platform/upstream/libvpx.git
11 years agoMerge "Constrain encoder motion search range"
Jingning Han [Mon, 18 Nov 2013 20:35:34 +0000 (12:35 -0800)]
Merge "Constrain encoder motion search range"

11 years agoConstrain encoder motion search range
Jingning Han [Sat, 16 Nov 2013 04:32:03 +0000 (20:32 -0800)]
Constrain encoder motion search range

Explicitly constrain the upper limit of motion search range (in the
unit of full pixel) to be [-1023, +1023]. It is intended to control
the effective motion search range for 4K sequences.

Change-Id: I645539c70885eec0f155781f439d97d333336e88

11 years agoMerge "Do horizontal loopfiltering in parallel"
Yunqing Wang [Mon, 18 Nov 2013 18:03:41 +0000 (10:03 -0800)]
Merge "Do horizontal loopfiltering in parallel"

11 years agoMerge "Add support for VC++2013"
Yaowu Xu [Mon, 18 Nov 2013 17:32:19 +0000 (09:32 -0800)]
Merge "Add support for VC++2013"

11 years agopartition context update speedup
Jim Bankoski [Sun, 17 Nov 2013 14:58:08 +0000 (06:58 -0800)]
partition context update speedup

This removes a lot of operations in setting partition context...

Change-Id: I365e6f5607ece85190cb21443988816dfa510ce3

11 years agovpxdec: Restore IVF support.
Tom Finegan [Sat, 16 Nov 2013 16:31:20 +0000 (08:31 -0800)]
vpxdec: Restore IVF support.

Refactored IVF frame reading code out into ivf_read_frame(). Forgot
to actually make the function call in read_frame().

Change-Id: Ie9f6917e70bd26d0352a761932465c60a29a1f81

11 years agoDo horizontal loopfiltering in parallel
Yunqing Wang [Wed, 13 Nov 2013 00:51:15 +0000 (16:51 -0800)]
Do horizontal loopfiltering in parallel

This patch followed "Rewrite filter_selectively_horiz for parallel
loopfiltering" commit, and added x86 SSE2 optimization to do
16-pixel filtering in parallel. Also, corrected the declaration
of aligned arrays. For 8-pixel-in-parallel case, improved the
calculation of the masks and filters. Updated the threshold loading
since the thresholds were already duplicated. Updated neon C functions
to call neon loopfilters twice.

Using tulip clip, tests showed it gave a ~1.5% decoder speed gain.

Change-Id: Id02638626ac27a4b0e0b09d71792a24c0499bd35

11 years agoMerge "Let the idct vp9_idct32x32_34_add = vp9_idct32x32_1024_add on arm until we...
hkuang [Fri, 15 Nov 2013 23:45:49 +0000 (15:45 -0800)]
Merge "Let the idct vp9_idct32x32_34_add = vp9_idct32x32_1024_add on arm until we implenment real vp9_idct32x32_34_add_neon."

11 years agoMerge "Renamed two files"
Yaowu Xu [Fri, 15 Nov 2013 23:20:23 +0000 (15:20 -0800)]
Merge "Renamed two files"

11 years agoLet the idct vp9_idct32x32_34_add = vp9_idct32x32_1024_add
hkuang [Fri, 15 Nov 2013 22:43:22 +0000 (14:43 -0800)]
Let the idct vp9_idct32x32_34_add = vp9_idct32x32_1024_add
on arm until we implenment real vp9_idct32x32_34_add_neon.

This issue is due to commit 47665452f0da3c11427ecb4852535e1787bb0c5b
Merge "Add 32x32 idct function for eob<=34 case".

Change-Id: I56b5f0abc20e7dd1bba521f78a995e85d65ea296

11 years agoMerge "Removing vp9_encodeintra.{h, c} files."
Dmitry Kovalev [Fri, 15 Nov 2013 21:38:20 +0000 (13:38 -0800)]
Merge "Removing vp9_encodeintra.{h, c} files."

