Scott LaVarnway [Fri, 19 Oct 2012 23:02:50 +0000 (16:02 -0700)]
Re-enable specialized sse2 block loop filter functions
vp8_loop_filter_bh and vp8_loop_filter_bv
Change-Id: I0a77d38c701a6cce96cb8371c7f3ab7e2e7f878c
Scott LaVarnway [Fri, 19 Oct 2012 22:52:12 +0000 (15:52 -0700)]
sse2 intrinsic version of vp8_mbloop_filter_vertical_edge()
First sse2 version of vp8_mbloop_filter_vertical_edge(). For now,
intrinsics are being used until the bitstream is finalized. This function
will be revisited later for further performance improvements.
For the test clip used, a 34+% decoder performance improvement
was seen. This will vary depending on material.
Change-Id: I455b438bc8d8af76cf7533ac42eda5f689b21f7c
Scott LaVarnway [Thu, 18 Oct 2012 21:29:26 +0000 (14:29 -0700)]
sse2 intrinsic version of vp8_mbloop_filter_horizontal_edge()
First sse2 version of vp8_mbloop_filter_horizontal_edge(). For now,
intrinsics are being used until the bitstream is finalized. This function
will be revisited later for further performance improvements.
For the test clip used, a 31+% decoder performance improvement
was seen. This will vary depending on material.
Change-Id: I03ed3a7182478bdd1f094644ff3e0442625600e7
Scott LaVarnway [Wed, 17 Oct 2012 16:57:23 +0000 (09:57 -0700)]
Merge "Removed the loopfilter rtcd invoke macro code" into experimental
Jim Bankoski [Wed, 17 Oct 2012 16:18:54 +0000 (09:18 -0700)]
removed obselete build dependency
this commit fixes the build on windows with visual studio 2008.
Change-Id: I0baa4044e9e54237da29f2e17332ea6f766dbbec
Scott LaVarnway [Tue, 16 Oct 2012 23:19:35 +0000 (16:19 -0700)]
Removed the loopfilter rtcd invoke macro code
Change-Id: I446b2ffcbe732ffb112dbd97a4799272d4c01a84
Jim Bankoski [Tue, 16 Oct 2012 20:49:22 +0000 (13:49 -0700)]
Merge "removed the recon rtcd invoke macro code (unrevert)" into experimental
Ronald S. Bultje [Tue, 16 Oct 2012 19:41:56 +0000 (12:41 -0700)]
Fix a mismatch if a keyframe contains skip-blocks.
Change-Id: I0681d3183f51627be8c2bb76f343b7270f9116d8
Jim Bankoski [Sun, 14 Oct 2012 01:49:44 +0000 (18:49 -0700)]
removed the recon rtcd invoke macro code (unrevert)
This reinstates reverted commit
2113a831575d81faeadd9966e256d58b6b2b1633
Change-Id: I9a9af13497d1e58d4f467e3e083fddf06b1b786c
Ronald S. Bultje [Tue, 16 Oct 2012 02:50:20 +0000 (19:50 -0700)]
Merge "Remove test from B_MODE_INFO." into experimental
Ronald S. Bultje [Tue, 16 Oct 2012 02:50:06 +0000 (19:50 -0700)]
Merge "Remove mode_rdopt from MB_MODE_INFO." into experimental
Ronald S. Bultje [Tue, 16 Oct 2012 02:49:37 +0000 (19:49 -0700)]
Merge "Properly use chroma block object for chroma coefficient costing." into experimental
Yaowu Xu [Mon, 15 Oct 2012 22:02:36 +0000 (15:02 -0700)]
Merge "Changed to use real pixels only for evaluating MVs" into experimental
Yaowu Xu [Mon, 15 Oct 2012 21:30:15 +0000 (14:30 -0700)]
Changed to use real pixels only for evaluating MVs
The commit changed to avoid using pixels from extended border in
in evaluating and select best reference motion vector.
Change-Id: I39b758889373e42ed2889d59744388e5b9c1a20a
Ronald S. Bultje [Mon, 15 Oct 2012 20:54:05 +0000 (13:54 -0700)]
Remove test from B_MODE_INFO.
