Scott LaVarnway [Fri, 1 Aug 2014 13:36:18 +0000 (06:36 -0700)]
Added vp9_sad8x8_neon()
Change-Id: I3be8911121ef9a5f39f6c1a2e28f9e00972e0624
Tom Finegan [Thu, 31 Jul 2014 04:07:43 +0000 (21:07 -0700)]
Merge "vp9_spatial_svc_encoder.sh: Prevent failures when ssvc is unavailable."
James Zern [Thu, 31 Jul 2014 00:49:01 +0000 (17:49 -0700)]
Merge "configure.sh: avoid test -a/-o for compatibility"
Tom Finegan [Wed, 30 Jul 2014 22:27:08 +0000 (15:27 -0700)]
vp9_spatial_svc_encoder.sh: Prevent failures when ssvc is unavailable.
Avoid running the tests when CONFIG_SPATIAL_SVC is 0.
Change-Id: I9b87ed206a9ba7de6827594fba275bd3f60faf2d
Jingning Han [Wed, 30 Jul 2014 21:34:42 +0000 (14:34 -0700)]
Merge "Chessboard pattern partition search"
Pengchong Jin [Wed, 30 Jul 2014 20:35:02 +0000 (13:35 -0700)]
Merge "Early termination after partition NONE is done in RD."
Johann [Wed, 30 Jul 2014 20:24:16 +0000 (13:24 -0700)]
Merge "Remove copy frame neon"
Johann [Wed, 30 Jul 2014 20:23:18 +0000 (13:23 -0700)]
Merge "Remove unused copy src frame function"
Johann [Wed, 30 Jul 2014 20:22:50 +0000 (13:22 -0700)]
Merge "Add test for vp8 copy frame"
James Zern [Wed, 30 Jul 2014 19:03:45 +0000 (12:03 -0700)]
Merge "vp9_cx_iface: defer compressed data buffer alloc"
Pengchong Jin [Wed, 30 Jul 2014 02:49:47 +0000 (19:49 -0700)]
Early termination after partition NONE is done in RD.
This patch allows the encoder to skip the search for partition
SPLIT, HORZ, VERT after the search for partition NONE is done
in RD optimization. It uses the first pass block-wise statistics
to make the decision. If all 16x16 blocks in the current partition
have zero motions and small residues from the frist pass statistics,
and it has small difference variance, further partition search is
skipped.
For speed 2 setting, experiments on general youtube clips show that
the speedup varies from 1% - 10%, 5% on average. On the performance
side in PSNR, derf 0.004%, yt -0.059%, hd -0.106%, stdhd 0.032%.
For hard stdhd clips:
park_joy_1080p, 502952 ms -> 503307 ms (-0.07%)
pedestrian_area_1080p, 227049 ms -> 220531 ms (+3%)
This feature is under the compilation flag CONFIG_FP_MB_STATS and
it is off in current setting.
Change-Id: I554537e9242178263b65ebe14a04f9c221b58bae
Frank Galligan [Wed, 30 Jul 2014 18:12:20 +0000 (11:12 -0700)]
Merge "Neon version of vp9_quantize_fp()"
James Zern [Wed, 30 Jul 2014 18:04:26 +0000 (11:04 -0700)]
Merge changes I3513ae17,Ia0b51165,I461d8b3d
* changes:
make testdata: look for other variants of sha1sum
make testdata: fail if a sha1 fails or is missing
test-data.sha1: add missing checksums
James Zern [Wed, 30 Jul 2014 17:55:11 +0000 (10:55 -0700)]
Merge "decode_to_md5.sh: fix local md5 variable assignment"
James Zern [Wed, 30 Jul 2014 17:50:01 +0000 (10:50 -0700)]
decode_to_md5.sh: fix local md5 variable assignment
extract only the md5 + quote the result
fixes:
test/examples.sh: 47: local: img-176x144-0029.i420: bad variable name
Change-Id: I81c6a83c8a4e792a520fd7046c8eedcbd4af9a0c
Marco Paniconi [Wed, 30 Jul 2014 17:33:35 +0000 (10:33 -0700)]
Merge "Fix to Visual Studio build error."