11 years agoRenamed two files
Yaowu Xu [Fri, 15 Nov 2013 20:48:43 +0000 (12:48 -0800)]
Renamed two files

from vp9_decodframe.{c,h} to vp9_decodeframe.{c,h}

Change-Id: I21ac4b14fc90246e3f16bd90c52c12d126d791f8

11 years agoRemoving vp9_encodeintra.{h, c} files.
Dmitry Kovalev [Fri, 15 Nov 2013 20:17:16 +0000 (12:17 -0800)]
Removing vp9_encodeintra.{h, c} files.

There was only one function in *.c file, so moving it to vp9_encodemb.c.

Change-Id: I728859d08b3d6c05c33c1c5b21f0ea1d0e0f83af

11 years agoMerge "vpx[dec|enc]: Extract IVF support from the apps."
Tom Finegan [Fri, 15 Nov 2013 19:21:57 +0000 (11:21 -0800)]
Merge "vpx[dec|enc]: Extract IVF support from the apps."

11 years agoPrepare 1.3.0 release:
Paweł Hajdan [Fri, 15 Nov 2013 18:48:09 +0000 (10:48 -0800)]
Prepare 1.3.0 release:

- update .mailmap and AUTHORS
- update CHANGELOG

Change-Id: I3d75e8a6f87cabcbebc20e3ee5f3cd7db4d2e84d

11 years agoMerge "Cleaning up vp9_loopfilter.c file."
Dmitry Kovalev [Fri, 15 Nov 2013 18:10:59 +0000 (10:10 -0800)]
Merge "Cleaning up vp9_loopfilter.c file."

11 years agoMerge "Generate GAS output for Chromium"
Johann [Fri, 15 Nov 2013 17:17:37 +0000 (09:17 -0800)]
Merge "Generate GAS output for Chromium"

11 years agoMerge "Build iOS targets with XCode 5"
Johann [Fri, 15 Nov 2013 17:17:28 +0000 (09:17 -0800)]
Merge "Build iOS targets with XCode 5"

11 years agoMerge "Fix coding format in vp9_idct"
Jingning Han [Fri, 15 Nov 2013 16:59:14 +0000 (08:59 -0800)]
Merge "Fix coding format in vp9_idct"

11 years agoGenerate GAS output for Chromium
Johann [Wed, 13 Nov 2013 23:41:40 +0000 (15:41 -0800)]
Generate GAS output for Chromium

Upstream changes to account for differences in clang
syntax for Chromium iOS builds.

Since most of these are incompatible with XCode clang,
hide them behind a flag.

Change-Id: Idafcbcd4eb01b1ada6277da2d2edfd6c04b579fd

11 years agoBuild iOS targets with XCode 5
Johann [Wed, 13 Nov 2013 22:05:27 +0000 (14:05 -0800)]
Build iOS targets with XCode 5

Thanks to armando.dicianno for the patch

Fixes issue #603

Change-Id: I91ec030e5674b82c05115ed4dfbdb8ccbff22b66

11 years agovpx[dec|enc]: Extract IVF support from the apps.
Tom Finegan [Thu, 14 Nov 2013 20:37:42 +0000 (12:37 -0800)]
vpx[dec|enc]: Extract IVF support from the apps.

- Move IVF reading support into ivfdec.c and ivfdec.h
- Move IVF writing support into ivfenc.c and ivfenc.h
- Removed IVF writing code from the SVC example in favor of ivfenc.

Change-Id: I70adf6240d0320fdd232d8546ed573f0f68dd793

11 years agopartition plane context speed up
Jim Bankoski [Fri, 15 Nov 2013 16:00:43 +0000 (08:00 -0800)]
partition plane context speed up

Removes silly operations inside loop.

Change-Id: I9eeab1e914e715a887f86cf1089de508e2364165

11 years agoMerge "loop filter assert cleanout"
Jim Bankoski [Fri, 15 Nov 2013 15:48:36 +0000 (07:48 -0800)]
Merge "loop filter assert cleanout"

11 years agoMerge "Cleaning up vp9_tile_common.{h, c} files."
Dmitry Kovalev [Fri, 15 Nov 2013 04:55:01 +0000 (20:55 -0800)]
Merge "Cleaning up vp9_tile_common.{h, c} files."

