Jingning Han [Mon, 30 Jun 2014 19:20:30 +0000 (12:20 -0700)]
Merge "Remove unused set_mode_info function"
hkuang [Mon, 30 Jun 2014 18:32:11 +0000 (11:32 -0700)]
Merge "Fix a bug in VP9Worker which leads to unit test hang."
Jingning Han [Mon, 30 Jun 2014 16:49:48 +0000 (09:49 -0700)]
Remove unused set_mode_info function
When the frame is intra coded only, the encoder takes the RD
coding flow. Hence the function set_mode_info is not practically
in use. This commit removes it and the associated conditional
branches.
Change-Id: I1e42659ceb55b771ba712d1cdecacb446aa6460d
hkuang [Sat, 28 Jun 2014 17:19:35 +0000 (10:19 -0700)]
Fix a bug in VP9Worker which leads to unit test hang.
This fixes the hang in VP9/InvalidFileTest.ReturnCode/3
due to worker->had_error has not been reset after getting
error.
Change-Id: Ia3608225094758a2bd88f6ae4dd9dfd93bbaad27
Yunqing Wang [Sat, 28 Jun 2014 00:44:32 +0000 (17:44 -0700)]
Enable encode breakout in real time
For real time speed 7, once encode breakout is on(i.e. encoding
setting --static-thresh=1), a proper encode breakout threshold
is set to speed up the encoder.
Set --static-thresh=1, RTC set borg test showed a slight overall
psnr loss of 0.162%, but ssim gain of 0.287%. The average speedup
on RTC set is 6%, and for some clips, the speedup can be 10+%.
Change-Id: Id522d9ce779ff7c699936d13d0c47083de4afb85
Yunqing Wang [Thu, 1 May 2014 22:14:39 +0000 (15:14 -0700)]
Decide the partitioning threshold from the variance histogram
Before encoding a frame, calculate and store each 16x16 block's
variance of source difference between last and current frame.
Find partitioning threshold T for the frame from its variance
histogram, and then use T to make partition decisions.
Comparing with fixed 16x16 partitioning, rtc set test showed an
overall psnr gain of 3.242%, and ssim gain of 3.751%. The best
psnr gain is 8.653%.
The overall encoding speed didn't change much. It got faster for
some clips(for example, 12% speedup for vidyo1), and a little
slower for others.
Also, a minor modification was made in datarate unit test.
Change-Id: Ie290743aa3814e83607b93831b667a2a49d0932c
Jim Bankoski [Mon, 30 Jun 2014 16:14:23 +0000 (09:14 -0700)]
Merge "initialize bit buffer structure to avoid warning error"
Marco Paniconi [Mon, 30 Jun 2014 15:30:54 +0000 (08:30 -0700)]
Merge "vp8 denoiser fix: free memory of denoiser->denoiser_state."
Tim Kopp [Mon, 30 Jun 2014 15:29:32 +0000 (08:29 -0700)]
Merge "Implemented motion compensation for VP9 denoiser"
Jim Bankoski [Mon, 30 Jun 2014 15:08:53 +0000 (08:08 -0700)]
Merge "silence unused parm warning for worker thread in loop filter"
Jim Bankoski [Mon, 30 Jun 2014 15:08:41 +0000 (08:08 -0700)]
Merge "remove unused parms from rd_pick_inter_mode_sb_seg_skip"
Jim Bankoski [Mon, 30 Jun 2014 15:05:15 +0000 (08:05 -0700)]
initialize bit buffer structure to avoid warning error
Change-Id: I38bb2801ad3f059d5e2eb6513eec92397c67abcd
Marco Paniconi [Sun, 29 Jun 2014 23:43:56 +0000 (16:43 -0700)]
vp8 denoiser fix: free memory of denoiser->denoiser_state.
