Scott LaVarnway [Wed, 24 Jan 2018 13:58:08 +0000 (13:58 +0000)]
Merge "vp9_quantize_fp_avx2()"
Johann Koenig [Wed, 24 Jan 2018 04:03:23 +0000 (04:03 +0000)]
Merge "update .clang-format for v5.0.0"
James Zern [Wed, 24 Jan 2018 02:12:00 +0000 (02:12 +0000)]
Merge "vpx_codec_enc_init_multi: fix segfault w/vp9"
Jerome Jiang [Wed, 24 Jan 2018 01:12:44 +0000 (01:12 +0000)]
Merge "Fix frame sizes in pkt to support spatial layers."
Shiyou Yin [Wed, 24 Jan 2018 01:06:42 +0000 (01:06 +0000)]
Merge "vp8: [loongson] fix bug of type conflict."
James Zern [Tue, 23 Jan 2018 21:01:23 +0000 (13:01 -0800)]
vpx_codec_enc_init_multi: fix segfault w/vp9
vp9 does not support multi-res encoding, the request should not crash.
+ encode_api_test: unconditionally expose multi-res test
vpx_codec_enc_init_multi should fail independent of
CONFIG_MULTI_RES_ENCODING if not for the same reason.
Change-Id: I44fc58ef70ee4e0e482cb6a5736885f4cb2a8517
Johann [Tue, 23 Jan 2018 21:37:42 +0000 (13:37 -0800)]
update .clang-format for v5.0.0
Change-Id: Id43e8ce9cf3790b728683acc9686e246ccaa90cf
Linfeng Zhang [Tue, 23 Jan 2018 17:47:34 +0000 (09:47 -0800)]
Add vp9_highbd_iht16x16_256_add_sse4_1()
BUG=webm:1413
Change-Id: I8d7eeae1bd219eb848c1a86071046a477f7a91af
Linfeng Zhang [Tue, 23 Jan 2018 18:28:59 +0000 (18:28 +0000)]
Merge "Add "vpx_" prefix to 2 idct x86 functions"
Jerome Jiang [Tue, 23 Jan 2018 17:25:28 +0000 (17:25 +0000)]
Merge "Fix crash invalid params for vp8 multres. Add test."
Linfeng Zhang [Tue, 23 Jan 2018 17:17:38 +0000 (09:17 -0800)]
Add "vpx_" prefix to 2 idct x86 functions
Change-Id: I4f3052d8748e16b06e9155f8daf22f867dfaa7a3
Linfeng Zhang [Tue, 23 Jan 2018 17:04:20 +0000 (17:04 +0000)]
Merge "Add vp9_highbd_iht8x8_64_add_sse4_1()"
Shiyou Yin [Tue, 23 Jan 2018 03:33:25 +0000 (11:33 +0800)]
vp8: [loongson] fix bug of type conflict.
In commit
577d4fa79, int8_t was used to replace char. This will result in a
compilation error, for int8_t was typedefined to signed char, but not char.
Change-Id: I5c9837e01b0b58688a7741f5c9a99a76ca887e4a
Jerome Jiang [Mon, 22 Jan 2018 23:34:59 +0000 (15:34 -0800)]
Fix frame sizes in pkt to support spatial layers.
Add test for svc frame sizes in pkt.
BUG=webm:1485
Change-Id: I983dc229e526d72d22360d7f3016d8358d6beae7
Jerome Jiang [Wed, 17 Jan 2018 19:23:55 +0000 (11:23 -0800)]
Fix crash invalid params for vp8 multres. Add test.
Fix is from the patch in the issue.
Release memories allocated before early exit.
BUG=webm:1482
Change-Id: I64952af99c58241496e03fa55da09fd129a07c77
Johann Koenig [Mon, 22 Jan 2018 19:38:42 +0000 (19:38 +0000)]
Merge changes from topic "clang-format"
* changes:
clang-format v5.0.0 vp9/
remove spurious comments
clang-format v5.0.0 vp8/
clang-format v5.0.0 vpx_dsp/
clang-format v5.0.0 mem_ops.h
clang-format v5.0.0 vpx_util/vpx_atomic.h
clang-format v5.0.0 y4minput.c
clang-format v5.0.0 vpxenc.c
clang-format v5.0.0 examples/
clang-format v5.0.0 test/
Linfeng Zhang [Thu, 18 Jan 2018 23:48:19 +0000 (15:48 -0800)]
Add vp9_highbd_iht8x8_64_add_sse4_1()
BUG=webm:1413
Change-Id: Id9038226902b2d793fc6c17ac81bb104c1a18988
Scott LaVarnway [Fri, 17 Nov 2017 20:34:24 +0000 (12:34 -0800)]
vp9_quantize_fp_avx2()
Started from vp9_quantize_fp_sse2 and tweaked to use avx2.
