748 Commits

Author SHA1 Message Date
Yaowu Xu
6043bfdb03 Change to use correct variable in for-loop
Change-Id: I252c2f06dfe256d2d33fd1abc42aaadf50273cc8
2016-11-01 09:54:05 -07:00
Jingning Han
ae81f8b2ab Merge "Make txfm_partition_update support rectangular tx_size" into nextgenv2 2016-11-01 16:51:03 +00:00
Jingning Han
2e4f129b42 Merge "Use get_entropy_context() in select_tx_block" into nextgenv2 2016-11-01 16:50:55 +00:00
Yaowu Xu
b386f0b762 Fix a compiler warning with --enable-adapt-scan
Change-Id: I93b191a522ed3e3ca9a363beab4292f64e869610
2016-11-01 09:40:12 -07:00
Yaowu Xu
a5924740a2 Resolve build issue --enable-aom-qm
Change-Id: I9f52ddb53b39cefd2e0ee7144203e1f3958d01aa
2016-11-01 09:32:03 -07:00
Yaowu Xu
fd601e346c Merge "Rename av1_convolve.[hc] to convolve.[hc]" into nextgenv2 2016-11-01 02:25:19 +00:00
Yaowu Xu
0279e91576 Merge "decodemv.c: relocate a function" into nextgenv2 2016-11-01 02:24:51 +00:00
Yaowu Xu
8040aaf9b1 Merge "Fix a bad merge" into nextgenv2 2016-11-01 02:24:38 +00:00
Yaowu Xu
6557ea9fe2 Rename av1_convolve.[hc] to convolve.[hc]
Change-Id: I2047adc4c147201ce0ce3c533fe2861cbff1002c
2016-10-31 17:17:37 -07:00
Jingning Han
7956bd64d7 Make txfm_partition_update support rectangular tx_size
Change-Id: I7d2414a8766141d5109b599271179bc505c772d3
2016-10-31 16:46:30 -07:00
Yaowu Xu
b24e115bc6 decodemv.c: relocate a function
Change-Id: I932dd9c8b43a20d248c00847b19dff88e6eb11be
2016-10-31 16:45:37 -07:00
Jingning Han
ce059e86fb Use get_entropy_context() in select_tx_block
Replace redundant separate handling to retrieve the context value.

Change-Id: I18dde4599cd08ffe33a78694ec377487609de1b1
2016-10-31 16:27:28 -07:00
Yaowu Xu
e86288d2de Fix a bad merge
Change-Id: I4615e8e64d75b1f4277d2221ec94c5d4f1830aa4
2016-10-31 15:56:38 -07:00
Jingning Han
e29fc1daef Merge "Refactor max_blocks_wide/high computation" into nextgenv2 2016-10-31 22:20:16 +00:00
Jingning Han
609f5c63ac Merge "Remove unused tx_size tables" into nextgenv2 2016-10-31 22:20:08 +00:00
Jingning Han
6511159787 Merge "Replace get_tx2d_size() with direct tx_size_2d[] table access" into nextgenv2 2016-10-31 22:19:59 +00:00
Jingning Han
6491b97350 Merge "Support rectangular tx_size in recursive txfm syntax coding" into nextgenv2 2016-10-31 22:19:46 +00:00
Yaowu Xu
292bd65510 Merge changes from topic 'fix_ec_adapt' into nextgenv2
* changes:
  Reverse order of CLPF and dering
  Refactor: read_tx_size_probs()
  Fix compiling issues with --enable-ec-adapt
  Fixes compilation error on Windows/Visual Studio
2016-10-31 22:18:52 +00:00
Steinar Midtskogen
5d56f4d69a Reverse order of CLPF and dering
Low latency:
PSNR YCbCr:     -0.15%      0.11%      0.12%
   PSNRHVS:     -0.25%
      SSIM:     -0.26%
    MSSSIM:     -0.26%
 CIEDE2000:     -0.03%

High latency:
PSNR YCbCr:     -0.18%      0.18%      0.07%
   PSNRHVS:     -0.20%
      SSIM:     -0.21%
    MSSSIM:     -0.21%
 CIEDE2000:     -0.03%

Change-Id: Ieb86d9ba353220de6454bdc15cea825944b6385b
2016-10-31 12:50:11 -07:00
Jingning Han
f65b870e27 Refactor max_blocks_wide/high computation
Factor common codes that show up in multiple places.