Change-Id: I86451ddfa90d871e46e5b4c97794dbef65293c6c
Jim Bankoski [Sun, 29 Jun 2014 16:30:59 +0000 (09:30 -0700)]
silence unused parm warning for worker thread in loop filter
Change-Id: Id51468f99f8970b8795ce2d254344f4b8d7817d0
Jim Bankoski [Sun, 29 Jun 2014 16:23:21 +0000 (09:23 -0700)]
remove unused parms from rd_pick_inter_mode_sb_seg_skip
Change-Id: I7f989d197444d166133ad91eb23ac1033109f58d
James Zern [Sat, 28 Jun 2014 17:21:38 +0000 (10:21 -0700)]
Merge "vp9: disable postproc buffer alloc when unnecessary"
James Zern [Sat, 28 Jun 2014 05:19:16 +0000 (22:19 -0700)]
Merge "Add a test using VP9E_SET_LOSSLESS"
Alex Converse [Fri, 27 Jun 2014 23:09:06 +0000 (16:09 -0700)]
Add a test using VP9E_SET_LOSSLESS
Change-Id: Ia9ac079d65fbd0a134e8d69c1f4ce171fe94bf64
James Zern [Fri, 27 Jun 2014 22:33:39 +0000 (15:33 -0700)]
configure: defer x32 check until post toolchain setup
x32 would never be detected unless an explicit CC= was used
fixes:
...: -E: command not found
Change-Id: I465fd7b0979f782053bdda240ad68edbca5658e5
James Zern [Fri, 27 Jun 2014 22:08:07 +0000 (15:08 -0700)]
vp9: disable postproc buffer alloc when unnecessary
the buffer is only used in encoding and only when
CONFIG_INTERNAL_STATS or CONFIG_VP9_POSTPROC is enabled.
a future change should decouple this from the frame buffer allocation
and make it conditional based on runtime flags when the above config
options are enabled.
reduces decode heap usage by at least 12%
Change-Id: Id0b97620d4936afefa538d3aadf32106743d9caf
James Zern [Sat, 28 Jun 2014 03:53:57 +0000 (20:53 -0700)]
Merge "Revert "Revert "Revert 3 patches from Hangyu to get Chrome to build:"""
James Zern [Sat, 28 Jun 2014 00:31:01 +0000 (17:31 -0700)]
Revert "Revert "Revert 3 patches from Hangyu to get Chrome to build:""
This reverts commit
b336356198b8ada50fbb59f04f11cefceaf5ff95.
This causes a hang in:
VP9/InvalidFileTest.ReturnCode/3
the change to test/user_priv_test.cc remains with a minor update
Change-Id: I4a8a272ca37ea329b0f413f0b1cd827a238bd9fd
Yaowu Xu [Fri, 27 Jun 2014 23:47:16 +0000 (16:47 -0700)]
Merge "Allow encoder to set lpf level to 0"
Yaowu Xu [Fri, 27 Jun 2014 23:47:06 +0000 (16:47 -0700)]
Merge "Added a new speed 7 in rt mode"
Alex Converse [Fri, 27 Jun 2014 23:32:52 +0000 (16:32 -0700)]
Merge "Add a CPU speed test with screen content."
James Zern [Fri, 27 Jun 2014 23:20:38 +0000 (16:20 -0700)]
Merge "iosbuild.sh Add arm64 support to VPX.framework."
Tom Finegan [Fri, 27 Jun 2014 23:17:35 +0000 (16:17 -0700)]
Merge "configure.sh: Tweak default of use_x86inc for 64-bit, non-pic, and Darwin."
Tim Kopp [Fri, 27 Jun 2014 22:42:14 +0000 (15:42 -0700)]
Merge "fix: Only do spatial SVC when there are > 1 layers"
Tim Kopp [Fri, 27 Jun 2014 22:41:51 +0000 (15:41 -0700)]
Merge "VP9 denoiser: implemented update_frame_stats()"
Yaowu Xu [Fri, 27 Jun 2014 22:23:02 +0000 (15:23 -0700)]
Allow encoder to set lpf level to 0
As a way to speed-up rtc encoding at speed 7.
Change-Id: Ie36a010392cf7b741dc130df21a4e733622a75b7
Yaowu Xu [Fri, 27 Jun 2014 17:05:24 +0000 (10:05 -0700)]
Added a new speed 7 in rt mode
To experiment with different speed/quality compromises.