Change-Id: Ic2da50cc9d73896c7ef2f3cd3db5b1c5d7795b8b
Johann [Fri, 12 Jan 2018 17:39:53 +0000 (09:39 -0800)]
clang-format v5.0.0 vp9/
Remove trailing commas to keep multiple elements on one line.
Add blank lines to prevent comments from being treated as blocks.
clang-format guards for struct with a comment in the middle.
Change-Id: I3bcb8313ae8aaf69179249a13b4087b1272cdbc0
Johann [Fri, 12 Jan 2018 17:14:30 +0000 (09:14 -0800)]
remove spurious comments
These don't appear to make any sense given their context. The
commit log also does not reveal anything.
Discovered due to spurious clang-format indenting:
https://bugs.llvm.org/show_bug.cgi?id=35930
Change-Id: I732a66056ba4c05e3e132a2f236fe10f7a282900
Johann [Fri, 12 Jan 2018 17:12:42 +0000 (09:12 -0800)]
clang-format v5.0.0 vp8/
Allow*OnASingleLine appears to no longer apply to
typedef structs.
Adjust closing parenthesis/opening brace on functions.
Remove trailing commas to keep multiple elements on one line.
Change-Id: I6e535a8ddb15c9b3de8216ce8ddb2a18241af46c
Johann [Fri, 12 Jan 2018 16:42:52 +0000 (08:42 -0800)]
clang-format v5.0.0 vpx_dsp/
Remove comments above #define statements because they get
indented unnecessarily.
https://bugs.llvm.org/show_bug.cgi?id=35930
Add blank lines to prevent comments from being treated as
blocks.
Change-Id: I04dce21b2a10e13b8dc07411a0019c098f6dd705
Marco [Tue, 16 Jan 2018 19:37:40 +0000 (11:37 -0800)]
vp8: Fix to multi-res-encoder for skipping streams.
For the vp8 simulcast/multi-res-encoder:
Add flags to keep track of the disabling/skipping of
streams for the multi-res-encoder. And if the lower spatial
stream is skipped for a given stream, disable the motion
vector reuse for that stream.
Also remove the condition of forcing same frame type
across all streams.
This fix allows for the skipping/disabling of the base
or middle layer streams.
Change-Id: Idfa94b32b6d2256932f6602cde19579b8e50a8bd
Vignesh Venkatasubramanian [Tue, 16 Jan 2018 23:42:43 +0000 (15:42 -0800)]
vp8dx.h: Add macro for skipping loop filter
Without this applications cannot use the vpx_codec_control macro
for VP9_SET_SKIP_LOOP_FILTER. The tests only cover the underscored
version vpx_codec_control_().
Change-Id: I3e6c1888307b76636fdc1a8deae70b5c14238163
Paul Wilkins [Fri, 12 Jan 2018 18:45:28 +0000 (18:45 +0000)]
Merge "Add zoom break out for kf boost loop."
Paul Wilkins [Fri, 12 Jan 2018 18:45:13 +0000 (18:45 +0000)]
Merge "Fix kf detection in some slide shows."
Johann [Thu, 11 Jan 2018 23:20:08 +0000 (15:20 -0800)]
clang-format v5.0.0 mem_ops.h
Remove trailing empty line to keep the comment from being indented.
https://bugs.llvm.org/show_bug.cgi?id=35930
Change-Id: I6c51f7afb4cc47f03a190b4c90e29e4ff1e0c689
Johann [Thu, 11 Jan 2018 23:17:29 +0000 (15:17 -0800)]
clang-format v5.0.0 vpx_util/vpx_atomic.h
Allow*OnASingleLine appears to no longer apply to
typedef structs.