It is essentially a duplicate of mode for RD-only purposes. Removing it
saves us 4 bytes per B_MODE_INFO, or ~0.5MB for a 1080p video encode.
Change-Id: I0a54db5f51658b3946d7efb1ca6e8cfbda0cdf88
Ronald S. Bultje [Mon, 15 Oct 2012 20:49:45 +0000 (13:49 -0700)]
Remove mode_rdopt from MB_MODE_INFO.
The variable is essentially a duplicate of mode for RD-only purposes.
Removing it gives identical results, and saves 4 bytes per macroblock
(i.e. 32.5kB for a 1080p HD video encode).
Change-Id: I22d5058fdb80ab0b69862caee825e9d86bb148b3
Ronald S. Bultje [Mon, 15 Oct 2012 20:36:00 +0000 (13:36 -0700)]
Properly use chroma block object for chroma coefficient costing.
Change-Id: Ieb38c7aae91dbaca4a8add204fa84e1cfc459933
Ronald S. Bultje [Mon, 15 Oct 2012 17:52:13 +0000 (10:52 -0700)]
Add a new token stuffing function vp8_stuff_mb().
This way a caller doesn't need to implement the logic for which (and how
many) tokens to write out to stuff one macroblock worth of EOBs. Make
the actual function implementations static, since they are now only used
in tokenize.c; also do some minor stylistic changes so it follows the
style guide a little more closely; use PLANE_TYPE where appropriate,
remove old (stale) frame_type function arguments; hardcode plane type
where only a single one is possible (2nd order DC or U/V EOB stuffing);
support stuffing 8x8/4x4 transform EOBs with no 2nd order DC.
Change-Id: Ia448e251d19a4e3182eddeb9edd034bd7dc16fa3
Ronald S. Bultje [Sun, 14 Oct 2012 22:29:56 +0000 (15:29 -0700)]
Add and consistently use PLANE_TYPE.
Change the macros PLANE_TYPE_{Y_NO_DC,Y2,UV,Y_WITH_DC} to a typed enum,
and use this typed enum consistently across all places where relevant.
In places where the type is implied (e.g. in functions that only handle
second order planes or chroma planes), remove it as a function argument
and instead hardcode the proper enum in the code directly.
Change-Id: I93652b4a36aa43163d49c732b0bf5c4442738c47
Ronald S. Bultje [Sat, 13 Oct 2012 18:46:21 +0000 (11:46 -0700)]
Merge duplicate loops in tokenization code.
Also merge the three occurrences of 4x4 chroma block writing into a
single function, and call that function instead of duplicating the
4x4 chroma tokenization code in 3 places.
Change-Id: I7913538d1029f709b0e3ae49fff1148d3be9eeb9
Ronald S. Bultje [Sat, 13 Oct 2012 16:27:54 +0000 (09:27 -0700)]
Minor refactoring in encodeintra.c.
Merge code blocks for different transform sizes; use MACROBLOCKD as a
temp variable where that leads to smaller overall source code; remove
duplicate code under #if CONFIG_HYBRIDTRANSFORM/#else blocks. Some style
changes to make it follow the style guide a little better.
Change-Id: I1870a06dae298243db46e14c6729c96c66196525
Ronald S. Bultje [Sat, 13 Oct 2012 15:31:02 +0000 (08:31 -0700)]
Make optimize_mb_16x16() static.
It is not used outside this file.
Change-Id: Id0947180edab4c5f1f50589863350b21e97c25f1
Ronald S. Bultje [Sat, 13 Oct 2012 15:29:59 +0000 (08:29 -0700)]
Remove assignment in vp8_optimize_mby_16x16().
The result of the assignment is unused.
Change-Id: Id94b790baa6451f24e9aeda6f036390388d93921
Ronald S. Bultje [Sat, 13 Oct 2012 15:26:05 +0000 (08:26 -0700)]
Remove duplicate code in optimize_mb_NxN vs. vp8_optimize_mby/uv_NxN().
Instead, just call vp8_optimize_mby/uv_NxN() inside optimize_mb_NxN().
Change-Id: Ief941064d86db4277ba02e990a26819144b55f7c
Ronald S. Bultje [Sat, 13 Oct 2012 15:15:51 +0000 (08:15 -0700)]
Remove duplicate or unused code in encoder/encodemb.c.