Change-Id: I0a72213a151f74bdad926d59f86f0a28d00968fc
2016-10-31 12:39:36 -07:00
Jingning Han
393a60d208 Remove unused tx_size tables
Change-Id: I04367fb68e8fd027f4b9d945f4001e5ab346d098
2016-10-31 12:39:33 -07:00
Jingning Han
7e9929736c Replace get_tx2d_size() with direct tx_size_2d[] table access
Change-Id: I20040cdb5d9fdbf6c50082e5e17b4cfbd1926b13
2016-10-31 12:39:33 -07:00
Jingning Han
42a0fb369d Support rectangular tx_size in recursive txfm syntax coding
Change-Id: I40aa342ffa5b6effe8b124b94783e5f0bd2f2a81
2016-10-31 12:38:07 -07:00
Jingning Han
a98d80fdaa Merge "Use the actual transform block size for loop filter selection" into nextgenv2 2016-10-31 19:09:07 +00:00
Yaowu Xu
efc7535beb Refactor: read_tx_size_probs()
Change-Id: Ibdedd9b8e0b6646b882bc159856ac7c7e7073149
2016-10-31 09:46:42 -07:00
Yaowu Xu
750955b4c1 mvref_common.c: apply clang-format
Change-Id: I755bfb11a57e92e3a68855a53e95efe526f198fd
2016-10-31 09:13:53 -07:00
Yaowu Xu
1aceffa06c Fix compiling issues with --enable-ec-adapt
Change-Id: I52e2c84ce43d36f78806c54b214f9e5b07c5f0f5
2016-10-31 09:13:53 -07:00
Yaowu Xu
09a4265725 Merge "simp-mv-pred integration with ref-mv" into nextgenv2 2016-10-30 19:54:41 +00:00
Yaowu Xu
aa70234e82 Merge "Fix the top-right reference block location" into nextgenv2 2016-10-30 19:54:23 +00:00
Yaowu Xu
ca5e18b750 Merge "Upsample reference frames after size dependent speed features are calculated." into nextgenv2 2016-10-30 19:54:17 +00:00
Yaowu Xu
e7a64cc9ec Merge "Let is_interp_needed always return 1" into nextgenv2 2016-10-30 19:53:40 +00:00
Yaowu Xu
02d33bfeeb Merge "Centralize EC_MULTISYMBOL error checking." into nextgenv2 2016-10-30 19:53:29 +00:00
Yaowu Xu
a2d2a1858e Merge "EC_ADAPT: refactor and fix MinArfFreq unit tests." into nextgenv2 2016-10-30 17:16:19 +00:00
Yaowu Xu
46fcecc395 Merge "EC_ADAPT: send updates for the correct nodes." into nextgenv2 2016-10-30 17:15:40 +00:00
Yaowu Xu
2ae4214618 Merge "Add ec_multisymbol for common daala_ec and rans code" into nextgenv2 2016-10-30 17:15:28 +00:00
Yaowu Xu
cfab447bf8 Merge "Fix ec_adapt+daala_ec test failure" into nextgenv2 2016-10-30 16:22:05 +00:00
Yaowu Xu
eaafb17d41 Merge "Add EC_ADAPT experiment for symbol-adaptive entropy coding." into nextgenv2 2016-10-30 16:21:47 +00:00
Deng
ca8d24d4e1 simp-mv-pred integration with ref-mv
This commit adds simp-mv-pred experiment. The experiment is to work on
top of ref-mv experiment to save memory bandwidth and reduce the size
of line buffer needed in ref-mv experiment.

When compared to ref-mv, this experiment showed:
low-delay BDR gain: 0.03%
High-delay BDR gain: 0.01%
memory/memory bandwidth saving: 40%
local memory/gate count saving: 20%

Change-Id: Ic4006e041fc58ede411da83d0d730c464ebe1749
2016-10-29 22:26:48 -07:00
Jingning Han
ea9cf097c9 Fix the top-right reference block location
This commit fixes the top-right reference block location for block
sizes above 8x8. It improves the coding performance of ref-mv:

lowres 0.08%
midres 0.15%

Thanks to jiafeng@ for finding this issue.

Change-Id: I70750fc7b18bf0126d3e07abc1b63ca5a160193e
2016-10-29 22:26:48 -07:00
Thomas Daede
919bd6abd7 Upsample reference frames after size dependent speed features are calculated.
This prevents a crash if the upsample_refs speed feature is
changed as part of set_size_dependent_vars, when the recode
loop is enabled.