Jingning Han [Tue, 29 Jul 2014 16:50:03 +0000 (09:50 -0700)]
Chessboard pattern partition search
This commit enables a chessboard pattern constrained partition
search for 720p and above resolutions. The scheme applies stricter
partition search to alternative blocks based on its above/left
neighboring blocks' partition range, as well as that of the
collocated blocks in the previous frame. It is currently turned
on at 16x16 block size level. The chessboard pattern is flipped
per coding frame.
The speed 3 runtime is reduced:
park_joy_1080p, 652832 ms -> 607738 ms (7% speed-up)
pedestrian_area_1080p, 215998 ms -> 200589 ms (8% speed-up)
The compression performance is changed:
hd -0.223%
stdhd -0.295%
Change-Id: I2d4d123ae89f7171562f618febb4d81789575b19
Jingning Han [Wed, 30 Jul 2014 17:05:26 +0000 (10:05 -0700)]
Merge "Clean up max/min allowed block size in rd_pick_partition"
Jingning Han [Wed, 30 Jul 2014 17:05:03 +0000 (10:05 -0700)]
Merge "Use frame index directly in get_chessboard_index"
Jim Bankoski [Wed, 30 Jul 2014 16:56:33 +0000 (09:56 -0700)]
Merge "clear up cfg unused warning in vp9_pick_inter_mode"
Scott LaVarnway [Mon, 28 Jul 2014 12:47:55 +0000 (05:47 -0700)]
Neon version of vp9_quantize_fp()
On a Nexus 7, vpxenc (in realtime mode, speed -12)
reported a performance improvement of ~12.4%
Change-Id: Id29d215acf58bb108489e218a259adf74b4768d7
Marco Paniconi [Wed, 30 Jul 2014 16:13:11 +0000 (09:13 -0700)]
Fix to Visual Studio build error.
signed/unsigned mismatch from commit: 918011.
Change-Id: I6b9bb3618f8e72ea4cc0e6ba8fb9b13676985ed0
Jim Bankoski [Wed, 30 Jul 2014 15:55:22 +0000 (08:55 -0700)]
clear up cfg unused warning in vp9_pick_inter_mode
Change-Id: Iefcf0a25aaf5e44e8e791839aa82d876555025e0
Scott LaVarnway [Tue, 29 Jul 2014 23:47:34 +0000 (16:47 -0700)]
Neon version of vp9_sub_pixel_variance16x16(),
vp9_variance16x16(), and vp9_get16x16var().
On a Nexus 7, vpxenc (in realtime mode, speed -12)
reported a performance improvement of ~16.7%.
Change-Id: Ib163aa99f56e680194aabe00dacdd7f0899a4ecb
James Zern [Wed, 30 Jul 2014 06:25:15 +0000 (23:25 -0700)]
configure.sh: avoid test -a/-o for compatibility
Change-Id: I0bea5cbccd6938fb967299c1d903b60dd2096ccc
James Zern [Wed, 30 Jul 2014 03:43:56 +0000 (20:43 -0700)]
make testdata: look for other variants of sha1sum
it's shasum / sha1 on osx, bsd variants
Change-Id: I3513ae1740adfb05e9b5777d27d445dafbd22413
James Zern [Wed, 30 Jul 2014 03:31:23 +0000 (20:31 -0700)]
make testdata: fail if a sha1 fails or is missing
Change-Id: Ia0b51165d260e283f5338f290a39718e462aed55
James Zern [Wed, 30 Jul 2014 03:31:10 +0000 (20:31 -0700)]
test-data.sha1: add missing checksums
Change-Id: I461d8b3d7cb8f5d787f189dac9032220b089f05f
James Zern [Wed, 30 Jul 2014 03:07:05 +0000 (20:07 -0700)]
Merge "Remove 'local' from configure and configure.sh"
James Zern [Wed, 30 Jul 2014 02:36:11 +0000 (19:36 -0700)]
Merge "vpx_once: implement once() for OS/2"
James Zern [Wed, 30 Jul 2014 02:21:24 +0000 (19:21 -0700)]
Merge "Makefile: add -mstackrealign to CFLAGS on OS/2\r \r This prevents SIGSEGV of test_libvpx.\r \r Change-Id: I788743841469f4141bc8d29b1d1a8683cb00655c\r"
KO Myung-Hun [Tue, 22 Jul 2014 02:09:27 +0000 (11:09 +0900)]
Remove 'local' from configure and configure.sh
'local' is not posix.
