Scott LaVarnway [Sat, 12 Jan 2013 01:11:04 +0000 (17:11 -0800)]
WIP: Added sse2 version of vp9_mb_lpf_horizontal_edge_w
and vp9_mb_lpf_vertical_edge_w_sse2. This was quickly done so we can
run some tests over the weekend. Future commits will optimize/refactor these
functions further.
The decoder performance improved by ~17% for the clip used.
Change-Id: I612687cd5a7670ee840a0cbc3c68dc2b84d4af76
Scott LaVarnway [Fri, 11 Jan 2013 22:54:14 +0000 (14:54 -0800)]
Initial sse2 version of the wide loopfilters
Updated the rtcd_defs and used the sse2 uv version
of the loopfilter. The performance improved by ~8%
for the test clip used.
Change-Id: I5a0bca3b6674198d40ca4a77b8cc722ddde79c36
Yaowu Xu [Fri, 11 Jan 2013 19:15:43 +0000 (11:15 -0800)]
Merge "Reduce the usage of widerlpf" into experimental
Jim Bankoski [Fri, 11 Jan 2013 17:45:45 +0000 (09:45 -0800)]
rtcd for new wider loop filters
Change-Id: I8826bcdcf72ba6d86bde31cd13902a710399805c
Yaowu Xu [Thu, 10 Jan 2013 18:55:07 +0000 (10:55 -0800)]
Reduce the usage of widerlpf
The commit changed to not to use wider lpf within a superblock when
32x32 transform is used for the block.
The commit also changed to use the shorter version of loop filtering:
for UV planes.
Change-Id: I344c1fb9a3be9d1200782a788bcb0b001fedcff8
Ronald S. Bultje [Thu, 10 Jan 2013 16:36:42 +0000 (08:36 -0800)]
Fix wrong pitch argument in dct32x32 unit test.
Change-Id: Id9474a1686daebfa3d004e21823bf1888ec9e534
Ronald S. Bultje [Thu, 10 Jan 2013 16:23:59 +0000 (08:23 -0800)]
Merge tx32x32 experiment.
Change-Id: I615651e4c7b09e576a341ad425cf80c393637833
Ronald S. Bultje [Thu, 10 Jan 2013 16:18:33 +0000 (08:18 -0800)]
Merge "Merge superblocks64 experiment." into experimental
Ronald S. Bultje [Thu, 10 Jan 2013 01:21:28 +0000 (17:21 -0800)]
Merge superblocks64 experiment.
Change-Id: If6c88752dffdb566f8d4322f135145270716fb8e
Yaowu Xu [Wed, 19 Dec 2012 23:53:03 +0000 (15:53 -0800)]
experiment a wider loop filter for MB border
when larger transforms are used
Change-Id: I25251442b44bf251df4c25a1c1fcf71fb2ad913b
Adrian Grange [Tue, 8 Jan 2013 22:14:01 +0000 (14:14 -0800)]
New prediction filter
This patch removes the old pred-filter experiment and replaces it
with one that is implemented using the switchable filter framework.
If the pred-filter experiment is enabled, three interopolation
filters are tested during mode selection; the standard 8-tap
interpolation filter, a sharp 8-tap filter and a (new) 8-tap
smoothing filter.
The 6-tap filter code has been preserved for now and if the
enable-6tap experiment is enabled (in addition to the pred-filter
experiment) the original 6-tap filter replaces the new 8-tap smooth
filter in the switchable mode.
The new experiment applies the prediction filter in cases of a
fractional-pel motion vector. Future patches will apply the filter
where the mv is pel-aligned and also to intra predicted blocks.
Change-Id: I08e8cba978f2bbf3019f8413f376b8e2cd85eba4
Deb Mukherjee [Tue, 8 Jan 2013 20:18:16 +0000 (12:18 -0800)]
Adds 64x64 hybrid dct/dwt transform
This is to add to the 64x64 transform experiment as an alternative to
a 64x64 DCT.
Two levels of wavelet decomposition is used on a 64x64 block, followed
by 16x16 DCT on the four lowest subbands. The highest three subbands
are left untransformed after the first level DWT.