Also make some minor stylistic changes to bring the code closer to
the style guide. Remove distinction between inter and intra transform
functions, since both do exactly the same thing except for the check
against SPLITMV for the second-order transform. Remove some commented
out debug code. Remove 8x8/16x16 transform code in encode_inter16x16y(),
since the first-pass only uses 4x4 anyway.
Change-Id: Ife54816ff759825b9141f95dc2ba43c253c14dba
Ronald S. Bultje [Sat, 13 Oct 2012 05:42:06 +0000 (22:42 -0700)]
Remove duplicate or unused code in encoder/quantize.c.
Also make some minor stylistic changes to bring the code closer to
the style guide. Remove checks against i8x8/bpred in the mb-codepath,
since these do individual block reconstruction and thus don't go through
this codepath.
Change-Id: I4dfcf8f78746f4647a206475acf731837aa4fd47
Ronald S. Bultje [Sat, 13 Oct 2012 04:59:50 +0000 (21:59 -0700)]
Remove duplicate or unused code in common/invtrans.c.
Also make some minor stylistic changes to bring the code closer to
the style guide.
Change-Id: Idafa280724f49868ed6b7d65b7fd9227d781994b
Ronald S. Bultje [Sat, 13 Oct 2012 04:41:58 +0000 (21:41 -0700)]
Add a _4x4 sufix to all 4x4-transform-related functions.
This includes trellis optimization, forward/inverse transform,
quantization, tokenization and stuffing functions.
Change-Id: Ibd34132e1bf0cd667671a57b3f25b3d361b9bf8a
Ronald S. Bultje [Sat, 13 Oct 2012 04:08:03 +0000 (21:08 -0700)]
Replace vp8_inverse_htransform_b() by direct call to vp8_ihtllm_c().
Change-Id: I55ee7566341b04d7e7340fdab95189f82231a792
Jim Bankoski [Sun, 14 Oct 2012 03:29:04 +0000 (20:29 -0700)]
Revert "removed the recon. rtcd invoke macro code"
This reverts commit
2113a831575d81faeadd9966e256d58b6b2b1633
Jim Bankoski [Sun, 14 Oct 2012 01:49:44 +0000 (18:49 -0700)]
removed the recon. rtcd invoke macro code
Code clean up - removed rtcd
Change-Id: Id963ecf53c370b1d99484ef18d6befeed7e0c748
Adrian Grange [Fri, 12 Oct 2012 22:08:24 +0000 (15:08 -0700)]
Merge "Remove reverting of tx-select if only a single txfm-size is used." into experimental
Yaowu Xu [Fri, 12 Oct 2012 18:38:10 +0000 (11:38 -0700)]
removed the dependency cross key frame boundary
This allowed decoding from any key frame after skipping the frames
prior.
Change-Id: I096fccad5346d75ae50a017c94eb8d772a7e8b00
Yaowu Xu [Fri, 12 Oct 2012 19:23:18 +0000 (12:23 -0700)]
Merge "added the ablility to skip first n frames for vpxdec" into experimental
Ronald S. Bultje [Fri, 12 Oct 2012 01:19:20 +0000 (18:19 -0700)]
Remove reverting of tx-select if only a single txfm-size is used.
Entropy coding takes care of this anyway, and this causes changes to
the txfm size assigned to skip blocks, which can affect the loopfilter
output, thus causing encoder/decoding mismatches.
Change-Id: I591a8d8a4758a507986b751a9f83e6d76e406998
Ronald S. Bultje [Fri, 12 Oct 2012 18:06:55 +0000 (11:06 -0700)]
Remove CONFIG_TX16X16, since the experiment was merged.
Change-Id: Ib74824c824f7ebb279e7df8eb32440d4d65fbd79
Yaowu Xu [Thu, 11 Oct 2012 22:12:35 +0000 (15:12 -0700)]
added the ablility to skip first n frames for vpxdec
Change-Id: Id7d1f25078be92dd6cda246738ad6fb50579f555
Ronald S. Bultje [Thu, 11 Oct 2012 17:13:48 +0000 (10:13 -0700)]
Add encoder/decoder mismatch information to internal stats file.