Change-Id: I46d7acfa574cbef686cee1b35022f5781485c3fa
KO Myung-Hun [Sat, 26 Jul 2014 05:53:59 +0000 (14:53 +0900)]
vpx_once: implement once() for OS/2
Change-Id: I9f736f299490464bbdbb6cd24ee6f5b46ad45ec6
James Zern [Tue, 22 Jul 2014 01:27:58 +0000 (18:27 -0700)]
vp9_cx_iface: defer compressed data buffer alloc
currently the only way to know if multiple alt-refs are enabled is to
inspect the encoder instance.
this reduces the size of the allocation by 75% when not using multiple
alt-refs
Change-Id: Ie4baa240c2897e64b766c6ad229674884b5a65b6
Pengchong Jin [Tue, 29 Jul 2014 21:27:24 +0000 (14:27 -0700)]
Merge "Remove the redundant index computation in the first pass"
Marco Paniconi [Tue, 29 Jul 2014 18:23:25 +0000 (11:23 -0700)]
Merge "vp8: Add an aggressive denoising mode."
Marco Paniconi [Thu, 24 Jul 2014 19:54:38 +0000 (12:54 -0700)]
vp8: Add an aggressive denoising mode.
Change-Id: Ie4686e1b15af6bcc8d59d585bbeb996f38224522
Jingning Han [Tue, 29 Jul 2014 18:03:00 +0000 (11:03 -0700)]
Clean up max/min allowed block size in rd_pick_partition
This commit replace the repetitive retrieve of max and min allowed
partition from speed_feature with local variables max_size and
min_size.
Change-Id: Ib06f11f16615e4876e4dd5fb6a968c6bf5f7b216
Jingning Han [Fri, 25 Jul 2014 23:43:27 +0000 (16:43 -0700)]
Use frame index directly in get_chessboard_index
The get_chessboard_index() used to call the entire VP9_COMMON
struct pointer to retrieve the chessboard pattern index. This cl
makes it call the frame index directly.
Change-Id: I3cad9d209ea2e77a358085a04fe1ff0ddec5ba03
Scott LaVarnway [Mon, 28 Jul 2014 22:21:10 +0000 (15:21 -0700)]
Added vp9_fdct8x8_neon(), vp9_fdct8x8_1_neon()
On a Nexus 7, vpxenc (in realtime mode, speed -12)
reported a performance improvement of ~3.7%.
Change-Id: I428c72c40df82c6d537955e320a8debf99343004
KO Myung-Hun [Tue, 22 Jul 2014 01:48:36 +0000 (10:48 +0900)]
Makefile: add -mstackrealign to CFLAGS on OS/2
This prevents SIGSEGV of test_libvpx.
Change-Id: I788743841469f4141bc8d29b1d1a8683cb00655c
Pengchong Jin [Mon, 28 Jul 2014 23:04:36 +0000 (16:04 -0700)]
Remove the redundant index computation in the first pass
Remove the redundant index computation when store the first
pass block-wise statistics. Currently, a single byte is
allocated for a 16x16 blocks, and all the frame statistics
saved during the first pass will be kept in memory for use
in the second pass. For a 1920x1080 300-frame clip, it will
take about 2.3 MB memory. This feature is off in current
setting.
Change-Id: I135a95b348ec093d54c6a07e1e8237626909e3bd
levytamar82 [Thu, 24 Jul 2014 07:42:13 +0000 (00:42 -0700)]
Fix bug 805
Remove all the redundant dct functions (dct4x4, dct8x8)
in avx2 except dct32x32 those functions were copied originally from dct_sse2
Change-Id: I742576fbf5175f3ac09f2076976a9247b259323e
Pengchong Jin [Mon, 28 Jul 2014 21:49:05 +0000 (14:49 -0700)]
Merge "Store block-wise statistics obtained in the first pass"
Jim Bankoski [Mon, 28 Jul 2014 16:47:57 +0000 (09:47 -0700)]
Merge "Fix reference frame size restrictions."