Change-Id: I3d48d5800468d655191933894df6b46e15adca56
Ronald S. Bultje [Tue, 8 Jan 2013 21:31:37 +0000 (13:31 -0800)]
Merge "Merge superblocks (32x32) experiment." into experimental
Yunqing Wang [Tue, 8 Jan 2013 20:59:08 +0000 (12:59 -0800)]
Merge "vp9_sub_pixel_variance16x2 SSE2 optimization" into experimental
Ronald S. Bultje [Tue, 8 Jan 2013 18:29:22 +0000 (10:29 -0800)]
Merge superblocks (32x32) experiment.
Change-Id: I0df99742029834a85c4933652b0587cf5b6b2587
Yunqing Wang [Tue, 8 Jan 2013 18:44:19 +0000 (10:44 -0800)]
vp9_sub_pixel_variance16x2 SSE2 optimization
About 5% decoder speedup.
Change-Id: Ib6687d337af758a536a0e7e289f400990f1f9794
John Koleszar [Tue, 8 Jan 2013 18:11:26 +0000 (10:11 -0800)]
Merge vp9-preview changes into experimental branch
Incorportate vp9-preview changes by merging master branch into experimental.
Conflicts:
test/test.mk
vp9/common/vp9_filter.c
vp9/common/vp9_idctllm.c
vp9/common/vp9_invtrans.h
vp9/common/vp9_mbpitch.c
vp9/common/vp9_rtcd_defs.sh
vp9/common/vp9_systemdependent.h
vp9/common/vp9_type_aliases.h
vp9/common/x86/vp9_asm_stubs.c
vp9/common/x86/vp9_subpixel_mmx.asm
vp9/decoder/vp9_decodframe.c
vp9/decoder/vp9_dequantize.c
vp9/decoder/vp9_dequantize.h
vp9/decoder/vp9_onyxd_int.h
vp9/encoder/vp9_bitstream.c
vp9/encoder/vp9_encodeframe.c
vp9/encoder/vp9_rdopt.c
Change-Id: I17f51c3666d1b59cf1a699f87607cbc5d30a87c5
Yaowu Xu [Fri, 14 Dec 2012 07:53:11 +0000 (23:53 -0800)]
reset segement map on key frame
This is to fix a decoder crash when decoder skips a number of frame to
continue decoding from a later key frame.
Change-Id: I3ba116eba6c3440e0528a21f53745f694302e4ad
Yaowu Xu [Tue, 8 Jan 2013 16:40:03 +0000 (08:40 -0800)]
Merge "minor loop filter refactoring and cleanup" into experimental
Yaowu Xu [Wed, 19 Dec 2012 19:34:49 +0000 (11:34 -0800)]
minor loop filter refactoring and cleanup
This commit did a couple of minor cleanup/refactoring to prepare for
futher loop filter experiments. It merged y_only version of loop filter
function into the regular one, which makes sure that same logic is used
for functions for picking level and for actual loop filtering.
Change-Id: Id10c94dccd45f58e5310bacfdf6ee63cbb60b86f
Ronald S. Bultje [Mon, 7 Jan 2013 19:41:49 +0000 (11:41 -0800)]
Remove a few redundant function arguments in encodeframe.c.
Also reindent a block of code that was misindented after addition of
the tx32x32 experiment.
Change-Id: Ic3e4aae3effd8a40136da68c9f382af03632ba08
Ronald S. Bultje [Mon, 7 Jan 2013 19:02:14 +0000 (11:02 -0800)]
Re-enable support for static_threshold (encode_breakout).
Change-Id: Ibd7380f478d3127f9db91d0a4fd2fd0dfde961ab
Ronald S. Bultje [Fri, 4 Jan 2013 20:05:40 +0000 (12:05 -0800)]
Don't use tx32x32 for macroblocks.
Change-Id: Ib674e0153ca360867ab7a20ba291ac9171a01250
Ronald S. Bultje [Sun, 6 Jan 2013 02:20:25 +0000 (18:20 -0800)]
64x64 blocksize support.
3.2% gains on std/hd, 1.0% gains on hd.