Change-Id: If10db1c30c74ee31dad1a0b1926964e850f15fd2
Johann [Thu, 11 Jan 2018 23:08:25 +0000 (15:08 -0800)]
clang-format v5.0.0 y4minput.c
Remove trailing empty line to keep the comment from being indented.
https://bugs.llvm.org/show_bug.cgi?id=35930
Change-Id: If0f0862623b3fa3ae49e850edbbed52c2b4c6672
Johann [Thu, 11 Jan 2018 23:06:32 +0000 (15:06 -0800)]
clang-format v5.0.0 vpxenc.c
Treat the formatted string as one distinct parameter to fprintf
Change-Id: I62cfd5657c4cefc6b3fa45247ba9f33515a292b1
Johann [Thu, 11 Jan 2018 22:55:06 +0000 (14:55 -0800)]
clang-format v5.0.0 examples/
Attempts to group () expression on their own line.
Change-Id: I404f9dd1a91aaa2100925c90162bcdefbead5ad2
Johann [Thu, 11 Jan 2018 22:05:35 +0000 (14:05 -0800)]
clang-format v5.0.0 test/
Remove trailing commas to keep multiple elements on one line.
Remove trailing empty lines to keep comments from being indented.
https://bugs.llvm.org/show_bug.cgi?id=35930
Change-Id: I0a66dde95f2a304f13cb85a2e9197afca20051e8
Scott LaVarnway [Thu, 11 Jan 2018 23:05:33 +0000 (23:05 +0000)]
Merge "Add quantize_fp_32x32_nz_c()"
Marco Paniconi [Thu, 11 Jan 2018 22:08:49 +0000 (22:08 +0000)]
Merge "vp9: Skip encoding of enhancement layers on the fly."
Johann [Thu, 28 Dec 2017 22:57:02 +0000 (14:57 -0800)]
adopt some clang 5.0.0 formatting
At least the changes that don't conflict with 4.0.1
Change-Id: I9b6a7c14dadc0738cd0f628a10ece90fc7ee89fd
Marco [Wed, 10 Jan 2018 00:35:29 +0000 (16:35 -0800)]
vp9: Skip encoding of enhancement layers on the fly.
For SVC: if an enhancement layer (spatial_layer > 0)
has 0 bandwidth, skip/drop the encoding of the layer.
This allows the application to dynamically disable
higher layers for SVC.
Add flag to signal the skip encoding, this is needed
to modify the packing of the superframe when the top
layer is skipped/dropped.
Also moved some updates (current_video_frame counter and
the last_avg_frame_bandwidth) to the postencode_update_drop_frame().
Added datarate unittest for dynamically going from 3 to 2
and then back to 3 spatial layers.
Change-Id: Idaccdb4aca25ba1d822ed1b4219f94e2e8640d43
Vlad Tsyrklevich [Wed, 10 Jan 2018 23:53:09 +0000 (15:53 -0800)]
[CFI] Remove function pointer casts
Control Flow Integrity [1] indirect call checking verifies that function
pointers only call valid functions with a matching type signature. This
change eliminates some function pointer casts that I missed in my last
CL https://crrev.com/c/780144.
BUG=chromium:776905
[1] https://www.chromium.org/developers/testing/control-flow-integrity
Change-Id: I1c7adbdfffa4fe0b62e993bfb31d06e64b022d66
Paul Wilkins [Wed, 10 Jan 2018 17:33:00 +0000 (17:33 +0000)]
Merge "Fix bug in use of zoom metric as part of arf breakout."
paulwilkins [Tue, 9 Jan 2018 15:51:28 +0000 (15:51 +0000)]
Add zoom break out for kf boost loop.
Adds a breakout threshold to key frame boost loop.
This reduces the boost somewhat in cases where there is a
significant zoom component. In tests most clips no effect
but a sizable gain for some clips like station.
Change-Id: I8b7a4d57f7ce5f4e3faab3f5688f7e4d61679b9a
paulwilkins [Wed, 10 Jan 2018 15:03:21 +0000 (15:03 +0000)]
Fix kf detection in some slide shows.
This fix improves detection of key frames in slide shows.
In particular it helps if the slides are pictures of varying formats
as in a sample provided by yclin@.
