]> granicus.if.org Git - libvpx/log
libvpx
8 years agoMerge "av1/common/filter.h: apply clang-format" into nextgenv2
Yaowu Xu [Wed, 26 Oct 2016 23:43:02 +0000 (23:43 +0000)]
Merge "av1/common/filter.h: apply clang-format" into nextgenv2

8 years agoMerge "Fix aom_fdct32x32_avx2 output as CONFIG_AOM_HIGHBITDEPTH=1" into nextgenv2
Yi Luo [Wed, 26 Oct 2016 22:42:16 +0000 (22:42 +0000)]
Merge "Fix aom_fdct32x32_avx2 output as CONFIG_AOM_HIGHBITDEPTH=1" into nextgenv2

8 years agoMerge "Synchronize tx_size counts in the decoder" into nextgenv2
Jingning Han [Wed, 26 Oct 2016 21:46:18 +0000 (21:46 +0000)]
Merge "Synchronize tx_size counts in the decoder" into nextgenv2

8 years agoMerge "Fix incorrect merge of forward txfm function declarations" into nextgenv2
Yi Luo [Wed, 26 Oct 2016 19:22:52 +0000 (19:22 +0000)]
Merge "Fix incorrect merge of forward txfm function declarations" into nextgenv2

8 years agoMerge changes I56cddcb4,I40c5a652 into nextgenv2
Urvang Joshi [Wed, 26 Oct 2016 18:28:56 +0000 (18:28 +0000)]
Merge changes I56cddcb4,I40c5a652 into nextgenv2

* changes:
  Mark bogus palette color probabilities as zero
  get_palette_color_context: code cleanup

8 years agoSynchronize tx_size counts in the decoder
Jingning Han [Wed, 26 Oct 2016 18:04:31 +0000 (11:04 -0700)]
Synchronize tx_size counts in the decoder

Make both encoder and decoder use depth index for frame count.

Change-Id: I96dddffc0a83ad5e4e2847b15391e01ba01ee502

8 years agoMerge "av1/convolve.[hc],av1_convolve_test: add missing copyright" into nextgenv2
Angie Chiang [Wed, 26 Oct 2016 17:51:28 +0000 (17:51 +0000)]
Merge "av1/convolve.[hc],av1_convolve_test: add missing copyright" into nextgenv2

8 years agoMerge "Use has_subpel_mv_component in av1_is_interp_needed" into nextgenv2
Angie Chiang [Wed, 26 Oct 2016 17:50:54 +0000 (17:50 +0000)]
Merge "Use has_subpel_mv_component in av1_is_interp_needed" into nextgenv2

8 years agoMerge "Revise precision clamping in GM param refinement" into nextgenv2
Sarah Parker [Wed, 26 Oct 2016 17:33:47 +0000 (17:33 +0000)]
Merge "Revise precision clamping in GM param refinement" into nextgenv2

8 years agoFix incorrect merge of forward txfm function declarations
Yi Luo [Tue, 25 Oct 2016 23:13:07 +0000 (16:13 -0700)]
Fix incorrect merge of forward txfm function declarations

- Restore the fwd txfm HBD function declarations exposure.

Change-Id: I1e33df6297fd37e242f4b73c8ab97063b9feb7c6

8 years agoConvert tx_size to relative depth to fetch tx_size_cost
Jingning Han [Tue, 25 Oct 2016 23:28:49 +0000 (16:28 -0700)]
Convert tx_size to relative depth to fetch tx_size_cost

Use the relative transform partition depth as index to fetch the
tx_size_cost value.

Change-Id: I7d5119817baa96f23c32828065ff3175bb9f75cf

8 years agoAllow backward probability model update from tx_size=0
Jingning Han [Tue, 25 Oct 2016 23:15:02 +0000 (16:15 -0700)]
Allow backward probability model update from tx_size=0

Replace enum items with range definitions.