Change-Id: Ia9d4b85243554d620498a327da37c356e752b07f
Minghai Shang [Fri, 27 Jun 2014 18:48:22 +0000 (11:48 -0700)]
Merge "[spatial svc] Remove encoding modes since we only need one mode at this time."
Minghai Shang [Fri, 27 Jun 2014 18:48:16 +0000 (11:48 -0700)]
Merge "[spatial svc]Remove key frame quantizer settings since key frame is decided by rate control"
Alex Converse [Fri, 27 Jun 2014 18:09:14 +0000 (11:09 -0700)]
Merge "Verify that the ouput of q0 is lossless in cpu speed test."
Jim Bankoski [Fri, 27 Jun 2014 18:05:21 +0000 (11:05 -0700)]
Merge "Better validation of invalid files"
Alex Converse [Thu, 26 Jun 2014 23:21:14 +0000 (16:21 -0700)]
Add a CPU speed test with screen content.
Encoding screen content exercises various fast skip paths that are
missed by natural video content.
Change-Id: Ie359884ef9be89cbe5dda6d82f1f79360604a090
Alex Converse [Thu, 26 Jun 2014 23:19:40 +0000 (16:19 -0700)]
Verify that the ouput of q0 is lossless in cpu speed test.
Change-Id: Ib94ee638eb486f1368f2e3098d25c1d3d28fc265
Marco Paniconi [Fri, 27 Jun 2014 17:14:56 +0000 (10:14 -0700)]
Merge "vp8: bugfix in postproc vp8_de_noise."
Alex Converse [Fri, 27 Jun 2014 17:07:03 +0000 (10:07 -0700)]
Merge "Add realtime coverage to cpu speed test"
Alex Converse [Fri, 27 Jun 2014 17:06:54 +0000 (10:06 -0700)]
Merge "Use UV prediction when deciding to skip in for lossless."
Jim Bankoski [Fri, 27 Jun 2014 17:03:15 +0000 (10:03 -0700)]
Better validation of invalid files
This patch checks that a decoder never tries to reference frame that's
outside the range of 2x to 1/16th the size of this frame. Any attempt
to do so causes a failure.
Change-Id: I5c98fa7bb95ac4f29146f29dd92b62fe96164e4c
Marco Paniconi [Fri, 27 Jun 2014 16:20:58 +0000 (09:20 -0700)]
vp8: bugfix in postproc vp8_de_noise.
Set the proper number of mb_rows/cols.
Also remove warnings (unused variable) when configured with temporal-denoising disabled.
Change-Id: I8abd2372394ee55295feb87a66efd294ea6989d0
Tim Kopp [Tue, 17 Jun 2014 20:01:34 +0000 (13:01 -0700)]
Implemented motion compensation for VP9 denoiser
Change-Id: Iee21eb0ecc5a1fe2c56fb3df0cee0ead6d139ed1
Tim Kopp [Thu, 26 Jun 2014 22:28:08 +0000 (15:28 -0700)]
fix: Only do spatial SVC when there are > 1 layers
Bug introduced in I930dced169c9d53f8044d2754a04332138347409. If
svc.number_temporal_layers == 1 and svc.number_spatial_layers == 1, the system
attempt to do spatial SVC. It no longer does that.
Change-Id: Ie6b130a72b1eea40c547c9a64447e40695f811c5
Tim Kopp [Tue, 17 Jun 2014 19:51:39 +0000 (12:51 -0700)]
VP9 denoiser: implemented update_frame_stats()
Also added reset_frame_stats()
Change-Id: I8e6ca00dbd5fa85cd39485d81c9343c0ff207d6c
Yaowu Xu [Fri, 27 Jun 2014 14:07:02 +0000 (07:07 -0700)]
Merge "Multi-arf: Change ref buffer for primary arf."
Scott LaVarnway [Fri, 27 Jun 2014 12:00:26 +0000 (05:00 -0700)]
denoising_sse2.c visual studio warnings/errors fix
Fixed reported vs warnings/errors.