This change does not impact any of the clips in our standard tests
but for the example slide show test clip helped global psnr by
several db and resolved a serious visual quality issue.
Change-Id: Iaeeeed55dc0bb50aeacd4996ed660ced06374603
Linfeng Zhang [Fri, 5 Jan 2018 17:57:56 +0000 (09:57 -0800)]
Add vp9_highbd_iht4x4_16_add_sse4_1()
BUG=webm:1413
Change-Id: I14930d0af24370a44ab359de5bba5512eef4e29f
Linfeng Zhang [Fri, 5 Jan 2018 17:40:19 +0000 (09:40 -0800)]
Update dct_test.cc
Make 8-bit functions testing available in high bitdepth.
Change-Id: Ic030c75aa4c6b649c52426abb4bb2122882de0fe
Linfeng Zhang [Mon, 8 Jan 2018 17:15:45 +0000 (17:15 +0000)]
Merge "Update iadst4_sse2()"
Marco Paniconi [Fri, 5 Jan 2018 19:01:31 +0000 (19:01 +0000)]
Merge "vp9-svc: Use eightap_smooth for downsampling at low resol."
Marco [Fri, 5 Jan 2018 17:33:27 +0000 (09:33 -0800)]
vp9-svc: Use eightap_smooth for downsampling at low resol.
Switch from bilinear to eighttap_smooth for frame-level
downsampling at low resolutions (<= 320x240).
avgPSNR/SSIM metrics increase from ~0.5-2% (all clips positive gain),
for 2 and 3 spatial layer SVC, with 3 temporal layers.
Small/negligible increase in encoding time (< 1%).
Change-Id: I758472fc4fddd51d87f13c9d1a1cd4986ef5d41f
paulwilkins [Thu, 4 Jan 2018 16:02:42 +0000 (16:02 +0000)]
Fix bug in use of zoom metric as part of arf breakout.
The in/out (or zoom metrics) in accumulate_frame_motion_stats()
are in effect a % of the blocks that have a motion vector pointing
either towards or away from the center. As such they are already
normalized in terms of image size and the thresholds against which
these are tested should be image size independent.
In practice a zoom either in or out is an indicator for a shorter group
length so the abs value is more important as a breakout clause.
This patch fixes the threshold test. Clips without noticeable zoom show
no effect but some with strong zooms such as "station" show a big
gain (5-10%). Average psnr-hvs gain on hdres set was 0.292%
Change-Id: I4f97a72b0e273e4e844ade15285749c32cd81c1c
Marco [Thu, 4 Jan 2018 22:20:59 +0000 (14:20 -0800)]
vp9: Increase convergence speed of noise estimation.
Increase the recursive average factor from 15/16 to 3/4
to make the noise estimation respond faster.
Small/neglible change on low noise content, but better
denoising for noisy content.
Also encoder speedup of ~2-3% observed on some noisy clips.
Change-Id: I9dd02fe961ca24b411fe4c2732f814bf1e9a7f9f
Linfeng Zhang [Fri, 29 Dec 2017 00:47:57 +0000 (16:47 -0800)]
Update iadst4_sse2()
Change-Id: I21ff81df0d6898170a3b80b3b5220f9f3ac7f4e8
Scott LaVarnway [Fri, 22 Dec 2017 18:05:33 +0000 (10:05 -0800)]
Add quantize_fp_32x32_nz_c()
This c version uses the shortcuts found in the
vp9_quantize_fp_32x32_ssse3 function.
Change-Id: I2e983adb00064e070b7f2b1ac088cc58cf778137
Scott LaVarnway [Mon, 18 Dec 2017 14:31:46 +0000 (06:31 -0800)]
Add vp9_quantize_fp_nz_c() -- 2
This c version uses the shortcuts found in the x86
vp9_quantize_fp functions.
The test was updated to use the correct quant/round range.
Change-Id: Ie5871f710d9eb39047d8d9f48b907c0633e1f830
James Zern [Thu, 21 Dec 2017 23:02:32 +0000 (23:02 +0000)]
Merge "vp9_quantize_ssse3_x86_64: fix out of bounds write"
Ralph Giles [Thu, 21 Dec 2017 21:48:55 +0000 (13:48 -0800)]
Don't force inlining for msvc targets.