Change-Id: I481d5df23d8a4fc650a5bcba956554490b2bd200
Adrian Grange [Fri, 4 Jan 2013 17:00:47 +0000 (09:00 -0800)]
Fix mode selection infinite loop bug
Mode selection for SBs could enter an infinite loop because
the interpolation filter mode index was not being reset
correctly.
Change-Id: I4bbe726f29ef5b6836e94884067c46084713cc11
Paul Wilkins [Thu, 3 Jan 2013 15:14:36 +0000 (15:14 +0000)]
Further change to mv reference search.
This experimental change reorders the search so
that all possible references that match the target
reference frame are tested first and these in order
of distance from the current block. These will usually
be the highest scoring candidates.
If we do not find enough good candidates this way
we try non matching cases. These will usually be lower
scoring candidates.
The change in order together with breakouts when
we have found enough candidates should reduce
the computational cost and especially reduce the number
of sort operations.
Quality Results:
Std Hd +0.228%, Hd +0.074%, YT +0.046%, derf +0.137%
This effect is probably due to the fact that more distant
weak candidates are now less likely to get "promoted" over
near candidates even if they are repeated.
Change-Id: Iec37e77d88a48ad0ee1f315b14327a95d63f81f6
Marco Paniconi [Thu, 3 Jan 2013 23:47:34 +0000 (15:47 -0800)]
Use seg/ref/mode offsets in loop_filter_partial.
Use the proper seg/mode/ref filter offsets when selecting the
frame loop filter level for fast mode (pick_filter_level_fast).
Change-Id: I2473e2131c800ad19755cb6211ad735fecfe2ac0
Yaowu Xu [Thu, 3 Jan 2013 17:57:07 +0000 (09:57 -0800)]
Merge "make cost_coeffs() and tokenize_b() consistent" into experimental
Yaowu Xu [Thu, 3 Jan 2013 17:33:21 +0000 (09:33 -0800)]
Merge "Merge cost_coeffs_2x2() into cost_coeffs()" into experimental
Yaowu Xu [Thu, 3 Jan 2013 17:00:30 +0000 (09:00 -0800)]
make cost_coeffs() and tokenize_b() consistent
Change-Id: I7cdb5c32a1400f88ec36d08ea982e38b77731602
Adrian Grange [Thu, 20 Dec 2012 22:56:19 +0000 (14:56 -0800)]
New interpolation filter selection algorithm
Old Scheme:
When SWITCHABLE filter selection is enabled the encoder
evaluates the use of each interpolation filter type and
selects the best one to use at the MB level. A frame-
level flag can be set to force the use of a particular
filter type for all MBs in a frame if it is more efficient
to encode that way. The logic here involved a Q dependent
threshold that assumed that the second 8-tap filter was
a high-pass filter. However, this requires a trip around
the recode loop. If the frame-level flag indicates use
of a particular filter, the other filters are not
evaluated in the pick_mode loop.
New Scheme:
Each filter type is evaluated at the MB level and a record
of the best filter is kept, irrespective of what filter
is signaled at the frame-level. Once all MBs have been
encoded, a decision is made as to what frame-level mode
to set for the *next* frame. If one filter is used by 80%
or more of the MBs, then this filter is forced since it
is assumed that this will be more efficient if the
next frame has similar characteristics. i.e. there is a
one-frame lag between measuring the filter selection and
setting the frame-level mode to use.
Change-Id: I6a7e7ced8f27e120fafb99db2dc9c6293f8d20f7
Yaowu Xu [Thu, 3 Jan 2013 16:00:00 +0000 (08:00 -0800)]
Merge cost_coeffs_2x2() into cost_coeffs()
Remove special case function cost_coeffs_2x2() and change function
cost_coeffs() to handle 2nd order haar block as it is handle all
other block types already.
Change-Id: I2aac6f81ee0ae9e03d6a8da4f8681d69b79ce41f
Yunqing Wang [Wed, 2 Jan 2013 19:45:27 +0000 (11:45 -0800)]
Merge "Switch the order of calculating 2-D inverse transform" into experimental
Yunqing Wang [Wed, 2 Jan 2013 19:45:19 +0000 (11:45 -0800)]
Merge "Skip finding best ref_mvs when the mode is ZEROMV" into experimental
Paul Wilkins [Wed, 2 Jan 2013 18:06:00 +0000 (18:06 +0000)]
Change INT64_MAX to LLONG_MAX
This is needed to make the windows build work after
the removal of vp9_type_alisases.h.