11 years agoFix coding format in vp9_idct
Jingning Han [Fri, 15 Nov 2013 00:04:31 +0000 (16:04 -0800)]
Fix coding format in vp9_idct

Change-Id: If97ae16a4478717933345b6b9d5bc1b417b8dd84

11 years agofix scalling bug by buffer auto-reallocation
Adrian Grange [Thu, 14 Nov 2013 22:55:49 +0000 (14:55 -0800)]
fix scalling bug by buffer auto-reallocation

Change-Id: Ib748eb287520c794631697204da6ebe19523ce95

11 years agoCleaning up vp9_loopfilter.c file.
Dmitry Kovalev [Thu, 14 Nov 2013 23:04:14 +0000 (15:04 -0800)]
Cleaning up vp9_loopfilter.c file.

Change-Id: Ic6770072f80dfb54d2725ed96370d4f243a9f474

11 years agoCleaning up vp9_tile_common.{h, c} files.
Dmitry Kovalev [Thu, 14 Nov 2013 21:40:56 +0000 (13:40 -0800)]
Cleaning up vp9_tile_common.{h, c} files.

Change-Id: I9d18f351abe7614107f34f47eeb38a234a9937c9

11 years agoRemoving unused coefband_trans_8x8plus array from VP9Decompressor.
Dmitry Kovalev [Thu, 14 Nov 2013 21:37:18 +0000 (13:37 -0800)]
Removing unused coefband_trans_8x8plus array from VP9Decompressor.

Change-Id: Ic1367d767705377402ebfec0705f9f553a834400

11 years agoTweaks to vpx_test android make file
Joshua Litt [Thu, 14 Nov 2013 00:30:31 +0000 (16:30 -0800)]
Tweaks to vpx_test android make file

  * Change from thumb mode to arm mode improves test time significantly
  * Direct inclusion of test.mk allows for unit test configuration via
    configure script
Change-Id: Id58d3ba8289374528756a672459d8334afe20e2a

11 years agoloop filter assert cleanout
Jim Bankoski [Thu, 14 Nov 2013 20:25:32 +0000 (12:25 -0800)]
loop filter assert cleanout

Change-Id: I4e2ad4b7342681e6ac236356ef3a4927a54f105b

11 years agoAdd support for VC++2013
Yaowu Xu [Wed, 13 Nov 2013 19:36:32 +0000 (11:36 -0800)]
Add support for VC++2013

Change-Id: Iccc15aacee1fee9995dfd974bc89913891ad67a9

11 years agoMerge "Eliminating usage of txfrm_block_to_raster_block() from encode_block()."
Dmitry Kovalev [Thu, 14 Nov 2013 18:12:54 +0000 (10:12 -0800)]
Merge "Eliminating usage of txfrm_block_to_raster_block() from encode_block()."

11 years agoMerge "Enable 4x4 DCT/ADST transform unit test"
Jingning Han [Thu, 14 Nov 2013 17:26:33 +0000 (09:26 -0800)]
Merge "Enable 4x4 DCT/ADST transform unit test"

11 years agoMerge "For CBR, keep rate-correction damping factor to 2."
Marco Paniconi [Thu, 14 Nov 2013 16:11:42 +0000 (08:11 -0800)]
Merge "For CBR, keep rate-correction damping factor to 2."

11 years agoSimplifies band-getting with a static array
Deb Mukherjee [Thu, 14 Nov 2013 00:02:06 +0000 (16:02 -0800)]
Simplifies band-getting with a static array

Simplifies the code by implementing band mapping with static arrays.
A lot of the code complexity introduced in a previous patch
disappears.

Change-Id: Ia3fac36e594fb5ad2d55ae141c58bba4c55c2d28

11 years agoEliminating usage of txfrm_block_to_raster_block() from encode_block().
Dmitry Kovalev [Thu, 14 Nov 2013 03:33:12 +0000 (19:33 -0800)]
Eliminating usage of txfrm_block_to_raster_block() from encode_block().

Change-Id: I7d11f1b6075a1115cdc2dcd605225b9c9c9b39c7

11 years agoMerge "Cleaning up decode_coefs() function."
Dmitry Kovalev [Thu, 14 Nov 2013 02:39:14 +0000 (18:39 -0800)]
Merge "Cleaning up decode_coefs() function."