Change-Id: Iba2b7cac657db5fb6177cb5c9e6f40ec0125d926

8 years agoMerge "Add depth to tx_size mapper to bit-stream coding" into nextgenv2
Jingning Han [Wed, 26 Oct 2016 17:11:56 +0000 (17:11 +0000)]
Merge "Add depth to tx_size mapper to bit-stream coding" into nextgenv2

8 years agoMerge "Simplify interpolation filter search in handle_inter_mode()" into nextgenv2
Angie Chiang [Wed, 26 Oct 2016 16:51:57 +0000 (16:51 +0000)]
Merge "Simplify interpolation filter search in handle_inter_mode()" into nextgenv2

8 years agoMerge "Change 2 motion search counts to be tile data" into nextgenv2
Yunqing Wang [Wed, 26 Oct 2016 16:17:42 +0000 (16:17 +0000)]
Merge "Change 2 motion search counts to be tile data" into nextgenv2

8 years agoav1/common/filter.h: apply clang-format
Yaowu Xu [Wed, 26 Oct 2016 16:05:58 +0000 (09:05 -0700)]
av1/common/filter.h: apply clang-format

Change-Id: I37f0d1fbcc6f262ae287290e2e6f5648ad0113c8

8 years agoAdd depth to tx_size mapper to bit-stream coding
Jingning Han [Tue, 25 Oct 2016 23:05:02 +0000 (16:05 -0700)]
Add depth to tx_size mapper to bit-stream coding

It serves as a helper function to integrate various transform coding
options.

Change-Id: I64e7d0c88ea10137fa1ff1072d865eb0054c2a25

8 years agoRevise precision clamping in GM param refinement
Sarah Parker [Tue, 25 Oct 2016 21:42:47 +0000 (14:42 -0700)]
Revise precision clamping in GM param refinement

This ensures that the parameter refinement never
results in a motion parameter value that exceeds the number
of alloted bits in the bitstream. It accounts for all of
the necessary precision shifts required to make global motion compatible
with the warped motion library. It also accounts for the
zero-centering that is applied to global motion parameters that are
naturally centered around one.
Change-Id: If429c93bb90b66fdff0edc07ecd9fc078077d303

8 years agoMerge "Refactor transform size coding" into nextgenv2
Jingning Han [Wed, 26 Oct 2016 01:12:03 +0000 (01:12 +0000)]
Merge "Refactor transform size coding" into nextgenv2

8 years agoUse has_subpel_mv_component in av1_is_interp_needed
Angie Chiang [Tue, 25 Oct 2016 23:56:48 +0000 (16:56 -0700)]
Use has_subpel_mv_component in av1_is_interp_needed

Change-Id: I8980df4512de605aaa6a67c1f05e544f69a12e96

8 years agoSimplify interpolation filter search in handle_inter_mode()
Angie Chiang [Tue, 25 Oct 2016 19:19:16 +0000 (12:19 -0700)]
Simplify interpolation filter search in handle_inter_mode()

                  BDRate
ext_interp  lowres -0.001%
dual_filter lowres  0.001%

Change-Id: Ic24165d554c300eaa0188ee8cb88d320b74125aa

8 years agoav1/convolve.[hc],av1_convolve_test: add missing copyright
Angie Chiang [Tue, 25 Oct 2016 22:09:56 +0000 (15:09 -0700)]
av1/convolve.[hc],av1_convolve_test: add missing copyright

Change-Id: Ie84bdf90c31b12977d32baacfc8086c1fdd96e65

8 years agoRefactor transform size coding
Jingning Han [Tue, 25 Oct 2016 22:35:29 +0000 (15:35 -0700)]
Refactor transform size coding

Introduce the transform block partition depth macro definition.

Change-Id: I218dc77a77c8e967da4d270d4ec0d7691b712a5f

8 years agoMerge "Refactor tx_size use case in block encoding stage" into nextgenv2
Jingning Han [Tue, 25 Oct 2016 22:29:20 +0000 (22:29 +0000)]
Merge "Refactor tx_size use case in block encoding stage" into nextgenv2

8 years agoMerge "Refactor tokenize_vartx to use aligned transform block size fetch" into nextgenv2
Jingning Han [Tue, 25 Oct 2016 22:29:02 +0000 (22:29 +0000)]
Merge "Refactor tokenize_vartx to use aligned transform block size fetch" into nextgenv2