Change-Id: I8addf38e9f3c8b864e0e30a8916a26e0264dd02c
Paul Wilkins [Mon, 10 Dec 2012 12:38:48 +0000 (12:38 +0000)]
Added update-able mv-ref probabilities.
Part of NEW_MVREF experiment.
Added update-able probabilities.
Change-Id: I5a4fcf4aaed1d0d1dac980f69d535639a3d59401
Yunqing Wang [Fri, 28 Dec 2012 00:04:44 +0000 (16:04 -0800)]
Skip finding best ref_mvs when the mode is ZEROMV
Read mode before calling vp9_find_best_ref_mvs(). If the mode is
ZEROMV, the best ref_mvs are not needed. Then, we can skip calling
vp9_find_best_ref_mvs().
Change-Id: I5baa3658dd3f1c7107211cbbbcf919b4584be2e2
Yunqing Wang [Thu, 27 Dec 2012 21:48:17 +0000 (13:48 -0800)]
Switch the order of calculating 2-D inverse transform
The 2-D inverse transform X = M1*Z*Transposed_M2 was calculated
in 2 steps from left to right:
1. Vertical transform: Y = M1*Z
2. Horizontal transform: X= Y*Transposed_M2
In SIMD, a transpose is needed in vertical transform.
Here, switched the calculation order to do it from right to left.
In this way, we could eliminate that transpose by writing the
intermediate results out to their transposed positions.
Change-Id: I34dfe5eb01292f6e363712420d99475e2e81e12c
John Koleszar [Fri, 21 Dec 2012 23:20:41 +0000 (15:20 -0800)]
Merge branch 'vp9-preview' of review:webm/libvpx
Merge the vp9-preview branch into master.
Change-Id: If700b9054676f24bed9deb59050af546c1ca5296
John Koleszar [Sun, 23 Dec 2012 15:20:10 +0000 (07:20 -0800)]
Build fixes to merge vp9-preview into master
Various fixups to resolve issues when building vp9-preview under the more stringent
checks placed on the experimental branch.
Change-Id: I21749de83552e1e75c799003f849e6a0f1a35b07
Yunqing Wang [Tue, 18 Dec 2012 20:09:46 +0000 (12:09 -0800)]
Fix a warning
Fixed the warning: the size of array ‘intermediate_buffer’ can’t
be evaluated [-Wvla].
Change-Id: Ibcffd6969bd71cee0c10f7cf18960e58cd0bd915
Scott LaVarnway [Fri, 21 Dec 2012 22:41:49 +0000 (14:41 -0800)]
Removed mmx versions of vp9_bilinear_predict filters
These filters will not work with VP9.
Change-Id: Ic26c77961084fcea6bfa97f4cd95afdea2282e85
John Koleszar [Fri, 21 Dec 2012 22:14:06 +0000 (14:14 -0800)]
Update CHANGELOG for v1.2.0 release
Change-Id: Ie0525d300c94bfde2947b1cc2572bb7d37e17998
John Koleszar [Fri, 21 Dec 2012 22:21:50 +0000 (14:21 -0800)]
Merge "add emmintrin_compat.h for builds with gcc < 4" into vp9-preview
John Koleszar [Fri, 21 Dec 2012 22:20:15 +0000 (14:20 -0800)]
Merge "configure: test for -msse4" into vp9-preview
Jim Bankoski [Fri, 21 Dec 2012 21:09:04 +0000 (13:09 -0800)]
fixed sizes of global arrays
Change-Id: Ibc077cf1c1da0c86063f88c6d3073c6876989119
John Koleszar [Thu, 20 Dec 2012 23:39:43 +0000 (15:39 -0800)]
Fix MSVS build for removed vp9/common/vp9_onyxd.h
Change-Id: I75ad0b4ca5b53b5bf759cc26a484ec196d275279
James Zern [Thu, 20 Dec 2012 22:46:55 +0000 (14:46 -0800)]
add emmintrin_compat.h for builds with gcc < 4
Change-Id: If7822e6fcd0d3568b934032322b19ba3e401df26
James Zern [Thu, 20 Dec 2012 22:05:19 +0000 (14:05 -0800)]
configure: test for -msse4
not present in < GCC 4.2
Change-Id: I7b1a3078fd782644b1c432ce682d2160ffc2263a
Jim Bankoski [Thu, 20 Dec 2012 17:40:18 +0000 (09:40 -0800)]
add private to assembly files to insure proper chromebuild
Change-Id: I6e43ca73f35401a974ed8ee27738d4318f09fd37
Deb Mukherjee [Tue, 27 Nov 2012 23:51:06 +0000 (15:51 -0800)]
New previous coef context experiment
Adds an experiment to derive the previous context of a coefficient
not just from the previous coefficient in the scan order but from a
combination of several neighboring coefficients previously encountered
in scan order. A precomputed table of neighbors for each location
for each scan type and block size is used. Currently 5 neighbors are
used.