Change-Id: I645e389bfe961879dd2001439a34fde2993868d9
2016-10-29 22:26:48 -07:00
Angie Chiang
a69ce1b314 Let is_interp_needed always return 1
This CL will cause
0.122% PSNR drop on lowres dataset
0.059% PSNR drop on midres dataset

However, it will facilitate hardware implementation.

Change-Id: I0a0713acacbfd571509a721337711c021915dd3c
2016-10-29 22:26:48 -07:00
Nathan E. Egge
baaaa16186 Centralize EC_MULTISYMBOL error checking.
The EC_ADAPT experiment cannot work unless EC_MULTISYMBOL is also
 enabled.
This patch replaces all individual checks with a centralized check in
 both the bitreader.h and bitwriter.h.

Change-Id: I418852d95c5012cc074ed65cd24997e08bc2aadd
2016-10-29 22:26:27 -07:00
Thomas Davies
6519bebf34 EC_ADAPT: refactor and fix MinArfFreq unit tests.
Ensure that cdfs are synced with pdfs after every
forward update.

Change-Id: I5677f78300156c8622f1728d7a343ff6c3a4ea64
2016-10-29 22:21:32 -07:00
Thomas Davies
09ebbfb39f EC_ADAPT: send updates for the correct nodes.
EOB and ZERO token are not currently adapted.

Change-Id: Ie7d657b71fcb157b09e40874fb06a8b7cd95cc70
2016-10-29 17:05:40 -07:00
Alex Converse
aca9feba82 Add ec_multisymbol for common daala_ec and rans code
The new ec_multisymbol experiment supersedes the rans experiment and is
used for multisymbol features that can be backed by either daala_ec or
rans.

This experiment is automatically enabled by ec_adapt and will try to
enable daala_ec or ans (in that order).

Change-Id: Ie75b4002b7a9d7f5f7b4d130c1aacb3dbe97e54f
2016-10-29 17:05:40 -07:00
Alex Converse
bc0a5bacb5 Fix ec_adapt+daala_ec test failure
AV1/AqSegmentTest.TestNoMisMatchAQ1/6 was failing with this experiment
pair.

BUG=aomedia:70

Change-Id: I8c53a043471a87a98a06687afce2e28891592362
2016-10-29 17:05:40 -07:00
Thomas
9ac5508f32 Add EC_ADAPT experiment for symbol-adaptive entropy coding.
This experiment performs symbol-by-symbol statistics
adaptation for non-binary symbols. It requires DAALA_EC or
RANS and ANS to be enabled. The adaptation is currently
based on a simple recursive filter and is taken from
Daala. It has an adaptation rate dependent on alphabet size,
taken from Daala. It applies wherever non-binary symbols
are encoded using Cumulative Probability Functions rather
than trees.

Where symbols are adapted, forward updates in the compressed
header are removed.

In the case of RANS coefficient token values are adapted,
with the exception of the zero token which remains a
binary symbol. In the case of DAALA_EC other values
such as inter and intra modes are adapted as CDFs are
provided in those cases.

The experiment is configured with:

./configure --enable-experimental --enable-daala-ec --enable-ec-adapt

or

./configure --enable-experimental --enable-ans --enable-rans \
    --enable-ec-adapt

EC_ADAPT is not currently compatible with tiles.

BDR results on Objective-1-fast give a small loss:

PSNR YCbCr:      0.51%      0.49%      0.48%
PSNRHVS:      0.50%
SSIM:      0.50%
MSSSIM:      0.51%
CIEDE2000:      0.50%

Change-Id: I3888718e42616f3fd87144de7f125228446ac984
2016-10-29 16:57:48 -07:00
Jingning Han
ee9264c923 Merge "Replace num_4x4_blocks_txsize_loopup table" into nextgenv2 2016-10-29 23:01:26 +00:00
Jingning Han
73d65a49a9 Merge "Refactor rate-distortion optimization of recursive transform partition" into nextgenv2 2016-10-29 23:01:14 +00:00
Jingning Han
9fb1d69e82 Use the actual transform block size for loop filter selection
Parse the recursive transform block partition to fetch the actual
transform size. Use this correct transform size to select the
corresponding loop filter kernel. This slightly improves the coding
performance of recursive transform partition for hdres to 0.14%.

Change-Id: Ibe8bc3fdd0d222a4f1fb8156c56a407bec052b9b
2016-10-29 15:59:55 -07:00