Change-Id: Ibb6ba26e8718d3af27553ca59443a8c6aec7749d
Ronald S. Bultje [Thu, 11 Oct 2012 00:18:22 +0000 (17:18 -0700)]
Use transform-size selection for i8x8_pred also.
Change-Id: Iecb282fc89f9b5145ef31c5eda294ad42bc32a5d
Deb Mukherjee [Wed, 10 Oct 2012 18:20:33 +0000 (11:20 -0700)]
Merge of the TX_16X16 experiment
Change-Id: I22aa803ffff330622cdb77277e7b196a9766f882
Scott LaVarnway [Wed, 10 Oct 2012 16:08:03 +0000 (09:08 -0700)]
Merge "Removing dr[16] from VP8D_COMP" into experimental
John Koleszar [Wed, 10 Oct 2012 15:49:50 +0000 (08:49 -0700)]
Merge "write_kfmodes: use update_skip_probs()" into experimental
Paul Wilkins [Wed, 10 Oct 2012 12:45:19 +0000 (05:45 -0700)]
Merge "convert copy16x16 to rtcd" into experimental
Jim Bankoski [Wed, 10 Oct 2012 00:09:08 +0000 (17:09 -0700)]
convert copy16x16 to rtcd
Convert copy16x16 from invoke to rtcd. The first in a long
string of converts.
Change-Id: I296b0aa32f40e9fb649f7a3cb914a4e5300cad63
Scott LaVarnway [Tue, 9 Oct 2012 22:54:32 +0000 (15:54 -0700)]
Removing dr[16] from VP8D_COMP
Currently not used. Also using this as a test commit into the
experimental branch.
Change-Id: I68c479461ae02f1c7a7f75fa02d1fc0db84bf198
John Koleszar [Tue, 9 Oct 2012 22:35:57 +0000 (15:35 -0700)]
write_kfmodes: use update_skip_probs()
Use the common update_skip_probs() function rather than duplicating its
logic in write_kf_modes().
Change-Id: I2890a28f6907cb79ffe0fb21d20f0ef98b85cdd9
Ronald S. Bultje [Tue, 9 Oct 2012 16:18:21 +0000 (09:18 -0700)]
Selectable transform size try 2.
Results: derf (vanilla or +hybridtx) +0.2% and (+hybrid16x16
or +tx16x16) +0.7%-0.8%; HD (vanilla or +hybridtx) +0.1-0.2%
and (+hybrid16x16 or +tx16x16) +1.4%, STD/HD (vanilla or +hybridtx)
about even, and (+hybrid16x16 or +tx16x16) +0.8-1.0%.
Change-Id: I03899e2f7a64e725a863f32e55366035ba77aa62
Deb Mukherjee [Fri, 5 Oct 2012 13:49:08 +0000 (06:49 -0700)]
Merge "Mismatch bug-fix when transform expts are off" into experimental
Deb Mukherjee [Fri, 5 Oct 2012 12:08:20 +0000 (05:08 -0700)]
Mismatch bug-fix when transform expts are off
A variety of issues related to tangling of various hybrid transform
experiments.
Change-Id: I84dac65ff513f8b024407760df55fffa95e198de
Paul Wilkins [Fri, 5 Oct 2012 10:16:46 +0000 (11:16 +0100)]
Fix SIMD unsafe use of floating point.
This commit fixes unsafe simd / floating point interactions arising
from the current hybrid and 16x16 transform implementation.
These led to a raft of bugs and issues when the project was
built using VS2008 for Win32 though they did not show up with
the unix builds.
Gerrit makes a meal out of presenting the fix but all I have actually
done is indent the body of each function that uses floating point by
one level and bracket with emms instructions using the function
vp8_clear_system_state(). See below.
function () {
vp8_clear_system_state();
{
... function body
}
vp8_clear_system_state();
}
This is almost certainly over the top in terms of number of emms
instructions but is a temporary measure pending implementation of
integer variants of each function to replace the floating point.
Limited testing suggests that this fixes the problems that arose for
Win32 VS2008 when the hybrid or 16x16 transforms were enabled.