Pengchong Jin [Fri, 25 Jul 2014 18:07:44 +0000 (11:07 -0700)]
Store block-wise statistics obtained in the first pass
Change-Id: I9956db2ba2f7d28f484daaf5022d8d1ef5db473c
Jim Bankoski [Mon, 28 Jul 2014 15:37:25 +0000 (08:37 -0700)]
Fix reference frame size restrictions.
The issue was introduced by commit g9f37d14 with adding explicit
restrictions on reference-frame scale factors. The restriction
is checked against aligned-by-8 frame dimensions, not against
original ones. So, for example, frame of 35×35 actually can refer
to frame of 70×70, but the new check won't allow this. It will
compare 35 vs 72 (not 70), so 2x downscale limit will be exceeded.
Change-Id: Ic663693034440f64ac8312cbff9e1e773a921060
Jim Bankoski [Mon, 28 Jul 2014 14:17:27 +0000 (07:17 -0700)]
Merge "This removes a test that fails on 32 bit"
Johann [Fri, 25 Jul 2014 04:52:16 +0000 (21:52 -0700)]
Remove copy frame neon
No performance benefit
Change-Id: I7f0cfa98ac2bb5b823d0972827b45f32cd3f6b4c
Johann [Wed, 16 Jul 2014 15:42:32 +0000 (08:42 -0700)]
Remove unused copy src frame function
Change-Id: I03798e387a656473b6ec004246e2b882c8c6bb84
Johann [Fri, 18 Jul 2014 20:53:53 +0000 (13:53 -0700)]
Add test for vp8 copy frame
Change-Id: Iddac0f263b283480315147c51deec2a2bbd7a1a7
Jingning Han [Fri, 25 Jul 2014 22:57:35 +0000 (15:57 -0700)]
Merge "Fix rd_pick_partition search loop for 4x4 blocks"
Jingning Han [Fri, 25 Jul 2014 22:56:53 +0000 (15:56 -0700)]
Merge "Fix potential ioc issue in vp9_get_prob for 4K above sizes"
Jingning Han [Fri, 25 Jul 2014 22:56:31 +0000 (15:56 -0700)]
Merge "Remove unnecessary conditional assignment"
Minghai Shang [Fri, 25 Jul 2014 20:24:27 +0000 (13:24 -0700)]
Merge "[spatial svc]Fix reference issues"
Joshua Litt [Fri, 25 Jul 2014 20:16:36 +0000 (13:16 -0700)]
Merge "Encode perf test"
Joshua Litt [Mon, 21 Jul 2014 17:57:16 +0000 (10:57 -0700)]
Encode perf test
Change-Id: I61993946fe5cbcd7ebeb6302efd9538bff022e5a
Yunqing Wang [Fri, 25 Jul 2014 17:53:13 +0000 (10:53 -0700)]
Merge "Allocate aligned source in variance test"
Alex Converse [Fri, 25 Jul 2014 17:51:48 +0000 (10:51 -0700)]
Merge "Refactor inter/intra_suberblock_yrd."
Yaowu Xu [Fri, 25 Jul 2014 15:49:39 +0000 (08:49 -0700)]
Merge "Fix allocation of context buffers on frame resize"
Yaowu Xu [Fri, 25 Jul 2014 15:48:48 +0000 (08:48 -0700)]
Merge "Changed validation of reference frame size"
Marco Paniconi [Fri, 25 Jul 2014 15:42:41 +0000 (08:42 -0700)]
Merge "vp8: reduce refresh_max_mbs_perframe with fewer layers."
Tim Kopp [Fri, 25 Jul 2014 14:53:51 +0000 (07:53 -0700)]
Merge "s/CONFIG_DENOISING/CONFIG_VP9_TEMPORAL_DENOISING"
Jingning Han [Fri, 25 Jul 2014 14:08:23 +0000 (07:08 -0700)]
Fix rd_pick_partition search loop for 4x4 blocks
The partition search for 4x4 blocks takes unnecessary steps to
reconstruct pixels and an extra partition type update. This commit
removes such operations. No visible compression/speed difference.