11 years agoEnable 4x4 DCT/ADST transform unit test
Jingning Han [Tue, 12 Nov 2013 20:47:32 +0000 (12:47 -0800)]
Enable 4x4 DCT/ADST transform unit test

This commit enables the  unit tests for 4x4 DCT and ADST transforms.
It covers tests of round-trip error check, coefficient match check,
coefficient overflow check, and inverse accuracy check.

Change-Id: Ibfea928ee48f0ebc088b7fdb0bf2d89a14161299

11 years agoCleaning up decode_coefs() function.
Dmitry Kovalev [Thu, 14 Nov 2013 00:46:21 +0000 (16:46 -0800)]
Cleaning up decode_coefs() function.

Removing vp9_read_and_apply_sign macro which was used only once.

Change-Id: I6a1625b720d89fc1291c99deccd6638b705f9b06

11 years agoFor CBR, keep rate-correction damping factor to 2.
Marco Paniconi [Thu, 14 Nov 2013 00:02:54 +0000 (16:02 -0800)]
For CBR, keep rate-correction damping factor to 2.

The switch to the rate-correction damping factor
in https://gerrit.chromium.org/gerrit/#/c/67536/ was not conditioned on CBR mode.

Change-Id: I2326704e8ac030a4f7b592dd3fedb94c7dd0644d

11 years agoMerge "Dual buffer encoding for intra modes"
Jingning Han [Wed, 13 Nov 2013 23:43:00 +0000 (15:43 -0800)]
Merge "Dual buffer encoding for intra modes"

11 years agoFix an overflow issue in SSE2 forward ADST
Jingning Han [Wed, 13 Nov 2013 22:48:38 +0000 (14:48 -0800)]
Fix an overflow issue in SSE2 forward ADST

The step that sums three input samples could potentially cause the
intermediate result go beyond 16 bit limit, when operating as the
second 1-D transform. This commit fixes the issue.

Change-Id: Iaf512449ac2d25ddd8a806d760afab362c62a516

11 years agoMerge "Replacing raster_block with block in the encoder."
Dmitry Kovalev [Wed, 13 Nov 2013 22:14:27 +0000 (14:14 -0800)]
Merge "Replacing raster_block with block in the encoder."

11 years agoMerge "Split macro strings on whitespace"
Johann [Wed, 13 Nov 2013 22:08:03 +0000 (14:08 -0800)]
Merge "Split macro strings on whitespace"

11 years agoMerge "Removing function pointers from inter prediction."
Dmitry Kovalev [Wed, 13 Nov 2013 21:54:15 +0000 (13:54 -0800)]
Merge "Removing function pointers from inter prediction."

11 years agoDual buffer encoding for intra modes
Jingning Han [Thu, 7 Nov 2013 22:56:58 +0000 (14:56 -0800)]
Dual buffer encoding for intra modes

Overall change (using dual buffer scheme for superblocks of both inter
and intra modes) reduces speed 2 runtime:
bluesky_1080p at 6000kbps:   263553ms -> 257441ms
riverbed_1080p at 8000kbps:  233230ms -> 225308ms.

Change-Id: Idf8d70f768a4b0d97b2a8506372c57b7b4022119

11 years agoMerge "Syncing write_modes_{b, sb} implementation with decode_modes_{b, sb}."
Dmitry Kovalev [Wed, 13 Nov 2013 18:47:46 +0000 (10:47 -0800)]
Merge "Syncing write_modes_{b, sb} implementation with decode_modes_{b, sb}."

11 years agoMerge "Optimizing set_contexts() function."
Dmitry Kovalev [Wed, 13 Nov 2013 18:01:05 +0000 (10:01 -0800)]
Merge "Optimizing set_contexts() function."

11 years agoMerge "Use 1D array to store super block filter levels"
Yunqing Wang [Wed, 13 Nov 2013 17:40:14 +0000 (09:40 -0800)]
Merge "Use 1D array to store super block filter levels"

11 years agoMerge "mips dsp-ase r2 vp9 decoder intra module optimizations (rebase)"
Johann [Wed, 13 Nov 2013 17:00:09 +0000 (09:00 -0800)]
Merge "mips dsp-ase r2 vp9 decoder intra module optimizations (rebase)"

11 years agomips dsp-ase r2 vp9 decoder intra module optimizations (rebase)
Parag Salasakar [Wed, 13 Nov 2013 05:47:14 +0000 (11:17 +0530)]
mips dsp-ase r2 vp9 decoder intra module optimizations (rebase)