Change-Id: I7c9a72bd79315246ed880578dec51e2b7c178442
John Koleszar [Wed, 3 Oct 2012 15:28:51 +0000 (08:28 -0700)]
Merge "all_builds.py: add ability to pass extra configure flags" into experimental
Deb Mukherjee [Mon, 10 Sep 2012 05:42:35 +0000 (22:42 -0700)]
Entropy coding for hybrid transform
Separates the entropy coding context models for 4x4, 8x8 and 16x16
ADST variants.
There is a small improvement for HD (hd/std-hd) by about 0.1-0.2%.
Results on derf/yt are about the same, probably because there is not
enough statistics.
Results may improve somewhat once the initial probability tables are
updated for the hybrid transforms which is coming soon.
Change-Id: Ic7c0c62dacc68ef551054fdb575be8b8507d32a8
Deb Mukherjee [Fri, 21 Sep 2012 21:20:15 +0000 (14:20 -0700)]
Separating 8x8DCT use from hybridtransform8x8
With this change, even if hybridtransform8x8 experiment is off,
8x8 dct is used for the I8x8 mode. However note that the gains
observed with the hybridtransform8x8 experiment will now be less,
since part of the gain is now merged in.
Change-Id: I9afb3880906fd0a1368a374041fc08efcf060c54
Yaowu Xu [Thu, 20 Sep 2012 19:41:59 +0000 (12:41 -0700)]
fixed inconsistency of transform size decision
for MBs using i8x8 prediciton mode when hybridtransform8x8 is not
enabled in configure.
Change-Id: I3e3c8bb1c0261108c1c71213ca9557b0ce558036
Adrian Grange [Mon, 17 Sep 2012 22:37:56 +0000 (15:37 -0700)]
Added default return value to transform selector
The non-void function didn't return a value in the default case.
Change-Id: Ie4c965315bedc189d2080dd65cf29d3347661946
Yaowu Xu [Fri, 14 Sep 2012 16:20:58 +0000 (09:20 -0700)]
removed redundant initialization
Change-Id: I34569375e520759860504c3dfb17a2c27ceb9aed
Yaowu Xu [Fri, 14 Sep 2012 15:58:02 +0000 (08:58 -0700)]
Use 3 rows and cols of pixels for ref_mv scoring
The commit changed to use 3 rows above and 3 cols from left for SAD
scoring for selecting the best reference motion vector. The change
helped std-hd set by >.2% on psnr/ssim metrics.
Change-Id: Ifad3b528d0b4b6e3c22518af789d76eff23c1520
Adrian Grange [Thu, 13 Sep 2012 18:37:55 +0000 (11:37 -0700)]
Removed redundant line.
Removed spurious line that has no effect.
Change-Id: I9dc90babbbfcaca049d4a40dc996edc7de123e01
Deb Mukherjee [Thu, 6 Sep 2012 16:07:42 +0000 (09:07 -0700)]
Adds feature for companded MV encoding
The high-precision (1/8) pel bit is turned off if the reference
MV is larger than a threshold. The motivation for this patch is
the intuition that if motion is likely large (as indicated by
the reference), there is likley to be more motion blur, and as
a result 1/8 pel precision would be wasteful both in rd sense
as well as computationally.
The feature is incorporated as part of the newmventropy experiment.
There is a modest RD improvement with the patch. Overall the
results with the newmventropy experiment with the threshold being
16 integer pels are:
derf: +0.279%
std-hd: +0.617%
hd: +1.299%
yt: +0.822%
With threshold 8 integer pels are:
derf: +0.295%
std-hd: +0.623%
hd: +1.365%
yt: +0.847%
Patch: rebased
Patch: rebase fixes
Change-Id: I4ed14600df3c457944e6541ed407cb6e91fe428b
Deb Mukherjee [Wed, 12 Sep 2012 02:36:28 +0000 (19:36 -0700)]
Cleanups for hybrid transform
Some cleanups on the transform size and type selection logic.