INLINE is defined as __forceinline for vs* configs, but is the
normal, compiler-discretion inline for gcc/clang configs. This
makes many functions very large when building for windows targets,
much larger than they are elsewhere.
Use '__inline' as a consistent definition to get consistent function
sizes. Although Visual Studio documentation says that 'inline' is
only available in C+ code. This is probably incorrect, since Visual
Studio 2017 accepts C99 'inline' even when passed /TC. Nevertheless,
this commit uses the recommended '__inline' for consistency.
Thanks to David Major for the diagnosis.
Change-Id: Ib0b31a3afcea77822c84fe3c6cd452add66d825a
James Zern [Thu, 21 Dec 2017 21:51:40 +0000 (16:51 -0500)]
vp9_quantize_ssse3_x86_64: fix out of bounds write
eob is a pointer to a uint16_t. previously the code would store 64-bits
causing a crash or test failure with the right stack layout.
Change-Id: Ibd653baf323db114f2444951b9d8b00c596bf15a
James Zern [Thu, 21 Dec 2017 14:03:20 +0000 (06:03 -0800)]
Revert "Add vp9_quantize_fp_nz_c()"
This reverts commit
86842855d30d6ca6befdcf5108003e027d90daa9.
SSSE3/VP9QuantizeTest.EOBCheck/1 fails on Mac and the build breaks under
visual studio due to a #if within another macro.
Change-Id: I475095a04aafcc714fade2b24e4df7b682be2cd1
Scott LaVarnway [Wed, 20 Dec 2017 23:15:11 +0000 (23:15 +0000)]
Merge "Add vp9_quantize_fp_nz_c()"
James Zern [Wed, 20 Dec 2017 20:22:34 +0000 (20:22 +0000)]
Merge "lpf_test: correct threshold ranges"
Marco [Fri, 15 Dec 2017 00:35:33 +0000 (16:35 -0800)]
vp9-svc: Add layer bitrate targeting to SVC datarate tests.
Modify and update the SVC datarate unittests to verify the
rate targeting for each spatial-temporal layer.
The current tests were only verifying the rate targeting
of the full SVC stream, not individual layers.
Also re-enabled a test that was disabled.
This is a stronger verification of the layered rate control
for SVC for 1 pass CBR encoding.
Added PostEncodeFrameHook, needed to get the layer_id and
update the layer buffer level.
Change-Id: I9fd54ad474686b20a6de3250d587e2cec194a56f
Scott LaVarnway [Mon, 18 Dec 2017 14:31:46 +0000 (06:31 -0800)]
Add vp9_quantize_fp_nz_c()
This c version uses the shortcuts found in the x86
vp9_quantize_fp functions.
The test was updated to use the correct quant/round range.
Change-Id: I5d19f8af2fddda8e50910249eafb740acb29415b
Marco [Fri, 8 Dec 2017 01:24:41 +0000 (17:24 -0800)]
vp9: Reset buffer level on large bitrate changes.
For a large change in the target avg_frame_bandwidth,
via the update in change_config()), reset the buffer_level
to optimal_level.
This fix prevents possible frame drops, where for example,
encoder suddenly goes from lower to higher bitrate.
Change-Id: I2f844c41d04c01240e85f574e59d2b9075c7eb6d
James Zern [Tue, 19 Dec 2017 07:17:45 +0000 (23:17 -0800)]
lpf_test: correct threshold ranges
the random number generator creates values from [0, range) add 1 to all
and make hev more realistic by mirroring its calculation of level >> 4,
i.e., [0, 3]
Change-Id: Ic19be5d7ba668deb17c96f143b739116a4b5d21c
Shiyou Yin [Fri, 15 Dec 2017 09:06:47 +0000 (17:06 +0800)]
vp8: [loongson] optimize loopfilter v2.
Optimize function vp8_mbloop_filter_vertical_edge_mmi and
function vp8_mbloop_filter_horizontal_edge_mmi.
Make full use of memory loading delay slot and reduce unnecessary
instructions.
Change-Id: I61da2c3a44c06044225461f46bf487d83cba6c16
Shiyou Yin [Fri, 15 Dec 2017 00:53:21 +0000 (00:53 +0000)]
Merge "vp8: [loongson] optimize sixtab predict v2."