Change-Id: Ib27fc4f3dbe01fe8adfa04a61aaba21b3480e75c

11 years agomips dsp-ase r2 vp9 decoder loopfilter module optimizations (rebase)
Parag Salasakar [Wed, 13 Nov 2013 05:23:16 +0000 (10:53 +0530)]
mips dsp-ase r2 vp9 decoder loopfilter module optimizations (rebase)

Change-Id: Ia7f640ca395e8deaac5986f19d11ab18d85eec2d

11 years agoMoving q_index from MACROBLOCKD to MACROBLOCK.
Dmitry Kovalev [Wed, 13 Nov 2013 01:28:27 +0000 (17:28 -0800)]
Moving q_index from MACROBLOCKD to MACROBLOCK.

Moving because q_index is used only by encoder.

Change-Id: I0b96175614ed4fd3d76ee56a0ba36258e1e896f6

11 years agoMerge "Enable dual buffer rd search and encoding scheme"
Jingning Han [Wed, 13 Nov 2013 02:11:41 +0000 (18:11 -0800)]
Merge "Enable dual buffer rd search and encoding scheme"

11 years agoMerge "Add test vector for row/col tile decoding"
Jingning Han [Wed, 13 Nov 2013 02:11:34 +0000 (18:11 -0800)]
Merge "Add test vector for row/col tile decoding"

11 years agoMerge "Calculating transform block offsets (x and y) only once."
Dmitry Kovalev [Wed, 13 Nov 2013 00:57:30 +0000 (16:57 -0800)]
Merge "Calculating transform block offsets (x and y) only once."

11 years agoMerge "Using max_tx_size instead of bsize when possible."
Dmitry Kovalev [Wed, 13 Nov 2013 00:54:30 +0000 (16:54 -0800)]
Merge "Using max_tx_size instead of bsize when possible."

11 years agoMerge "Moving {sb, mb, b, ab}_index from MACROBLOCKD to MACROBLOCK."
Dmitry Kovalev [Tue, 12 Nov 2013 23:59:28 +0000 (15:59 -0800)]
Merge "Moving {sb, mb, b, ab}_index from MACROBLOCKD to MACROBLOCK."

11 years agoMerge "Adding const to tree pointer inside vp9_extra_bit struct."
Dmitry Kovalev [Tue, 12 Nov 2013 23:48:07 +0000 (15:48 -0800)]
Merge "Adding const to tree pointer inside vp9_extra_bit struct."

11 years agoMerge "Added optimized vp9_idct32x32_34_add_dspr2"
Johann [Tue, 12 Nov 2013 23:30:00 +0000 (15:30 -0800)]
Merge "Added optimized vp9_idct32x32_34_add_dspr2"

11 years agoSplit macro strings on whitespace
Johann [Tue, 12 Nov 2013 20:26:45 +0000 (12:26 -0800)]
Split macro strings on whitespace

Match any whitespace instead of individual spaces. The macro
definitions in vp9/common/arm/neon/vp9_short_idct32x32_1_add_neon.asm
triggered this and treated spaces as arguments leading to lines like:
$8vld1$8.$88$8           {$8q8$8}, [$$89$8], $$8stride$8

Change-Id: I2d5718aba4614e4fd7b702e15c2a1bd80e656bd2

11 years agoAdding const to tree pointer inside vp9_extra_bit struct.
Dmitry Kovalev [Tue, 12 Nov 2013 22:21:15 +0000 (14:21 -0800)]
Adding const to tree pointer inside vp9_extra_bit struct.

Change-Id: I60e02fa3de930ff1f969687ab5af93dee40d86ad

11 years agoAndroid.mk file for vpx unittests
Joshua Litt [Mon, 11 Nov 2013 20:31:42 +0000 (12:31 -0800)]
Android.mk file for vpx unittests

These changes are to support automated regressions of vpx on android
new file:   test/android/Android.mk
new file:   test/android/README
new file:   test/android/get_files.py

Change-Id: I52c8e9daf3676a3561badbe710ec3a16fed72abd

11 years agoCalculating transform block offsets (x and y) only once.
Dmitry Kovalev [Mon, 11 Nov 2013 22:54:01 +0000 (14:54 -0800)]
Calculating transform block offsets (x and y) only once.