Thanks to Yue (yuec@) for finding this issue.
Change-Id: I3f83824aa3fd3717d63be0b280fa57258939a70a
James Zern [Fri, 25 Jul 2014 06:09:35 +0000 (23:09 -0700)]
Merge "rtcd.pl: check for auto_help availability"
Johann [Thu, 24 Jul 2014 16:32:01 +0000 (09:32 -0700)]
Remove neon version of vp8 extend borders
The code fails the unit test. Speed comparisons to the C are invalid
because the code frequently didn't correctly extend the right and
bottom portions of the frame.
Reduce maximum frame size on ARM devices to avoid OOM
Change-Id: Ia664c86406f0bb8120fd7ad401f32d0bd44994fb
Johann [Wed, 16 Jul 2014 10:47:27 +0000 (03:47 -0700)]
Add test for vp8 extend_borders
Change-Id: I5bfc965019815e631a2300eb29aa065fb2bc94aa
Yunqing Wang [Fri, 25 Jul 2014 00:11:58 +0000 (17:11 -0700)]
Allocate aligned source in variance test
The source buffer is an aligned buffer in VP9. Added the alignment
to make it consistent with libvpx.
Change-Id: I3ebb9d2e8555ed532951da479dd5cbbb8812e02d
Marco Paniconi [Thu, 24 Jul 2014 22:44:29 +0000 (15:44 -0700)]
vp8: reduce refresh_max_mbs_perframe with fewer layers.
Change-Id: Ib39cc0218a1f232c0cde8e99ca3319c17274bd48
Jingning Han [Thu, 24 Jul 2014 22:31:32 +0000 (15:31 -0700)]
Fix potential ioc issue in vp9_get_prob for 4K above sizes
This commit turns on the existing vp9_get_prob function using
64 bit in the intermediate step. It fixes the ioc issue for 4K
above frame sizes (issue 828).
Change-Id: I9f627f3beca2c522f73b38fd2a3e7eefdff01a7c
Jingning Han [Thu, 24 Jul 2014 21:43:06 +0000 (14:43 -0700)]
Remove unnecessary conditional assignment
The assignment of the variable mode_excluded in
vp9_rd_pick_inter_mode_sub8x8 takes redundant conditional jump.
This commit removes it.
Change-Id: Ie195fbe6e54ec2ade7093d562c456a2e93143704
Johann [Wed, 16 Jul 2014 09:30:21 +0000 (02:30 -0700)]
Set and use uv_crop_[width|height]
Ensure consistent border extension by rounding uv_crop_* at image
creation time. Where it was rounded problems could arise with the right
and bottom extensions.
When padding = 32, y_width = 64, and y_crop_width = 63:
(padding + width - crop_width + 1) / 2
32 + 64 - 63 + 1 should equal 32 *but*
32 + 1 + 1 equals 34 giving a right buffer of 17 instead of 16.
By calculating uv_crop_* earlier we round up at the appropriate time and
for the same values:
(y_crop_width + 1) / 2
63 + 1 / 2
64
(padding / 2) + uv_width - uv_crop_width
16 + 16 - 16
16
Change-Id: If866cd1b63444771440edb1432280ac83875969b
Yaowu Xu [Thu, 24 Jul 2014 16:12:46 +0000 (09:12 -0700)]
Changed validation of reference frame size
A previous change, https://gerrit.chromium.org/gerrit/#/c/70632,
introduced a size validation for reference frames to insuare the
input stream is a valid VP9 stream. However, the logic requiring
all reference frames have valid size turned out to be too strict.
In this commit, we modify the validation to require one of the
reference frame has valid dimension. In addition, the decoder
reports error whenever it detects the use of reference frame
with invalid scalig ratio.