Results are about 0.2% positive using a strategy where the max coef
magnitude from the 5 neigbors is used to derive the context.
Change-Id: Ie708b54d8e1898af742846ce2d1e2b0d89fd4ad5
Scott LaVarnway [Wed, 19 Dec 2012 22:13:25 +0000 (14:13 -0800)]
Disabled x86inc style assembly functions.... part 2
Missed a file
Change-Id: I33179de6755bc9eda9ad906e4fec6902ace435a5
John Koleszar [Wed, 19 Dec 2012 21:44:32 +0000 (13:44 -0800)]
make: fix dependency generation
Remove an extra level of escaping around the $@ variable to get valid output.
Prior to this change, modifying header files did not trigger a rebuild of
sources dependent on them.
Change-Id: I93ecc60371b705b64dc8a2583a5d31126fe3f851
John Koleszar [Fri, 16 Nov 2012 18:48:23 +0000 (10:48 -0800)]
Use boolcoder API instead of inlining
This patch changes the token packing to call the bool encoder API rather
than inlining it into the token packing function, and similarly removes
a special get_signed case from the detokenizer. This allows easier
experimentation with changing the bool coder as a whole.
Change-Id: I52c3625bbe4960b68cfb873b0e39ade0c82f9e91
Scott LaVarnway [Wed, 19 Dec 2012 19:53:43 +0000 (11:53 -0800)]
Disabled x86inc style assembly functions
Temporary fix for 32-bit mac build errors.
Change-Id: I2038f033cac16ea796097d0edd0f1c3da03246d7
Johann [Wed, 19 Dec 2012 00:30:47 +0000 (16:30 -0800)]
Add 'jni' to documention for ndk-build
ndk-build is intended to be launched from the application directory,
not the jni directory. Clarify the path in the comments.
Change-Id: Ie0faa89a13d967088a4fd2fd1c647962a9c04217
Ronald S. Bultje [Tue, 18 Dec 2012 23:31:19 +0000 (15:31 -0800)]
Use standard integer types for pixel values and coefficients.
For coefficients, use int16_t (instead of short); for pixel values in
16-bit intermediates, use uint16_t (instead of unsigned short); for all
others, use uint8_t (instead of unsigned char).
Change-Id: I3619cd9abf106c3742eccc2e2f5e89a62774f7da
Yaowu Xu [Tue, 18 Dec 2012 23:00:02 +0000 (15:00 -0800)]
Merge "Changed MAX_PSNR to 100" into experimental
Yaowu Xu [Tue, 18 Dec 2012 22:36:20 +0000 (14:36 -0800)]
Changed MAX_PSNR to 100
The MAX_PSNR was used to assign a "psnr" number when the mse is close
to zero. The direct assignment is used to prevent divide by zero in
computation. Changing it from 60 to 100 to be consistent against what
is being done in VP9
Change-Id: I4854ffc4961e59d372ec8005a0d52ca46e3c4c1a
Yaowu Xu [Tue, 18 Dec 2012 22:23:39 +0000 (14:23 -0800)]
Merge "correct logic in cnvcontext experiment for tx32x32" into experimental
Yaowu Xu [Tue, 18 Dec 2012 22:22:19 +0000 (14:22 -0800)]
Merge "Problem of over smoothing with intra modes." into vp9-preview
Ronald S. Bultje [Tue, 18 Dec 2012 22:17:46 +0000 (14:17 -0800)]
Merge "Give 4x4 scan and coef_band tables a _4x4 suffix." into experimental
Ronald S. Bultje [Tue, 18 Dec 2012 22:17:16 +0000 (14:17 -0800)]
Merge "Remove redundant "Prob" type (it's a duplicate of vp9_prob)." into experimental
Yaowu Xu [Tue, 18 Dec 2012 19:25:24 +0000 (11:25 -0800)]
correct logic in cnvcontext experiment for tx32x32
Change-Id: I004ded11983b7fda85793912ebc5c6f266dc5eb5
Yunqing Wang [Tue, 18 Dec 2012 20:58:18 +0000 (12:58 -0800)]
Fix uninitialized warning
Fixed uninitialized warning for txfm_size.