Change-Id: If2e9675459482242cf83b4f7de7634505e3f6dac
Yaowu Xu [Tue, 11 Sep 2012 22:00:51 +0000 (15:00 -0700)]
Merge "Revert "Hybrid transform cleanups"" into experimental
Debargha Mukherjee [Tue, 11 Sep 2012 21:52:20 +0000 (14:52 -0700)]
Revert "Hybrid transform cleanups"
This reverts commit
c5af82b7ed6d89a4a12003b8d1c6b4befcadce92
There was a mismatch bug introduced by this patch. Hence reverting until it is fixed.
Deb Mukherjee [Mon, 10 Sep 2012 22:39:37 +0000 (15:39 -0700)]
Merge "Hybrid transform cleanups" into experimental
Deb Mukherjee [Mon, 10 Sep 2012 06:10:30 +0000 (23:10 -0700)]
Hybrid transform cleanups
Some cleanups that will make it easier to maintain the code
and incorporate upcoming changes on entropy coding for the
hybrid transforms.
Change-Id: I44bdba368f7b8bf203161d7a6d3b1fc2c9e21a8f
Paul Wilkins [Fri, 7 Sep 2012 13:21:22 +0000 (14:21 +0100)]
Partial merge of NEWBESTREFMV and NEW_MVREF.
This commit merges those parts of the CONFIG_NEW_MVREF
that specifically relate to choosing a better set of candidate
MV references into the NEWBESTREFMV experiment.
CONFIG_NEW_MVREF will then be used for changes relating
to the explicit coding of a cost optimized MV reference in the
bitstream as part of MV coding.
Change-Id: Ied982c0ad72093eab29e38b8cd74d5c3d7458b10
Paul Wilkins [Fri, 7 Sep 2012 11:46:41 +0000 (12:46 +0100)]
MV reference changes
Extend experiment to use both vectors from MBs
coded using compound prediction as candidates.
In final sort only consider best 4 candidates
for now but make sure 0,0 is always one of them.
Other minor changes to new MV reference code.
Pass in Mv list to vp8_find_best_ref_mvs().
Change-Id: Ib96220c33c6b80bd1d5e0fbe8b68121be7997095
Deb Mukherjee [Thu, 26 Jul 2012 20:42:07 +0000 (13:42 -0700)]
New motion vector entropy coding
Adds a new experiment with redesigned/refactored motion vector entropy
coding. The patch also takes a first step towards separating the
integer and fractional pel components of a MV. However the fractional
pel encoding still depends on the integer pel part and so they are
not fully independent. Further experiments are in progress to see
how much they can be decoupled without affecting performance.
All components including entropy coding/decoding, costing for MV
search, forward updates and backward updates to probability tables,
have been implemented.
Results so far:
derf: +0.19%
std-hd: +0.28%
yt: +0.80%
hd: +1.15%
Patch: Simplifies the fractional pel models:
derf: +0.284%
std-hd: +0.289%
yt: +0.849%
hd: +1.254%
Patch: Some changes in the models, rebased.
derf: +0.330%
std-hd: +0.306%
yt: +0.816%
hd: +1.225%
Change-Id: I646b3c48f3587f4cc909639b78c3798da6402678
Jingning Han [Wed, 29 Aug 2012 18:25:38 +0000 (11:25 -0700)]
hybrid transform of 16x16 dimension
Enable ADST/DCT of dimension 16x16 for I16X16 modes. This change provides
benefits mostly for hd sequences.
Set up the framework for selectable transform dimension.
Also allowing quantization parameter threshold to control the use
of hybrid transform (This is currently disabled by setting threshold
always above the quantization parameter. Adaptive thresholding can
be built upon this, which will further improve the coding performance.)
The coding performance gains (with respect to the codec that has all
other configuration settings turned on) are
derf: 0.013
yt: 0.086
hd: 0.198
std-hd: 0.501
Change-Id: Ibb4263a61fc74e0b3c345f54d73e8c73552bf926
Ronald S. Bultje [Thu, 30 Aug 2012 16:11:41 +0000 (09:11 -0700)]
Merge "Add proper skip support to intra frames." into experimental
Ronald S. Bultje [Wed, 29 Aug 2012 17:43:20 +0000 (10:43 -0700)]
Add proper skip support to intra frames.