Change-Id: I4b5106bdc08fd4551339b968c13428a8f43122e2

11 years agoUse 1D array to store super block filter levels
Yunqing Wang [Tue, 12 Nov 2013 20:01:25 +0000 (12:01 -0800)]
Use 1D array to store super block filter levels

As Jim suggested, 1D array was used to store filter levels instead
of 2D array. This used shift_y in setup_mask directly, and saved
few cycles.

Change-Id: If61ab298784861f1806b1cd396d4e4e2e0f097b9

11 years agoAdd test vector for row/col tile decoding
Jingning Han [Mon, 11 Nov 2013 18:39:32 +0000 (10:39 -0800)]
Add test vector for row/col tile decoding

1920x1080

vp90-2-08-tile-4x4.webm
vp90-2-08-tile-4x4.webm.md5

vpxenc blue_sky_1080p25 --frame-parallel=0 \
--tile-columns=2 --tile-rows=2 \
--lossless=0 --error-resilient=0 \
--min-q=0 --max-q=30 --kf-min-dist=0 --kf-max-dist=9999 \
--cpu-used=0 --codec=vp9 -p 2 --good --fps=25/1 --limit=5

vp90-2-08-tile-4x1.webm
vp90-2-08-tile-4x1.webm.md5

vpxenc blue_sky_1080p25 --frame-parallel=0 \
--tile-columns=0 --tile-rows=2 \
--lossless=0 --error-resilient=0 \
--min-q=0 --max-q=30 --kf-min-dist=0 --kf-max-dist=9999 \
--cpu-used=0 --codec=vp9 -p 2 --good --fps=25/1 --limit=5

Change-Id: I77d0ccf9b721ae5cfa72781fe0d3c534ea43b756

11 years agoMerge "Removes conditional statements from band getting"
Deb Mukherjee [Tue, 12 Nov 2013 19:22:21 +0000 (11:22 -0800)]
Merge "Removes conditional statements from band getting"

11 years agoUse lowercase 'b' to branch
Johann [Tue, 12 Nov 2013 18:41:06 +0000 (10:41 -0800)]
Use lowercase 'b' to branch

iOS doesn't recognize B:
bad instruction `B idct32_pass_loop'

Change-Id: I3cf6aede4639f1d9efa97f7962fa287ba6feaaef

11 years agoMerge "Rewrite filter_selectively_horiz for parallel loopfiltering"
Yunqing Wang [Tue, 12 Nov 2013 18:20:49 +0000 (10:20 -0800)]
Merge "Rewrite filter_selectively_horiz for parallel loopfiltering"

11 years agoMerge "Improve loopfilter function"
Yunqing Wang [Tue, 12 Nov 2013 18:19:56 +0000 (10:19 -0800)]
Merge "Improve loopfilter function"

11 years agoRemoves conditional statements from band getting
Deb Mukherjee [Wed, 6 Nov 2013 01:25:38 +0000 (17:25 -0800)]
Removes conditional statements from band getting

Implements scan order to band map with arrays in both the encoder
and decoder to remove conditional statements.

Encoding seems to be about 1% faster at speed 0, tested on football.
Decoding seems to be about 0.5-1% faster on a set of 25 videos.

Change-Id: Idb233ca0b9e0efd790e30880642e8717e1c5c8dd

11 years agoMerge "Removing redundant assignment."
Dmitry Kovalev [Tue, 12 Nov 2013 02:37:22 +0000 (18:37 -0800)]
Merge "Removing redundant assignment."

11 years agoRemoving function pointers from inter prediction.
Dmitry Kovalev [Tue, 12 Nov 2013 02:37:00 +0000 (18:37 -0800)]
Removing function pointers from inter prediction.

Removing foreach_predicted_block_visitor and calling build_inter_predictors
directly.

Change-Id: I11bb3c872b99b47c2680b01b0dbcc01c558c4a2b

11 years agoEnable dual buffer rd search and encoding scheme
Jingning Han [Wed, 6 Nov 2013 05:07:08 +0000 (21:07 -0800)]
Enable dual buffer rd search and encoding scheme

This commit enables the dual buffer rate-distortion optimization
and encoding scheme. It stacks the original transform coefficients,
quantized levels, and reconstructed coefficients, in the rate-
distortion optimization search process, hence eliminates the need
to re-run residual generation, forward transform, and quantization
in the encoding stage.