Change-Id: Ia4c44d64a69dbd9e1738c6ebb48be29582b7e1d3
Paul Wilkins [Thu, 26 Jun 2014 10:26:25 +0000 (11:26 +0100)]
Multi-arf: Change ref buffer for primary arf.
For the primary arf in a group, if multiple arfs
are enabled and we were using arfs in the previous
group, then allow the second arf from the previous
group to be used as an additional reference.
Change-Id: Iaf41706a52f54ef21548026851cd77100d6aebda
Alex Converse [Thu, 26 Jun 2014 23:16:38 +0000 (16:16 -0700)]
Add realtime coverage to cpu speed test
Change-Id: I33e8d1788ce29ffc695f7e9b84167ef446b1b56c
Jingning Han [Wed, 25 Jun 2014 23:53:07 +0000 (16:53 -0700)]
Adaptive txfm size selection depending on residual sse/variance
This commit enables an adaptive transform size selection method
for speed -6. It uses largest transform size when the sse is more
than 4 times of variance, i.e., most energy is compacted in the
DC coefficient. Otherwise, use the default TX_8X8. It improves
the compression efficiency for rtc set of speed -6 by 0.8%, no
speed change observed.
Change-Id: Ie6ed1e728ff7bf88ebe940a60811361cdd19969c
Pengchong Jin [Thu, 26 Jun 2014 21:36:10 +0000 (14:36 -0700)]
Merge "Skip the partition search for the frame with no motion"
Alex Converse [Thu, 26 Jun 2014 19:55:37 +0000 (12:55 -0700)]
Use UV prediction when deciding to skip in for lossless.
Change-Id: Ic149749157d762039446d14472d40d9211c6451a
Marco Paniconi [Thu, 26 Jun 2014 20:03:49 +0000 (13:03 -0700)]
Merge "vp8: Add temporal denoising for UV-channel."
Tom Finegan [Thu, 26 Jun 2014 19:49:16 +0000 (12:49 -0700)]
Merge "Add arm64-darwin-gcc target."
Pengchong Jin [Mon, 23 Jun 2014 20:00:16 +0000 (13:00 -0700)]
Skip the partition search for the frame with no motion
This patch allows the encoder to skip the partition search for the
frame if it is an inter frame and only zero motion vectors have
been detected in the first pass. The partition size is directly
assigned according to the difference variance.
Borg tests show overall little performance changes in term of PSNR
(derf -0.027%, yt 0.152%, hd 0.078%, stdhd 0%). The worst case of
PSNR loss is -0.514% from yt. The best PSNR gain is 4.293% from yt.
The second pass encoding speedup for slideshow clips is 15%-40%.
Change-Id: I881f347d286553ee5594a9ea09ba1a61ac684045
Scott LaVarnway [Wed, 25 Jun 2014 18:28:02 +0000 (11:28 -0700)]
vp8: Add temporal denoising for UV-channel.
C version and sse2 version, and off by default.
For the test clip used, the sse2 performance improved by ~5.6%
Change-Id: Ic2d815968849db51b9d62085d7a490d0e01574f6
Tom Finegan [Thu, 26 Jun 2014 01:32:16 +0000 (18:32 -0700)]
Add arm64-darwin-gcc target.
Change-Id: I251cca45db54d2947868e2b9c35cb54f86b10706
Jingning Han [Thu, 26 Jun 2014 18:00:34 +0000 (11:00 -0700)]
Merge "Add const mark to const values in non-RD coding mode"
Jingning Han [Thu, 26 Jun 2014 18:00:25 +0000 (11:00 -0700)]
Merge "Enable real-time version reference motion vector search"
Jingning Han [Tue, 24 Jun 2014 17:45:31 +0000 (10:45 -0700)]
Enable real-time version reference motion vector search
This commit enables a fast reference motion vector search scheme.
It checks the nearest top and left neighboring blocks to decide the
most probable predicted motion vector. If it finds the two have
the same motion vectors, it then skip finding exterior range for
the second most probable motion vector, and correspondingly skips
the check for NEARMV.