8 years agoChange 2 motion search counts to be tile data
Yunqing Wang [Tue, 25 Oct 2016 22:15:23 +0000 (15:15 -0700)]
Change 2 motion search counts to be tile data

Imported changes from VP9:
https://chromium-review.googlesource.com/#/c/402551/
https://chromium-review.googlesource.com/#/c/403128/

Change-Id: I8570c867190a6fa641926431ce97f7d9d7da3528

8 years agoMerge "Use table fetch for block width in block_rd_txfm" into nextgenv2
Jingning Han [Tue, 25 Oct 2016 22:18:43 +0000 (22:18 +0000)]
Merge "Use table fetch for block width in block_rd_txfm" into nextgenv2

8 years agoMerge "update_state: quiet const warning w/global-motion" into nextgenv2
James Zern [Tue, 25 Oct 2016 22:15:39 +0000 (22:15 +0000)]
Merge "update_state: quiet const warning w/global-motion" into nextgenv2

8 years agoFix aom_fdct32x32_avx2 output as CONFIG_AOM_HIGHBITDEPTH=1
Yi Luo [Mon, 24 Oct 2016 23:30:55 +0000 (16:30 -0700)]
Fix aom_fdct32x32_avx2 output as CONFIG_AOM_HIGHBITDEPTH=1

- Change FDCT32x32_2D_AVX2 output parameter to tran_low_t.
- Add unit tests for CONFIG_AOM_HIGHBITDEPTH=1.
- Update TODO notes.
BUG=webm:1323

Change-Id: If4766c919a24231fce886de74658b6dd7a011246

8 years agoVertical scalers: Use signed int for src/dst pitch in parameters.
Urvang Joshi [Tue, 25 Oct 2016 20:00:22 +0000 (13:00 -0700)]
Vertical scalers: Use signed int for src/dst pitch in parameters.

This avoids explicitly casting them to 'int' later.
These methods were already called with signed int arguments for pitch,
so this also avoids int -> unsigned int -> int conversion.

Change-Id: I2129f5ceff8f2525a188ee3ae52f9fe7067bd2e3

8 years agoMerge "Fix unsigned type error in gen_scaler.c" into nextgenv2
Angie Chiang [Tue, 25 Oct 2016 19:43:09 +0000 (19:43 +0000)]
Merge "Fix unsigned type error in gen_scaler.c" into nextgenv2

8 years agoMerge "dkboolwriter.c: change copyright notice" into nextgenv2
Yaowu Xu [Tue, 25 Oct 2016 19:41:17 +0000 (19:41 +0000)]
Merge "dkboolwriter.c: change copyright notice" into nextgenv2

8 years agoMerge "7-bit interpolation filters" into nextgenv2
Yaowu Xu [Tue, 25 Oct 2016 19:41:07 +0000 (19:41 +0000)]
Merge "7-bit interpolation filters" into nextgenv2

8 years agoMerge "Use constrained tokenset with --enable-daala_ec." into nextgenv2
Yaowu Xu [Tue, 25 Oct 2016 19:40:51 +0000 (19:40 +0000)]
Merge "Use constrained tokenset with --enable-daala_ec." into nextgenv2

8 years agoMerge "Refactor dist_block() function" into nextgenv2
Jingning Han [Tue, 25 Oct 2016 19:39:22 +0000 (19:39 +0000)]
Merge "Refactor dist_block() function" into nextgenv2

8 years agoMerge changes I6faedb29,Ic6586114 into nextgenv2
Angie Chiang [Tue, 25 Oct 2016 19:36:49 +0000 (19:36 +0000)]
Merge changes I6faedb29,Ic6586114 into nextgenv2

* changes:
  Remove speed feature of ext_interp experiment
  Refactor: handle_inter()

8 years agoRefactor tx_size use case in block encoding stage
Jingning Han [Tue, 25 Oct 2016 19:35:43 +0000 (12:35 -0700)]
Refactor tx_size use case in block encoding stage

Change-Id: I56110d1fc94b335668e6b991442e9083bbaea8ee

8 years agoRefactor tokenize_vartx to use aligned transform block size fetch
Jingning Han [Tue, 25 Oct 2016 18:57:46 +0000 (11:57 -0700)]
Refactor tokenize_vartx to use aligned transform block size fetch

This prepares for the integration of rectangular transform size
into recursive transform block partition.