Johann Koenig [Thu, 14 Dec 2017 23:44:30 +0000 (23:44 +0000)]
Merge "add copyright to rtcd files"
Johann Koenig [Thu, 14 Dec 2017 23:44:04 +0000 (23:44 +0000)]
Merge "mark generated version header"
Johann [Thu, 14 Dec 2017 20:05:56 +0000 (12:05 -0800)]
add copyright to rtcd files
Allows them to pass the license check in chromium.
BUG=chromium:98319
Change-Id: Iefc1706152a549d8c4ae774c917596bf1c9492d8
Johann Koenig [Thu, 14 Dec 2017 21:19:59 +0000 (21:19 +0000)]
Merge "remove unused tools"
Johann Koenig [Thu, 14 Dec 2017 21:19:47 +0000 (21:19 +0000)]
Merge "fix typo in boilerplate"
Johann [Thu, 14 Dec 2017 20:27:11 +0000 (12:27 -0800)]
remove unused tools
all_builds.py has been more or less replaced by Jenkins.
author_first_release.sh is unused.
ftfy.sh has been obviated by having the whole tree clang-format clean.
Change-Id: I741315ad9042e6e901f07410e93f28371db703b2
Johann [Thu, 14 Dec 2017 19:57:49 +0000 (11:57 -0800)]
mark generated version header
Allows it to pass the license check in chromium.
BUG=chromium:98319
Change-Id: I5ba9c8c81ab9eb4168df09db9d2eab846e99e981
Johann [Thu, 14 Dec 2017 19:54:16 +0000 (11:54 -0800)]
fix typo in boilerplate
The extra 'e' was causing the chromium license check to flag this file.
BUG=chromium:98319
Change-Id: Ic875ba66370298bf998438d14ff5f7e760293706
Johann [Thu, 14 Dec 2017 19:48:46 +0000 (11:48 -0800)]
mark generated rtcd headers
Allows them to pass the license check in chromium.
BUG=chromium:98319
Change-Id: Ib37bf45bdac8cf1edc62037dea17b734a5e37fa7
Shiyou Yin [Thu, 14 Dec 2017 08:29:58 +0000 (16:29 +0800)]
vp8: [loongson] optimize sixtab predict v2.
1. Delete unnecessary zero setting process.
2. Optimize the method of calculating SSE in vpx_varianceWxH.
Change-Id: I8bab801416e7f4958c28c6d080e3cf785a50f82b
Marco [Wed, 13 Dec 2017 21:34:44 +0000 (13:34 -0800)]
vp9: Update to SVC datarate tests.
With recent fixes to rate control for SVC the
buffer underrun in the tests does not happen,
so comment and TODO can be removed.
Also, in some of these SVC tests, replace the HD clip
with the corresponding VGA clip, which has > 400 frames.
For the (niklas) HD clip: it has only 60 frames but the
test was running up to 300 frames. Fixed it to 60 frames.
Keep some tests with the HD clip, needed for the 4 thread
and 5 level scaling test.
Change-Id: I0a2356a908e8b2271c7a422eb8b15c0d56eec968
Marco Paniconi [Wed, 13 Dec 2017 21:03:40 +0000 (21:03 +0000)]
Merge "vp9: Reset rc flags on some configuration changes."
Marco [Wed, 13 Dec 2017 19:29:26 +0000 (11:29 -0800)]
vp9: Cleanup/remove TODO comment.
Change-Id: I2bd43e996909ad688b7e00b81ee19a5fc4df460b
Marco [Fri, 8 Dec 2017 01:24:41 +0000 (17:24 -0800)]
vp9: Reset rc flags on some configuration changes.
For large dynamic changes in target avg_frame_bandwidth, or
a change in resolution, via the update in change_config()),
reset the under/overshoot flags (rc_1_frame, rc_2_frame)
to prevent constraining the QP for the first few frames
following the change.
For SVC use the spatial stream avg_frame_bandwidth in
reset condition.
For the avg_frame_bandwidth condition, use fairly large
threshold (~50%) for now in reset.
This allows for better/faster QP response if, for example,
application dynamically changes bitrate by large amount.