The runtime of speed -5 goes down
pedestrian at 1080p 29377 ms -> 27783 ms
vidyo at 720p 11830 ms -> 10990 ms
i.e., 6%-8% speed-up.
For rtc set, the compression performance
goes down by about -1.3% for both speed -5 and -6.
Change-Id: I2a7794fa99734f739f8b30519ad4dfd511ab91a5
Jingning Han [Thu, 26 Jun 2014 16:41:21 +0000 (09:41 -0700)]
Add const mark to const values in non-RD coding mode
Change-Id: I65209fd1e06fc06833f6647cb028b414391a7017
Paul Wilkins [Thu, 26 Jun 2014 16:41:40 +0000 (09:41 -0700)]
Merge "Fix quality regression for multi arf off case."
Jingning Han [Thu, 26 Jun 2014 16:10:07 +0000 (09:10 -0700)]
Merge "Make non-RD intra mode search txfm size dependent"
Paul Wilkins [Thu, 26 Jun 2014 08:48:31 +0000 (09:48 +0100)]
Fix quality regression for multi arf off case.
Bug introduced during multiple iterations on: I3831*
gf_group->arf_update_idx[] cannot currently be used
to select the arf buffer index if buffer flipping on overlays
is enabled (still currently the case when multi arf OFF).
Change-Id: I4ce9ea08f1dd03ac3ad8b3e27375a91ee1d964dc
Paul Wilkins [Thu, 26 Jun 2014 08:55:00 +0000 (01:55 -0700)]
Merge "Dual arf: Name changes."
Tom Finegan [Thu, 26 Jun 2014 01:06:17 +0000 (18:06 -0700)]
configure.sh: Tweak default of use_x86inc for 64-bit, non-pic, and Darwin.
Mainly a readability fix.
Change-Id: Idfd4de9f25d9b9f9931d8e885d9d031473d5df6d
Tom Finegan [Thu, 26 Jun 2014 01:45:15 +0000 (18:45 -0700)]
iosbuild.sh Add arm64 support to VPX.framework.
Change-Id: I30b1813b744c11c0ca4e84554539cc524e7ee034
Jingning Han [Wed, 25 Jun 2014 18:41:49 +0000 (11:41 -0700)]
Make non-RD intra mode search txfm size dependent
This commit fixes the potential issue in the non-RD mode decision
flow that only checks part of the block to estimate the cost. It
was due to the use of fixed transform size, in replacing the
largest transform block size. This commit enables per transform
block cost estimation of the intra prediction mode in the non-RD
mode decision.
Change-Id: I14ff92065e193e3e731c2bbf7ec89db676f1e132
James Zern [Wed, 25 Jun 2014 21:26:18 +0000 (14:26 -0700)]
Merge changes I915beaef,I229dd6ca
* changes:
vp9cx.mk: move avx c files outside of x86inc block
test.mk: remove renamed file
James Zern [Wed, 25 Jun 2014 19:20:46 +0000 (12:20 -0700)]
vp9cx.mk: move avx c files outside of x86inc block
same reasoning as:
9f3a0db vp9_rtcd: correct avx2 references
these are all intrinsics, so don't depend on x86inc.asm
Change-Id: I915beaef318a28f64bfa5469e5efe90e4af5b827
hkuang [Wed, 25 Jun 2014 18:42:08 +0000 (11:42 -0700)]
Merge "Revert "Revert 3 patches from Hangyu to get Chrome to build:""
hkuang [Tue, 24 Jun 2014 21:22:09 +0000 (14:22 -0700)]
Revert "Revert 3 patches from Hangyu to get Chrome to build:"
This patch reverts the previous revert from Jim and also add a
variable user_priv in the FrameWorker to save the user_priv
passed from the application. In the decoder_get_frame function,
the user_priv will be binded with the img. This change is needed
or it will fail the unit test added here:
https://gerrit.chromium.org/gerrit/#/c/70610/
This reverts commit
9be46e4565f553460a1bbbf58d9f99067d3242ce.