Change-Id: I011bfad3a59a380a869ee552e91dae0394ec492e

11 years agoMerge "Allocate dual buffer sets for encoding"
Jingning Han [Tue, 12 Nov 2013 02:00:57 +0000 (18:00 -0800)]
Merge "Allocate dual buffer sets for encoding"

11 years agoRemoving redundant assignment.
Dmitry Kovalev [Tue, 12 Nov 2013 01:39:43 +0000 (17:39 -0800)]
Removing redundant assignment.

xd->mi_8x8 is assigned inside set_offsets() for each prediction block.

Change-Id: I20e5974a9eaf105e5a04fc7f99b7a93bd50e3d0a

11 years agoSyncing write_modes_{b, sb} implementation with decode_modes_{b, sb}.
Dmitry Kovalev [Tue, 12 Nov 2013 01:29:31 +0000 (17:29 -0800)]
Syncing write_modes_{b, sb} implementation with decode_modes_{b, sb}.

Change-Id: Iaee740ec3bfb2b5328c24f4641c285e5a4a046dc

11 years agoRewrite filter_selectively_horiz for parallel loopfiltering
Yunqing Wang [Sat, 9 Nov 2013 00:12:04 +0000 (16:12 -0800)]
Rewrite filter_selectively_horiz for parallel loopfiltering

Added loop filter mask checking, and made the caller function
ready for implementation of parallel loopfiltering in horizontal
direction.

Next, we need to go through the loopfilter functions (both c and
optimized versions), and provide 16-byte wide loopfiltering for
each filter type.

Change-Id: Ifef47e7ef9086ebc2fd6ca7ede8f27c9bbf79e66

11 years agoMerge "vpxenc: clean up; move stats code out of vpxenc.c"
Tom Finegan [Tue, 12 Nov 2013 01:01:57 +0000 (17:01 -0800)]
Merge "vpxenc: clean up; move stats code out of vpxenc.c"

11 years agoMerge "Cleaning up joint_motion_search function."
Dmitry Kovalev [Tue, 12 Nov 2013 00:34:39 +0000 (16:34 -0800)]
Merge "Cleaning up joint_motion_search function."

11 years agovpxenc: clean up; move stats code out of vpxenc.c
Tom Finegan [Fri, 8 Nov 2013 05:28:45 +0000 (21:28 -0800)]
vpxenc: clean up; move stats code out of vpxenc.c

Change-Id: I6b245ef5017c49373ba33e30f8fbf3d1b97409b5

11 years agoAllocate dual buffer sets for encoding
Jingning Han [Fri, 1 Nov 2013 19:53:37 +0000 (12:53 -0700)]
Allocate dual buffer sets for encoding

Allocate memory space of dual buffer sets that store the coeff, qcoeff,
dqcoeff, and eobs. Connect the pointers of macroblock_plane and
macroblockd_plane to the actual buffer in use accordingly.

Change-Id: I2f0b5f482ca879fae39095013eaf8901db20a5a4

11 years agoMerge "Decouple macroblockd_plane buffer usage"
Jingning Han [Tue, 12 Nov 2013 00:09:16 +0000 (16:09 -0800)]
Merge "Decouple macroblockd_plane buffer usage"

11 years agoCleaning up joint_motion_search function.
Dmitry Kovalev [Tue, 12 Nov 2013 00:04:02 +0000 (16:04 -0800)]
Cleaning up joint_motion_search function.

Change-Id: I70a0878b23bda0ac3ff8733b4c96d5c636bc551c

11 years agoMoving {sb, mb, b, ab}_index from MACROBLOCKD to MACROBLOCK.
Dmitry Kovalev [Mon, 11 Nov 2013 23:58:57 +0000 (15:58 -0800)]
Moving {sb, mb, b, ab}_index from MACROBLOCKD to MACROBLOCK.

We use {sb, mb, b, ab}_index only inside encoder, so moving them into
appropriate data structure.