Change-Id: Ib6e3761732d956949d79c9247e50dba744a535c0
Paul Wilkins [Tue, 12 Dec 2017 11:56:10 +0000 (11:56 +0000)]
Merge "Bug fix for second reference stats."
Jerome Jiang [Tue, 12 Dec 2017 05:27:11 +0000 (05:27 +0000)]
Merge "vp9 svc: Allow denoising next to highest resolution."
Jerome Jiang [Fri, 8 Dec 2017 18:38:18 +0000 (10:38 -0800)]
vp9 svc: Allow denoising next to highest resolution.
Denoise 2 spatial layes at most.
Add noise sensitivity level 2 for vp9 such that applications can control
whether to denoise the second highest spatial layer.
Add tests to cover this case.
Change-Id: Ic327d14b29adeba3f0dae547629f43b98d22997f
Jerome Jiang [Mon, 11 Dec 2017 18:27:17 +0000 (18:27 +0000)]
Merge "Fix build warnings for gcc 6.3"
paulwilkins [Mon, 4 Dec 2017 13:41:44 +0000 (13:41 +0000)]
Bug fix for second reference stats.
Immediately following a key frame the trailing second reference
error in the first pass stats will be based on a reference frame from
the prior key frame group and will thus usually be much larger.
This fix eliminates that effect (which typically triggers a short arf
group immediately after a key frame). It also changes the accounting
for the first frame in each new arf group.
This change gives large gains on a couple of clips that contain mid
sequence key frames (e.g. 6% on 1080P tennis). Overall there was
a net gain in PSNR and PSNR-HVS ~(0.05- 0.4%) and mixed results for
SSIM (+/- 0.2%).
Change-Id: I8e00538ac2c0b5c2e7e637903cac329ce5c2a375
Jerome Jiang [Wed, 6 Dec 2017 00:11:05 +0000 (16:11 -0800)]
Fix build warnings for gcc 6.3
Clean up some alias.
BUG=webm:1465
Change-Id: I99e186162db9f9e15375fef01564692434eda619
Jerome Jiang [Wed, 6 Dec 2017 22:37:15 +0000 (22:37 +0000)]
Merge "Add frame width & height to frame pkt. Add test."
Jerome Jiang [Wed, 29 Nov 2017 23:06:35 +0000 (15:06 -0800)]
Add frame width & height to frame pkt. Add test.
Used to return correct frame width and height when dynamic resizing happens.
BUG=webm:1474
Change-Id: Ia2043f7e1635b3821848a67b9b134f47f14b0f3a
Marco [Wed, 6 Dec 2017 19:21:29 +0000 (11:21 -0800)]
vp9-svc: Set downsampling filter for VGA layer.
Downsampling filter for SVC was set to subsample (phase 0)
for HD -> VGA, and bilinear averaging (phase 8) for VGA -> QVGA.
This change makes it bilinear averaging for HD -> VGA.
Given the recent commit 9f9d4f8, quality is improved with
this change: avgPSNR/SSIM up ~1-3% on HD clips in RTC set.
Speed decrease of ~1% for 3 layer SVC.
Change-Id: If834a320e372b8b922a6bf7cab4227703b1beae6
Marco Paniconi [Wed, 6 Dec 2017 19:18:51 +0000 (19:18 +0000)]
Merge "vp9: Nonrd-pickmode: move some early exits up."
Hui Su [Wed, 6 Dec 2017 18:46:19 +0000 (18:46 +0000)]
Merge "Add max luma picture width/height constraint in VP9 level"
Marco [Wed, 6 Dec 2017 18:11:54 +0000 (10:11 -0800)]
vp9: Nonrd-pickmode: move some early exits up.
Move the early exit checks on usable_ref_frame and
skip_ref_find_pref up before the check on flag_svc_subpel.
The code under flag_svc_subpel requires frame_mv to be set
for the golden/spatial reference, which is only set if the
both those exits don't pass.
No change in behavior.
Change-Id: Id304276c745eeb389ff85fa2dcf510d5976bc413
Marco [Fri, 1 Dec 2017 23:20:41 +0000 (15:20 -0800)]
vp9-svc: Allow for nonzero motion on spatial reference.