Change-Id: I42b7e802c3e84825d49f34e632361502641b7cbf
Yunqing Wang [Tue, 18 Dec 2012 20:09:46 +0000 (12:09 -0800)]
Fix a warning
Fixed the warning: the size of array ‘intermediate_buffer’ can’t
be evaluated [-Wvla].
Change-Id: Ibcffd6969bd71cee0c10f7cf18960e58cd0bd915
Ronald S. Bultje [Tue, 18 Dec 2012 18:49:10 +0000 (10:49 -0800)]
Give 4x4 scan and coef_band tables a _4x4 suffix.
This matches the names of tables for all other transform sizes.
Change-Id: Ia7681b7f8d34c97c27b0eb0e34d490cd0f8d02c6
Ronald S. Bultje [Tue, 18 Dec 2012 18:38:12 +0000 (10:38 -0800)]
Remove redundant "Prob" type (it's a duplicate of vp9_prob).
Change-Id: I9548891d7b8ff672a31579bcdce74e4cea529883
John Koleszar [Fri, 14 Dec 2012 20:35:33 +0000 (12:35 -0800)]
Remove vp9_type_aliases.h
Prefer the standard fixed-size integer typedefs.
Change-Id: Iad75582350669e49a8da3b7facb9c259e9514a5b
Yaowu Xu [Mon, 17 Dec 2012 19:10:16 +0000 (11:10 -0800)]
fixed a warning
where variable is used without initialization
Change-Id: Ic6b52623802641060cad4a72271050aeaf20ad5c
Paul Wilkins [Fri, 14 Dec 2012 17:49:46 +0000 (17:49 +0000)]
Problem of over smoothing with intra modes.
In some cases intra modes in inter frames give
an over smoothed appearance. Especially with
noisy but flat content.
Also in some cases there were problems with key
frame sizing again with very flat but noisy content.
These are temporary changes to help alleviate the
visual problems but will almost certainly hurt metric
results especially at the very low data rate end.
Change-Id: I11549179a19277ffc283d9788bc70168f2a8bdc9
Yaowu Xu [Fri, 14 Dec 2012 07:53:11 +0000 (23:53 -0800)]
reset segement map on key frame
This is to fix a decoder crash when decoder skips a number of frame to
continue decoding from a later key frame.
Change-Id: I3ba116eba6c3440e0528a21f53745f694302e4ad
Yaowu Xu [Fri, 14 Dec 2012 00:09:52 +0000 (16:09 -0800)]
prevents redefine of INT64_MAX
MSVC 2012 (_MSC_VER=1600) introduced the definition, this commit
prevents the redefinition of the macro
Change-Id: I7de92e7e9e865a342f2bcc4b071f8d3c9b2a508c
Yaowu Xu [Thu, 13 Dec 2012 23:57:31 +0000 (15:57 -0800)]
remove floating point inverse transforms
Change-Id: I9c651bd7c161974bf5f929446361b00d85e57a3f
Yaowu Xu [Thu, 13 Dec 2012 22:58:30 +0000 (14:58 -0800)]
fixed an encoder/decoder mismatch
The mismatch was caused by an improper merge of cleanup code around
tokenize_b() and stuff_b() with TX32X32 experiment.