Change-Id: I7524883fb29f42303fb46a5bc6772fbcf8781d1d
Ronald S. Bultje [Tue, 28 Aug 2012 16:51:52 +0000 (09:51 -0700)]
Remove some disabled code from pick_inter_mode_sb().
Change-Id: I323c429b453eec8c3b76c37f7e2a52da9dfe8d77
Ronald S. Bultje [Tue, 28 Aug 2012 01:04:53 +0000 (18:04 -0700)]
Disable 16x16 temp buffer for superblocks.
Change-Id: I3d585bae446c16fcf7a39d72c1bcd6468a7bb1a1
John Koleszar [Wed, 22 Aug 2012 17:48:20 +0000 (10:48 -0700)]
all_builds.py: add ability to pass extra configure flags
Useful to pass --target etc to this script on Jenkins.
Change-Id: I9b22339a211df543077bcc84bdef4460f3ca5c9c
John Koleszar [Fri, 24 Aug 2012 22:21:49 +0000 (15:21 -0700)]
Merge "Fix file permissions" into experimental
Ronald S. Bultje [Fri, 24 Aug 2012 19:10:04 +0000 (12:10 -0700)]
Merge "change to properly account for coef scaling in unit tests" into experimental
Yaowu Xu [Fri, 24 Aug 2012 16:59:49 +0000 (09:59 -0700)]
change to properly account for coef scaling in unit tests
The transform functions in experimental branch absorbed a scaling
factor of 4 to allow quantization steps closer to unit quantizer.
This commit added scaling code in between forward and inverse
transform to properly account for the scaling factor.
Change-Id: I9a573ddc1ffa74973b34800a5da1a56dbabe0949
Paul Wilkins [Fri, 24 Aug 2012 14:44:01 +0000 (15:44 +0100)]
New Motion Reference Search
Alternative strategy for finding a list of candidate motion
vectors to use as reference values in mv coding and as
nearest and near.
Sort by sad in vp8_find_best_ref_mvs() rather than just
pick the best. Allow 0,0 as a best ref option but not a
nearest or near unless there are no alternatives.
Encode/Decode verified on at least some clips.
Some commented out experimental and stats code still in place.
Gain over existing code averages about 1% on derf (alll metrics)
with improvement on all clips. Other test results pending.
The entropy coding of the mode (nearest/near etc) still
depends upon and requires the old "findnear" code so
this needs looking at and may provide room for further gains.
Change-Id: I871d7cba1d1c379c4bad9bcccce1fb19c46b8247
Ronald S. Bultje [Wed, 22 Aug 2012 23:05:21 +0000 (16:05 -0700)]
Remove duplicate code in C intra prediction.
Change-Id: If4eb0fd60cc5bc2f32254e82e8d6cfa4120cb558
John Koleszar [Wed, 22 Aug 2012 17:16:05 +0000 (10:16 -0700)]
Fix file permissions
Change-Id: I8d1a2c664cbeb939cf23bcd414b23e19da760a14
John Koleszar [Wed, 22 Aug 2012 17:03:06 +0000 (10:03 -0700)]
Merge "all_builds.py: move to tools/" into experimental
John Koleszar [Wed, 22 Aug 2012 17:02:45 +0000 (10:02 -0700)]
Merge "all_builds.py: support for sharding builds" into experimental
John Koleszar [Wed, 22 Aug 2012 17:01:39 +0000 (10:01 -0700)]
Merge "remove rotation experiment" into experimental
Christian Duvivier [Wed, 22 Aug 2012 00:17:42 +0000 (17:17 -0700)]
Merge "SSE2 version of vectorized 8-tap filtering." into experimental
Ronald S. Bultje [Wed, 22 Aug 2012 00:12:41 +0000 (17:12 -0700)]
Merge "Fix mismatch caused by a rebase failure." into experimental
Ronald S. Bultje [Tue, 21 Aug 2012 23:45:51 +0000 (16:45 -0700)]
Fix mismatch caused by a rebase failure.
Change-Id: Ifff660de60fb12d745092eafc17f6a49f6e1fb53
Christian Duvivier [Tue, 21 Aug 2012 06:49:43 +0000 (23:49 -0700)]
SSE2 version of vectorized 8-tap filtering.