For nonrd pickmode on a given spatial layer, the spatial
(golden) reference was always only using zeromv for prediction.
In this patch if the downsampling filter used for generating
the lower spatial layer is an averaging filter (nonzero phase),
we allow for subpel motion on the spatial (golden) reference to
compensate for the shift. This is done by forcing the testing of
nonzero motion mode to compensate for spatial downsampling shift.
Improvement for cases where the downsampling is averaging filter.
In the current code this is only done for generating
resolutions <= QVGA.
Improvement for avgPSNR/SSIM on RTC set for speed 7: ~1.2%.
Gain is larger (~2-3%) for VGA clips with 2 spatial layers.
~1% speed slowdown for 3 layer SVC on mac.
Change-Id: I9ec4fa20a38947934fc650594596c25280c3b289
Shiyou Yin [Mon, 4 Dec 2017 01:30:06 +0000 (01:30 +0000)]
Merge "vpx_dsp: [loongson] optimize variance v2."
Hui Su [Wed, 29 Nov 2017 01:38:47 +0000 (17:38 -0800)]
Add max luma picture width/height constraint in VP9 level
BUG=b/
65412009
Change-Id: I9e1478dcbd2ef9e97f5f8fb5a1c733b5f5cdf396
Johann [Fri, 1 Dec 2017 23:03:49 +0000 (15:03 -0800)]
filter out asm includes
Don't add include files to the archive. Avoids build failures for
Windows such as:
the input file 'libvpx_g.a(x86_abi_support.asm.o)' has no sections
Change-Id: If9c8e70c0ec913b7ad7dd6a08d4fa19011114ad2
Johann [Fri, 1 Dec 2017 22:31:04 +0000 (14:31 -0800)]
explicitly label .text sections
nasm should infer .text but does not for windows:
https://bugzilla.nasm.us/show_bug.cgi?id=3392451
Change-Id: Ib195465e5f33405f5ff61c4cf88aa2a72640cacb
Johann [Fri, 1 Dec 2017 20:46:37 +0000 (12:46 -0800)]
nasm defaults to -Ox
No need to specify default behaviour. The original change introducing nasm:
https://chromium.googlesource.com/webm/libvpx/+/
7be093ea4d50c8d38438f88cb9fa817c1c9de8dd
mentions requiring 2.0.9, which was the first release to default to this behaviour:
http://www.nasm.us/doc/nasmdoc2.html
"The -Ox mode is recommended for most uses, and is the default since NASM 2.09."
Change-Id: Ia914c4deede5aa447277b5189bb4fcf7e54c338d
Johann Koenig [Fri, 1 Dec 2017 22:07:03 +0000 (22:07 +0000)]
Merge "pass 'win64' instead of 'x64' to the assembler"
Johann [Fri, 1 Dec 2017 18:58:44 +0000 (10:58 -0800)]
pass 'win64' instead of 'x64' to the assembler
nasm does not accept x64
yasm has accepted (and appears to prefer) win64 at least as far back as
1.0.0:
http://yasm.tortall.net/releases/Release1.0.0.html
Change-Id: Ied881b1df0570da256b1bd7e131e7817e47f768f
Marco [Fri, 1 Dec 2017 18:30:45 +0000 (10:30 -0800)]
vp9-svc: Set num_inter_modes in non-rd pickmode.
Set num_inter_modes based on ref_mode_set_svc, which is
smaller set than ref_mode_set (which may use alt-ref).
No change in behavior.
Change-Id: I31169bb09028db230552c6fca0a86959d1ade692
Shiyou Yin [Fri, 1 Dec 2017 05:44:48 +0000 (13:44 +0800)]
vpx_dsp: [loongson] optimize variance v2.
1. Delete unnecessary zero setting process.
2. Optimize the method of calculating SSE in vpx_varianceWxH.
Change-Id: I58890c6a2ed1543379acb48e03e620c144f6515f
Kaustubh Raste [Fri, 1 Dec 2017 01:24:25 +0000 (01:24 +0000)]
Merge "mips msa optimize vpx_scaled_2d function"
Marco Paniconi [Fri, 1 Dec 2017 00:23:52 +0000 (00:23 +0000)]
Merge "Nonrd-pickmode: avoid duplicate computation of UV predictor."