Change-Id: Ib5c1036716354d9d321e11a60c1634c1cb8f9716

11 years agoMerge "Fixed memory leak with svc_test.cc"
Ivan Maltz [Mon, 11 Nov 2013 23:52:42 +0000 (15:52 -0800)]
Merge "Fixed memory leak with svc_test.cc"

11 years agoDecouple macroblockd_plane buffer usage
Jingning Han [Thu, 31 Oct 2013 19:21:49 +0000 (12:21 -0700)]
Decouple macroblockd_plane buffer usage

Make the macroblockd_plane contain dynamic buffer pointers instead
static pointers to the memory space allocated therein. The decoder
uses the buffer allocated in pbi, while encoder will use a dual
buffer approach for rate-distortion optimization search.

Change-Id: Ie6f24be2dcda35df7c15b4014e5ccf236fb3f76c

11 years agoFixed memory leak with svc_test.cc
Ivan Maltz [Mon, 11 Nov 2013 21:55:34 +0000 (13:55 -0800)]
Fixed memory leak with svc_test.cc

was not calling vpx_codec_destroy and delete(decoder_) in TearDown

Change-Id: Iff4fd24a260223d224b3ea3287cdf0227405492f

11 years agoReplacing raster_block with block in the encoder.
Dmitry Kovalev [Mon, 11 Nov 2013 23:18:48 +0000 (15:18 -0800)]
Replacing raster_block with block in the encoder.

We only used "ib" to call get_scan() function, which in turn calls
get_tx_type_4x4() function. The latter one only needs block index if
bsize < BLOCK_8X8 -- under that condition raster_block == block.

Change-Id: I697306a0c3cf937acdd4f5e623d4367c5acc0b2f

11 years agoFix a bug in the assembly code.
hkuang [Mon, 11 Nov 2013 20:49:12 +0000 (12:49 -0800)]
Fix a bug in the assembly code.

Change-Id: Ic416e3f8a11e82ee298e6f709b2119a9ddf1e2f8

11 years agoMerge "Replacing (raster_block >> tx_size) with (block >> (tx_size << 1))."
Dmitry Kovalev [Mon, 11 Nov 2013 20:27:42 +0000 (12:27 -0800)]
Merge "Replacing (raster_block >> tx_size) with (block >> (tx_size << 1))."

11 years agoMerge "vpx[dec|enc]: Clean up target OS based IO focused preproc abuse."
Tom Finegan [Mon, 11 Nov 2013 19:53:33 +0000 (11:53 -0800)]
Merge "vpx[dec|enc]: Clean up target OS based IO focused preproc abuse."

11 years agovpx[dec|enc]: Clean up target OS based IO focused preproc abuse.
Tom Finegan [Mon, 11 Nov 2013 14:48:18 +0000 (06:48 -0800)]
vpx[dec|enc]: Clean up target OS based IO focused preproc abuse.

Relocate it to tools_common.h so we can stop duping this code
everywhere.

BUG=https://code.google.com/p/webm/issues/detail?id=660

Change-Id: If8aa49b74d297273cd9fd7006b2767837055a359

11 years agoMerge "Localizing NEARESTMV special cases in the code."
Dmitry Kovalev [Mon, 11 Nov 2013 19:12:06 +0000 (11:12 -0800)]
Merge "Localizing NEARESTMV special cases in the code."

11 years agoMerge "Cleaning up vp9_quantize_b_c() function."
Dmitry Kovalev [Mon, 11 Nov 2013 19:02:32 +0000 (11:02 -0800)]
Merge "Cleaning up vp9_quantize_b_c() function."

11 years agoMerge "Cleaning up read_mv_probs() function."
Dmitry Kovalev [Mon, 11 Nov 2013 19:01:35 +0000 (11:01 -0800)]
Merge "Cleaning up read_mv_probs() function."

11 years agoMerge "Adding read_reference_mode() function."
Dmitry Kovalev [Mon, 11 Nov 2013 19:00:51 +0000 (11:00 -0800)]
Merge "Adding read_reference_mode() function."

11 years agoMerge "vpxenc: Remove support for IVF input."
Tom Finegan [Mon, 11 Nov 2013 15:57:42 +0000 (07:57 -0800)]
Merge "vpxenc: Remove support for IVF input."