Change-Id: I164eb43d10afa9bb2f4722de7a48faa770ba4ced

8 years agoMerge "Refactor tx_size use cases in blockd.c" into nextgenv2
Jingning Han [Tue, 25 Oct 2016 19:03:29 +0000 (19:03 +0000)]
Merge "Refactor tx_size use cases in blockd.c" into nextgenv2

8 years agoUse table fetch for block width in block_rd_txfm
Jingning Han [Tue, 25 Oct 2016 17:37:34 +0000 (10:37 -0700)]
Use table fetch for block width in block_rd_txfm

Make direct use of block_size_wide to fetch data for stride.

Change-Id: I0d8491e58cf00ea73c764d218cb56408b64d9ee7

8 years agodkboolwriter.c: change copyright notice
Yaowu Xu [Tue, 25 Oct 2016 17:32:33 +0000 (10:32 -0700)]
dkboolwriter.c: change copyright notice

Change-Id: I1d9349a07ffd85991fc5673354d3ceff3404b358

8 years agoRefactor dist_block() function
Jingning Han [Tue, 25 Oct 2016 17:15:39 +0000 (10:15 -0700)]
Refactor dist_block() function

Support automatic scale for mapping between transform block size
and pixel block size.

Change-Id: I141b0477a85c0dcc5f99b4e5d880cfccfae6d316

8 years ago7-bit interpolation filters
Arild Fuldseth [Fri, 26 Aug 2016 12:08:58 +0000 (14:08 +0200)]
7-bit interpolation filters

Purpose:
-Reduce dynamic range of interpolation filter coefficents from 8
bits to 7 bits.
-Inner product for 8-bit input data can be stored in a 16-bit signed
integer.

Impact on compression efficiency:
-Marginal improvement, typically less than 0.5% BDR.

Change-Id: I58d1408307ae7d2a6f9de8965c5877b258703199

8 years agoMerge "Refactor extrabits packing" into nextgenv2
Yaowu Xu [Tue, 25 Oct 2016 17:14:56 +0000 (17:14 +0000)]
Merge "Refactor extrabits packing" into nextgenv2

8 years agoMerge "Linearize extrabits writing." into nextgenv2
Yaowu Xu [Tue, 25 Oct 2016 17:14:43 +0000 (17:14 +0000)]
Merge "Linearize extrabits writing." into nextgenv2

8 years agoUse constrained tokenset with --enable-daala_ec.
Nathan E. Egge [Wed, 24 Aug 2016 18:52:42 +0000 (14:52 -0400)]
Use constrained tokenset with --enable-daala_ec.

Change-Id: Ia09edf92bf9f7ecacc65c232ac6e656cde236634

8 years agoRefactor tx_size use cases in blockd.c
Jingning Han [Tue, 25 Oct 2016 16:47:02 +0000 (09:47 -0700)]
Refactor tx_size use cases in blockd.c

Use table to replace the arithmetic computation for mapping between
transform block and pixel number. Support automatic scale of block
size and transform block size.

Change-Id: I84766850172265d4295f418383dbc5e6e5838ec8

8 years agoMerge "Fix compile error with --enable-accounting" into nextgenv2
Debargha Mukherjee [Tue, 25 Oct 2016 16:25:42 +0000 (16:25 +0000)]
Merge "Fix compile error with --enable-accounting" into nextgenv2

8 years agoMerge "Refactor: Add macro LOG_SWITCHABLE_FILTERS" into nextgenv2
Angie Chiang [Tue, 25 Oct 2016 16:24:23 +0000 (16:24 +0000)]
Merge "Refactor: Add macro LOG_SWITCHABLE_FILTERS" into nextgenv2

8 years agoRemove speed feature of ext_interp experiment
Angie Chiang [Mon, 24 Oct 2016 23:04:26 +0000 (16:04 -0700)]
Remove speed feature of ext_interp experiment