Change-Id: I376d9a12ee196faffdf3c792b59e6137c56132c1
James Zern [Wed, 25 Jun 2014 18:04:14 +0000 (11:04 -0700)]
test.mk: remove renamed file
vp90-2-15-fuzz-flicker.webm was renamed in:
c3db2d8 error check vp9 superframe parsing
Change-Id: I229dd6ca4c662802c457beea0f7b4128153a65dc
Alex Converse [Wed, 25 Jun 2014 17:51:57 +0000 (10:51 -0700)]
Merge "Allow lossless breakout in non-rd mode decision."
James Zern [Wed, 25 Jun 2014 17:44:36 +0000 (10:44 -0700)]
Merge "decode_test_driver: check HasFailure() in RunLoop"
James Zern [Wed, 25 Jun 2014 17:43:49 +0000 (10:43 -0700)]
Merge "decode_test_driver: fix type size warning"
Minghai Shang [Wed, 25 Jun 2014 17:42:04 +0000 (10:42 -0700)]
[spatial svc] Remove encoding modes since we only need one mode at this time.
Change-Id: I1b6ed2e82a7e3e8a9e790bf29d667a3f856a9518
Minghai Shang [Wed, 25 Jun 2014 17:36:24 +0000 (10:36 -0700)]
[spatial svc]Remove key frame quantizer settings since key frame is decided by rate control
Change-Id: I7eda0f5e678034f0e9c2ab481c517d2e9b280eb5
Minghai Shang [Wed, 25 Jun 2014 17:31:21 +0000 (10:31 -0700)]
Merge "[spatial svc]Don't skip motion search in first pass encoding"
Minghai Shang [Wed, 25 Jun 2014 17:31:05 +0000 (10:31 -0700)]
Merge "[spatial svc]Implement lag in frames for spatial svc"
Jingning Han [Wed, 25 Jun 2014 15:57:18 +0000 (08:57 -0700)]
Merge "Replace cpi->common with preset variable cm"
Jingning Han [Wed, 25 Jun 2014 15:57:08 +0000 (08:57 -0700)]
Merge "Add vp9_ prefix to mv_pred and setup_pred_block functions"
Yunqing Wang [Wed, 25 Jun 2014 15:18:33 +0000 (08:18 -0700)]
Merge "Reuse inter prediction result in real-time speed 6"
Paul Wilkins [Wed, 25 Jun 2014 09:31:32 +0000 (10:31 +0100)]
Dual arf: Name changes.
Cosmetic patch only in response to comments on
previous patches suggesting a couple of name changes
for consistency and clarity.
Change-Id: Ida3a359b0d5755345660d304a7697a3a3686b2a3
Paul Wilkins [Wed, 25 Jun 2014 09:35:56 +0000 (02:35 -0700)]
Merge "Dual ARF changes: Buffer index selection."
Paul Wilkins [Wed, 25 Jun 2014 09:35:45 +0000 (02:35 -0700)]
Merge "Adjust arf Q limits with multi-arf."
James Zern [Wed, 25 Jun 2014 05:44:24 +0000 (22:44 -0700)]
decode_test_driver: fix type size warning
like vpx_codec_decode(), vpx_codec_peek_stream_info() takes an unsigned
int, not size_t, parameter for buffer size
Change-Id: I4ce0e1fbbde461c2e1b8fcbaac3cd203ed707460
James Zern [Sat, 21 Jun 2014 23:55:29 +0000 (16:55 -0700)]
vp9: check tile column count
the max is 6. there are assumptions throughout the decode regarding
this; fixes a crash with a fuzzed bitstream
$ zzuf -s 5861 -r 0.01:0.05 -b 6- \
< vp90-2-00-quantizer-00.webm.ivf \
| dd of=invalid-vp90-2-00-quantizer-00.webm.ivf.s5861_r01-05_b6-.ivf \
bs=1 count=81883
Change-Id: I6af41bb34252e88bc156a4c27c80d505d45f5642
Alex Converse [Tue, 24 Jun 2014 23:39:20 +0000 (16:39 -0700)]
Allow lossless breakout in non-rd mode decision.