Change-Id: If8efc312244087556cfe00f1fcbdff811268ebad
James Zern [Thu, 24 Jul 2014 21:55:19 +0000 (14:55 -0700)]
rtcd.pl: check for auto_help availability
'auto_help' was added to Getopt::Long in 2.33
this isn't strictly necessary as an unrecognized option (--help) will
issue a warning and then print the usage
Change-Id: Ia757553a4e19d22a8eb70768a8866ab1a76a0eec
Adrian Grange [Thu, 24 Jul 2014 20:37:47 +0000 (13:37 -0700)]
Fix allocation of context buffers on frame resize
The patch:
https://gerrit.chromium.org/gerrit/#/c/70814/
changed the test that determined whether the context
frame buffers needed to be reallocated or not.
The code checked for a change in total frame area
to signal the need to reallocate context buffers.
However, the above_context buffer needs to be
resized i:xf only the width of the frame has increased.
Change-Id: Ib89d75651af252908144cf662578d84f16cf30e6
Tim Kopp [Thu, 24 Jul 2014 20:30:00 +0000 (13:30 -0700)]
s/CONFIG_DENOISING/CONFIG_VP9_TEMPORAL_DENOISING
This should prevent confusion with the VP8 CONFIG_TEMPORAL_DENOISING and other
flags.
Change-Id: I1fe4e2977895b7966841d861ab74317ad875b6c8
Alex Converse [Wed, 23 Jul 2014 00:57:40 +0000 (17:57 -0700)]
Refactor inter/intra_suberblock_yrd.
Move txfm_rd_in_plane into choose_tx_size_from_rd and cleanup callers.
Change-Id: I1df2d7dc984802bd5e204cbe881ada0d75fbb3f7
James Zern [Thu, 24 Jul 2014 05:45:51 +0000 (22:45 -0700)]
Merge "x86inc: fix compilation with NASM\r \r Change-Id: I5978921ab1ccad6648a5bde6ad023c3e3430f5f4\r"
KO Myung-Hun [Tue, 22 Jul 2014 01:47:20 +0000 (10:47 +0900)]
x86inc: fix compilation with NASM
Change-Id: I5978921ab1ccad6648a5bde6ad023c3e3430f5f4
KO Myung-Hun [Tue, 22 Jul 2014 11:19:52 +0000 (20:19 +0900)]
Makefile: strip debugging informations only
--strip-unneeded causes SIGSEGV when accessing g_executable_path. So
test_libvpx crashes due to SIGSEGV in ::testing::InitGoogleTest().
OS/2, aout, strip v2.23.2
Change-Id: I2718d082447ee0d9ad0c021b9156c50e1ac085a6
Minghai Shang [Wed, 23 Jul 2014 20:52:34 +0000 (13:52 -0700)]
[spatial svc]Fix reference issues
1. Remove last reference flag for first frame upper layers in one pass mode.
2. Disable refresh golden frame flag for key frames.
Change-Id: I44ac1bd2c795169e4fbfdd078ea79a1d33a204d6
Jingning Han [Wed, 23 Jul 2014 22:07:01 +0000 (15:07 -0700)]
Merge "Remove redundant argument entry in handle_inter_mode"
Jingning Han [Wed, 23 Jul 2014 22:06:52 +0000 (15:06 -0700)]
Merge "Use the chessboard pattern pred search in newmv mode"
Marco Paniconi [Wed, 23 Jul 2014 20:58:45 +0000 (13:58 -0700)]
Merge "Fix clang compiler warning in denoising_neon."
Yaowu Xu [Wed, 23 Jul 2014 19:46:11 +0000 (12:46 -0700)]
Merge "Moved call to vp9_clear_system_state() to a proper location"
Jingning Han [Wed, 23 Jul 2014 19:02:52 +0000 (12:02 -0700)]
Remove redundant argument entry in handle_inter_mode
The value of mode_excluded has been properly set in
vp9_rd_pick_inter_mode_sb(). It is redundant to send it in
handle_inter_mode() and re-set the value again.
Change-Id: I408d4731f2f42e0bcf3ae62e85757717bb410471
Jingning Han [Wed, 23 Jul 2014 18:47:56 +0000 (11:47 -0700)]
Use the chessboard pattern pred search in newmv mode
This commit extends the chessboard pattern prediction filter search.
If the above and left blocks have the same prediction filter type,
the encoder will skip the prediction filter type search and use the
reference one.
