merge "Add double precision warping for ransac" into nextgenv2

8 years agorename pred_mv_s8 to pred_mv
Yaowu Xu [Mon, 26 Sep 2016 16:13:38 +0000 (09:13 -0700)]
rename pred_mv_s8 to pred_mv

Change-Id: Ib1088c3fc80952074e098385fe5eb81742e7dc59

8 years agoMerge "minor format fix" into nextgenv2
Yaowu Xu [Mon, 26 Sep 2016 15:13:05 +0000 (15:13 +0000)]
Merge "minor format fix" into nextgenv2

8 years agoMerge "change to use aomedia copyright notice" into nextgenv2
Yaowu Xu [Mon, 26 Sep 2016 15:12:57 +0000 (15:12 +0000)]
Merge "change to use aomedia copyright notice" into nextgenv2

8 years agoFix for compile error with RECT_TX without EXT_TX
Peter de Rivaz [Mon, 26 Sep 2016 13:20:13 +0000 (14:20 +0100)]
Fix for compile error with RECT_TX without EXT_TX

Change-Id: I2f4e3fc877c03a5bee7f7fd1dc50e6a693697647

8 years agoMerge "enums.h: Combine related #defines into packed enums." into nextgenv2
Alex Converse [Sat, 24 Sep 2016 00:38:53 +0000 (00:38 +0000)]
Merge "enums.h: Combine related #defines into packed enums." into nextgenv2

8 years agominor format fix
Yaowu Xu [Fri, 23 Sep 2016 22:37:46 +0000 (15:37 -0700)]
minor format fix

Change-Id: Ia4a37d43a7110c84cda6ad317aa7f799e00bde82

8 years agochange to use aomedia copyright notice
Yaowu Xu [Fri, 23 Sep 2016 22:37:36 +0000 (15:37 -0700)]
change to use aomedia copyright notice

av1/common/allcommon.h
doc.mk

Change-Id: I7e08c9131ab1c0d7e7854f7e70b90397d041143a

8 years agoAdd double precision warping for ransac
Sarah Parker [Fri, 23 Sep 2016 18:17:27 +0000 (11:17 -0700)]
Add double precision warping for ransac

Change-Id: I32b6e2e6c8454ffb64e4a4ceb87070d175f05fe9

8 years agoMerge "Migrate bitwriter to the interface in aom/master" into nextgenv2
Alex Converse [Fri, 23 Sep 2016 01:18:30 +0000 (01:18 +0000)]
Merge "Migrate bitwriter to the interface in aom/master" into nextgenv2

8 years agoMerge "Fix bug in table for UV tx ize" into nextgenv2
Debargha Mukherjee [Thu, 22 Sep 2016 18:53:51 +0000 (18:53 +0000)]
Merge "Fix bug in table for UV tx ize" into nextgenv2

8 years agoFix bug in table for UV tx ize
Debargha Mukherjee [Thu, 22 Sep 2016 17:10:20 +0000 (10:10 -0700)]
Fix bug in table for UV tx ize

Change-Id: I086b79462b0933cf9dc1101ff71cbc71c7da2738

8 years agoenums.h: Combine related #defines into packed enums.
Urvang Joshi [Tue, 20 Sep 2016 18:36:33 +0000 (11:36 -0700)]
enums.h: Combine related #defines into packed enums.

enums for BLOCK_SIZE, TX_SIZE and PREDICTION_MODE.

Note: These were converted to #defines earlier to save on memory:
https://chromium-review.googlesource.com/#/c/269854/

But we, instead, use attribute 'packed' (see here:
https://gcc.gnu.org/onlinedocs/gcc/Common-Type-Attributes.html#Common-Type-Attributes)
to ensure that these enums use the smallest possible integer type,
and so use smallest memory when used in structs/arrays etc.

Change-Id: If1fc136686b28847109c9f3a06f8728165e7e475

8 years agobitstream_debug: build related cleanup
Angie Chiang [Wed, 21 Sep 2016 23:01:04 +0000 (16:01 -0700)]
bitstream_debug: build related cleanup

Move experimental config from debug_util.c/h to aom_util.mk to avoid
empty object.

Change-Id: Id7978ed6a342262bddaa4df8b53115e750fa1c2c

8 years agoMigrate bitwriter to the interface in aom/master
Alex Converse [Tue, 20 Sep 2016 23:39:01 +0000 (16:39 -0700)]
Migrate bitwriter to the interface in aom/master

Change-Id: I73d46229f0feea43cbe933e51da997833cce032b