This is very helpful for large moving windows in screencasts.
Change-Id: I91b5f9acb133281ee85ccd8f843e6bae5cadefca
Jingning Han [Mon, 23 Jun 2014 22:52:16 +0000 (15:52 -0700)]
Replace cpi->common with preset variable cm
This commit replaces a few use cases of cpi->common with preset
variable cm, to avoid unnecessary pointer fetch in the non-RD
coding mode.
Change-Id: I4038f1c1a47373b8fd7bc5d69af61346103702f6
Jingning Han [Mon, 23 Jun 2014 22:49:12 +0000 (15:49 -0700)]
Add vp9_ prefix to mv_pred and setup_pred_block functions
Make these two functions accessible by both RD and non-RD coding
modes.
Change-Id: Iecb39dbf3d65436286ea3c7ffaa9920d0b3aff85
Minghai Shang [Tue, 17 Jun 2014 21:58:25 +0000 (14:58 -0700)]
[spatial svc]Don't skip motion search in first pass encoding
Change-Id: Ia6bcdaf5a5b80e68176f60d8d00e9b5cf3f9bfe3
Minghai Shang [Tue, 24 Jun 2014 21:01:17 +0000 (14:01 -0700)]
[spatial svc]Implement lag in frames for spatial svc
Change-Id: I930dced169c9d53f8044d2754a04332138347409
Yunqing Wang [Fri, 20 Jun 2014 17:11:34 +0000 (10:11 -0700)]
Reuse inter prediction result in real-time speed 6
In real-time speed 6, no partition search is done. The inter
prediction results got from picking mode can be reused in the
following encoding process. A speed feature reuse_inter_pred_sby
is added to only enable the resue in speed 6.
This patch doesn't change encoding result. RTC set tests showed
that the encoding speed gain is 2% - 5%.
Change-Id: I3884780f64ef95dd8be10562926542528713b92c
hkuang [Mon, 23 Jun 2014 21:59:20 +0000 (14:59 -0700)]
Add unit test to test user_priv parameter.
Change-Id: I6ba6171e43e0a43331ee0a7b698590b143979c44
Adrian Grange [Wed, 18 Jun 2014 21:34:24 +0000 (14:34 -0700)]
Fix test on maximum downscaling limits
There is a normative scaling range of (x1/2, x16)
for VP9. This patch fixes the maximum downscaling
tests that are applied in the convolve function.
The code used a maximum downscaling limit of x1/5
for historic reasons related to the scalable
coding work. Since the downsampling in this
application is non-normative it will revert to
using a separate non-normative scaler.
Change-Id: Ide80ed712cee82fe5cb3c55076ac428295a6019f
Tim Kopp [Tue, 24 Jun 2014 16:48:11 +0000 (09:48 -0700)]
Merge "Fixed VP9 denoiser COPY_BLOCK case"
Paul Wilkins [Mon, 23 Jun 2014 12:07:24 +0000 (13:07 +0100)]
Dual ARF changes: Buffer index selection.
Add indirection to the section of buffer indices.
This is to help simplify things in the future if we
have other codec features that switch indices.
Limit the max GF interval for static sections to fit
the gf_group structures.
Change-Id: I38310daaf23fd906004c0e8ee3e99e15570f84cb
Paul Wilkins [Mon, 23 Jun 2014 13:59:46 +0000 (14:59 +0100)]
Adjust arf Q limits with multi-arf.
Adjust enforced minimum arf Q deltas for non primary arfs
in the middle of an arf/gf group.
Change-Id: Ie8034ffb3ac00f887d74ae1586d4cac91d6cace2
Paul Wilkins [Mon, 23 Jun 2014 11:17:07 +0000 (12:17 +0100)]
Further dual arf changes: multi_arf_allowed.
Add multi_arf_allowed flag.
Re-initialize buffer indices every kf.
Add some const indicators.
Change-Id: If86c39153517c427182691d2d4d4b7e90594be71