Change-Id: I225ae62f015983751f017386548d9c988c30664c
Yaowu Xu [Thu, 13 Dec 2012 22:51:27 +0000 (14:51 -0800)]
fixed build issue with round()
not defined in msvc
Change-Id: I8fe8462a0c2f636d8b43c0243832ca67578f3665
Deb Mukherjee [Thu, 13 Dec 2012 20:18:38 +0000 (12:18 -0800)]
Build fixes with teh super blcoks and 32x32 expts
Change-Id: I3c751f8d57ac7d3b754476dc6ce144d162534e6d
Johann [Thu, 13 Dec 2012 19:35:59 +0000 (11:35 -0800)]
Unit tests do not build for iOS
Change-Id: I63423f8f2ec152cb490ddd9e6a94e431a11858f6
Deb Mukherjee [Thu, 13 Dec 2012 19:04:56 +0000 (11:04 -0800)]
Merge "Further improvements on the hybrid dwt/dct expt" into experimental
Deb Mukherjee [Wed, 12 Dec 2012 01:06:35 +0000 (17:06 -0800)]
Further improvements on the hybrid dwt/dct expt
Modifies the scanning pattern and uses a floating point 16x16
dct implementation for now to handle scaling better.
Also experiments are in progress with 2/6 and 9/7 wavelets.
Results have improved to within ~0.25% of 32x32 dct for std-hd
and about 0.03% for derf. This difference can probably be bridged by
re-optimizing the entropy stats for these transforms. Currently
the stats used are common between 32x32 dct and dwt/dct.
Experiments are in progress with various scan pattern - wavelet
combinations.
Ideally the subbands should be tokenized separately, and an
experiment will be condcuted next on that.
Change-Id: Ia9cbfc2d63cb7a47e562b2cd9341caf962bcc110
Ronald S. Bultje [Thu, 13 Dec 2012 17:56:50 +0000 (09:56 -0800)]
Merge "New default coefficient/band probabilities." into experimental
Ronald S. Bultje [Wed, 12 Dec 2012 18:25:58 +0000 (10:25 -0800)]
New default coefficient/band probabilities.
Gives 0.5-0.6% improvement on derf and stdhd, and 1.1% on hd. The
old tables basically derive from times that we had only 4x4 or
only 4x4 and 8x8 DCTs.
Note that some values are filled with 128, because e.g. ADST ever
only occurs as Y-with-DC, as does 32x32; 16x16 ever only occurs
as Y-with-DC or as UV (as complement of 32x32 Y); and 8x8 Y2 ever
only has 4 coefficients max. If preferred, I can add values of
other tables in their place (e.g. use 4x4 2nd order high-frequency
probabilities for 8x8 2nd order), so that they make at least some
sense if we ever implement a larger 2nd order transform for the
8x8 DCT (etc.), please let me know
Change-Id: I917db356f2aff8865f528eb873c56ef43aa5ce22
Scott LaVarnway [Wed, 12 Dec 2012 23:49:39 +0000 (15:49 -0800)]
Improved vp9_ihtllm_c
As suggested by Yaowu, we can use eob to reduce the complexity
of the vp9_ihtllm_c function. For the 1080p test clip used, the decoder
performance improved by 17%.
Change-Id: I32486f2f06f9b8f60467d2a574209aa3a3daa435
Ronald S. Bultje [Wed, 12 Dec 2012 18:34:14 +0000 (10:34 -0800)]
Merge "Consistently use get_prob(), clip_prob() and newly added clip_pixel()." into experimental
Ronald S. Bultje [Mon, 10 Dec 2012 20:09:07 +0000 (12:09 -0800)]
Consistently use get_prob(), clip_prob() and newly added clip_pixel().
Add a function clip_pixel() to clip a pixel value to the [0,255] range
of allowed values, and use this where-ever appropriate (e.g. prediction,
reconstruction). Likewise, consistently use the recently added function
clip_prob(), which calculates a binary probability in the [1,255] range.
If possible, try to use get_prob() or its sister get_binary_prob() to
calculate binary probabilities, for consistency.