About 20% overall encoder speedup (vs. about 30% for sse4 version).
Change-Id: Ibf608a6a1bc94b14ec47e8046d3206b275b5a8bd
Daniel Kang [Tue, 21 Aug 2012 22:08:54 +0000 (15:08 -0700)]
Fix compiler warnings
Change-Id: I887cdcf90b87736f2a3805420aa41eed93324bee
Christian Duvivier [Tue, 21 Aug 2012 21:58:31 +0000 (14:58 -0700)]
Fix warning.
Change-Id: I99311811820169d62fda3febac4bafacfc7a538b
John Koleszar [Tue, 21 Aug 2012 18:18:58 +0000 (11:18 -0700)]
all_builds.py: move to tools/
Change-Id: I64c470fb5a4f32a862cfb5424fb95baac47fcc24
John Koleszar [Tue, 21 Aug 2012 18:18:38 +0000 (11:18 -0700)]
all_builds.py: support for sharding builds
Allow sharding of the builds with the --shard= and --shards= options.
Change-Id: I9d5552ad2edc0b1210e96f0e94ce7dfd645c45f8
John Koleszar [Tue, 21 Aug 2012 17:10:43 +0000 (10:10 -0700)]
Merge changes I96cd08c9,If8eeb8ff into experimental
* changes:
hybridtransform8x8: fix compilation
comp_intra_pred: fix compilation
John Koleszar [Tue, 21 Aug 2012 17:09:56 +0000 (10:09 -0700)]
remove rotation experiment
This is being reimplemented more generically in terms of affine
transforms.
Change-Id: I9300bfde5f8b93c708c64f59427087720f8ed782
John Koleszar [Tue, 21 Aug 2012 16:31:31 +0000 (09:31 -0700)]
hybridtransform8x8: fix compilation
corrected ifdefs to include CONFIG_HYBRIDTRANSFORM8X8 for missing
definitions
Change-Id: I96cd08c98d74b5ba10866f3fa30252073ce92fa5
Ronald S. Bultje [Tue, 21 Aug 2012 00:45:36 +0000 (17:45 -0700)]
Fix some compiler warnings.
Change-Id: Iccb56e3b966d1519ba49dfe4e1396479c6b54a6f
John Koleszar [Tue, 21 Aug 2012 00:38:44 +0000 (17:38 -0700)]
comp_intra_pred: fix compilation
Added missing parameters to calls to:
vp8_build_intra_predictors_internal
vp8_build_intra_predictors_mbuv_internal
Change-Id: If8eeb8ff23eff4572397b404fe61be5d0c950bbe
John Koleszar [Tue, 21 Aug 2012 00:09:30 +0000 (17:09 -0700)]
Merge "Fix typo." into experimental
Ronald S. Bultje [Tue, 21 Aug 2012 00:07:24 +0000 (17:07 -0700)]
Fix typo.
Change-Id: I95217ce52675ad00637fc8f60d4d0dd054efd512
Ronald S. Bultje [Tue, 21 Aug 2012 00:04:23 +0000 (17:04 -0700)]
Merge "Superblock coding." into experimental
Ronald S. Bultje [Mon, 20 Aug 2012 21:43:34 +0000 (14:43 -0700)]
Superblock coding.
This commit adds a pick_sb_mode() function which selects the best 32x32
superblock coding mode. Then it selects the best per-MB modes, compares
the two and encodes that in the bitstream.
The bitstream coding is rather simplistic right now. At the SB level,
we code a bit to indicate whether this block uses SB-coding (32x32
prediction) or MB-coding (anything else), and then we follow with the
actual modes. This could and should be modified in the future, but is
omitted from this commit because it will likely involve reorganizing
much more code rather than just adding SB coding, so it's better to let
that be judged on its own merits.
Gains on derf: about even, YT/HD: +0.75%, STD/HD: +1.5%.
Change-Id: Iae313a7cbd8f75b3c66d04a68b991cb096eaaba6
Christian Duvivier [Fri, 17 Aug 2012 22:37:51 +0000 (15:37 -0700)]
Fix warnings.
Change-Id: I01dd96da18fe22450bb372b10ba90eab52e29bdf