The overall chessboard pattern prediction filter type search reduces
speed 3 runtime for hard clips. Experiments on park joy at 1080p
and 15000 kbps show that the runtime goes from 723265 ms to 65832 ms,
i.e., about 10% speed-up. Compression performance wise, it affects
the coding quality by
Change-Id: I880975497c7ad166532e9eea9bf46684d77ff327
derf: -0.326%
yt: -0.257%
hd: -0.241%
stdhd: -0.417%
Jingning Han [Wed, 23 Jul 2014 17:34:54 +0000 (10:34 -0700)]
Merge "Remove redundant num_refs definition"
Marco Paniconi [Wed, 23 Jul 2014 16:42:09 +0000 (09:42 -0700)]
Fix clang compiler warning in denoising_neon.
Issue: https://code.google.com/p/webm/issues/detail?id=829
Change-Id: I580308f8aa4af194b5d8990a9692ebd18db68ee8
Jingning Han [Wed, 23 Jul 2014 16:29:04 +0000 (09:29 -0700)]
Remove redundant num_refs definition
Use is_comp_pred to replace the use case of num_refs.
Change-Id: I4d0c1e14d5f728428a2ae3d293cd2b4a8b2f31d8
Jingning Han [Wed, 23 Jul 2014 16:12:56 +0000 (09:12 -0700)]
Merge "Enable chessboard inter prediction filter type search"
James Zern [Wed, 23 Jul 2014 01:40:31 +0000 (18:40 -0700)]
Merge "frame_size_tests: reduce default 'large' frame size"
Jingning Han [Tue, 22 Jul 2014 23:32:20 +0000 (16:32 -0700)]
Enable chessboard inter prediction filter type search
This commit enables a chessboard pattern prediction filter type
search scheme for rate-distortion optimization speed-up. For the
inferred motion vector modes, the encoder can re-use its above/left
neighbor blocks' prediction filter type and skip a full test on
all possible filter types. Such operation is turned on/off
alternatively in a chessboard manner.
It is turned on in speed 3. For test clip pedestrian 1080p, the
runtime is reduced from 231500 ms -> 221700 ms. The compression
performance is changed:
derf: -0.147%
yt: -0.134%
hd: -0.079%
stdhd: -0.220%
Change-Id: I1912f278e7576c2dc632688e3ad7a257410c605a
James Zern [Tue, 22 Jul 2014 22:48:25 +0000 (15:48 -0700)]
frame_size_tests: reduce default 'large' frame size
VP9FrameSizeTestsLarge exposed an integer overflow in the VP9 encoder,
for now reduce the size to allow the tests to clear and prevent further
regressions.
4096x4096 -> 4096x2160
this should be restored after the bug is fixed:
https://code.google.com/p/webm/issues/detail?id=828
Change-Id: I47fdf0648f1d9a3951f731bbf0b727f85ada4fa1
Marco Paniconi [Tue, 22 Jul 2014 22:25:06 +0000 (15:25 -0700)]
Merge "vp8: Set default denoiser_decision to copy for UV channel."
Adrian Grange [Tue, 22 Jul 2014 22:17:27 +0000 (15:17 -0700)]
Merge "Fix get_frame_type function"
Tim Kopp [Tue, 22 Jul 2014 21:48:42 +0000 (14:48 -0700)]
Merge "VP9 denoiser bugfix in debugging code."
Jingning Han [Tue, 22 Jul 2014 21:21:34 +0000 (14:21 -0700)]
Merge "USE local best_filter variable in handle_inter_mode"
Tim Kopp [Tue, 22 Jul 2014 21:00:11 +0000 (14:00 -0700)]
VP9 denoiser bugfix in debugging code.
When OUTPUT_YUV_DENOISED is enabled the encoder outputs the uncompressed,
denoised video to a separate file. Moved the point at which the file is
written to in order to avoid an extra blank frame at the beginning of the video.
Change-Id: I805f6a912b18b3d9cae59b13c5b8108279439ce3
James Zern [Tue, 22 Jul 2014 20:17:40 +0000 (13:17 -0700)]
Merge "vp9_bitstream.c: cosmetics"