Since in some places, this means that binary probability calculations
are changed (we use {255,256}*count0/(total) in a range of places,
and all of these are now changed to use 256*count0+(total>>1)/total),
this changes the encoding result, so this patch warrants some extensive
testing.
Change-Id: Ibeeff8d886496839b8e0c0ace9ccc552351f7628
Yaowu Xu [Tue, 11 Dec 2012 21:51:56 +0000 (13:51 -0800)]
Merge "clean up tokenize_b() and stuff_b()" into experimental
Yaowu Xu [Thu, 6 Dec 2012 20:40:57 +0000 (12:40 -0800)]
clean up tokenize_b() and stuff_b()
Change-Id: I0c1be01aae933243311ad321b6c456adaec1a0f5
Johann [Tue, 11 Dec 2012 19:47:49 +0000 (11:47 -0800)]
Revert "Upstream build bug for chromium"
This reverts commit
8bb82fded57f6316fdfd103d565050d0f1286bc5.
This is an incorrect workaround. It has been fixed in the GYP files
upstream.
Change-Id: If42f997747ce878b874508fdf7ae5a73a6fa1b2b
Scott LaVarnway [Tue, 11 Dec 2012 19:00:53 +0000 (11:00 -0800)]
Merge "Bug fix: use correct count_mb_ref_frame_usage"
Yaowu Xu [Tue, 11 Dec 2012 17:46:36 +0000 (09:46 -0800)]
Merge "experiment with CONTEXT conversion" into experimental
Scott LaVarnway [Tue, 11 Dec 2012 01:38:55 +0000 (17:38 -0800)]
Bug fix: use correct count_mb_ref_frame_usage
Change-Id: I9702f3e9ed664c2537e7874698c944620b07fff8
Scott LaVarnway [Tue, 11 Dec 2012 01:32:58 +0000 (17:32 -0800)]
Moved error_bins to macroblock struct
Change-Id: Ic9956ddf1c2ddffcf7be7fdfc23ad9a2426fc47a
WIP: Fixing unsafe threading in VP8 encoder.
Frank Galligan [Mon, 10 Dec 2012 23:50:48 +0000 (15:50 -0800)]
Fix ads2gas script to look for ALIGN as a word.
Change-Id: I4efc4f4e87e8666b69257de82c5c5dd4aadee28c
Scott LaVarnway [Tue, 11 Dec 2012 00:22:57 +0000 (16:22 -0800)]
Merge "Moved zbin_mode_boost to macroblock struct"
Scott LaVarnway [Mon, 10 Dec 2012 23:23:11 +0000 (15:23 -0800)]
Update correct macroblock quantize_b function ptrs
WIP: Fixing unsafe threading in VP8 encoder.
Use the passed in macroblock instead of the macroblock located in
cpi.
Change-Id: I1bfa07de6ea463f2baeaae1bae5d950691bc4afc
Scott LaVarnway [Mon, 10 Dec 2012 18:33:01 +0000 (10:33 -0800)]
Moved zbin_mode_boost to macroblock struct
Fixing unsafe threading in VP8 encoder.
Change-Id: Ibf4c89a2043654834747811bc11eb283de0bb830
John Koleszar [Mon, 10 Dec 2012 20:07:59 +0000 (12:07 -0800)]
configure: add --enable-external-build support
First attempt at avoiding all the compile-time environment detection for
cases where you can generate the environments statically, as when the
real build is being performed by another build system.
Change-Id: Ie3cf95d71d6c5169900f31e263b84bc123cdf73f
Deb Mukherjee [Mon, 10 Dec 2012 20:28:06 +0000 (12:28 -0800)]
Merge "A bug fix related to switchable filters" into experimental
Deb Mukherjee [Mon, 10 Dec 2012 20:10:36 +0000 (12:10 -0800)]
A bug fix related to switchable filters
The switchable count update was mistakenly inside a macro.
Change-Id: Iec04c52ad57034b88312dbaf05eee1f47ce265b3
Scott LaVarnway [Mon, 10 Dec 2012 18:51:42 +0000 (10:51 -0800)]
Moved zbin_over_quant to macroblock struct
Change-Id: I76fe20ade099573997404b8733cf7f79e82fb21e
WIP: Fixing unsafe threading in VP8 encoder.