This is to facilitate the refactor process

Change-Id: I6faedb29129b47abefe20821dc3f32a43db149d8

8 years agoRefactor: Add macro LOG_SWITCHABLE_FILTERS
Angie Chiang [Mon, 24 Oct 2016 19:24:44 +0000 (12:24 -0700)]
Refactor: Add macro LOG_SWITCHABLE_FILTERS

Change-Id: I7593ff2f8949d8bc26ca1c8577faaefb09640b59

8 years agoRefactor: handle_inter()
Angie Chiang [Mon, 24 Oct 2016 22:57:36 +0000 (15:57 -0700)]
Refactor: handle_inter()

Make the parenthesis symmetric
Replace interpolation filter mode number by macro

Change-Id: Ic6586114c4cebe920b950e1b3adc8ebc764d4713

8 years agoMerge "Fix to make intra_only frames decodable out of order" into nextgenv2
Debargha Mukherjee [Tue, 25 Oct 2016 16:21:20 +0000 (16:21 +0000)]
Merge "Fix to make intra_only frames decodable out of order" into nextgenv2

8 years agoFix unsigned type error in gen_scaler.c
Angie Chiang [Tue, 25 Oct 2016 16:18:22 +0000 (09:18 -0700)]
Fix unsigned type error in gen_scaler.c

Avoid applying unary minus operator on unsigned type

Change-Id: Ibc60541837eef06810f5be0aaa7fef9edcc8f8a4

8 years agoMerge "update_state_supertx: rename a shadowed variable" into nextgenv2
Yaowu Xu [Tue, 25 Oct 2016 16:04:32 +0000 (16:04 +0000)]
Merge "update_state_supertx: rename a shadowed variable" into nextgenv2

8 years agoMerge "Move small fixes and refactoring for obmc pred from AV1" into nextgenv2
Yaowu Xu [Tue, 25 Oct 2016 16:03:47 +0000 (16:03 +0000)]
Merge "Move small fixes and refactoring for obmc pred from AV1" into nextgenv2

8 years agoMerge "adapt_scan experiment" into nextgenv2
Yaowu Xu [Tue, 25 Oct 2016 16:03:40 +0000 (16:03 +0000)]
Merge "adapt_scan experiment" into nextgenv2

8 years agoRefactor extrabits packing
Alex Converse [Tue, 26 Jul 2016 23:27:51 +0000 (16:27 -0700)]
Refactor extrabits packing

- Eliminate the awkward _av1 suffix/infix in local variable names.
- Lift bitdepth selection out of the token loop.

Change-Id: I26d3397464f7808e0481a804033a93ca4f01f5d5

8 years agoLinearize extrabits writing.
Alex Converse [Tue, 26 Jul 2016 22:35:42 +0000 (15:35 -0700)]
Linearize extrabits writing.

The decoder is already linear so changing these tables would just create
a mismatch.

Change-Id: Ib888c0dc273e089c38298f569bb35b6e4c32dd60

8 years agoAutomatically upgrade profile to match input chroma subsampling.
Thomas Daede [Mon, 19 Sep 2016 20:24:51 +0000 (13:24 -0700)]
Automatically upgrade profile to match input chroma subsampling.

This is a follow-up to 1195a396f6c53a5bd35559eed957c2aac855f7e.

Change-Id: I4db554e5d88290d55258062e29a1652707d91037

8 years agoRemove select_tx_size from struct macroblock
Yaowu Xu [Mon, 24 Oct 2016 19:55:54 +0000 (12:55 -0700)]
Remove select_tx_size from struct macroblock

It is no longer used anywhere.

Change-Id: I5d40664373d66821c5382f6155302b8719ce58c0

8 years agoAvoid unnecessary reencode in choose_largest_tx_size
Guillaume Martres [Mon, 22 Aug 2016 02:02:33 +0000 (19:02 -0700)]
Avoid unnecessary reencode in choose_largest_tx_size

This change is similar to the one done for choose_tx_size_from_rd in
daf841b4a10ece1b6831300d79f271d00f9d027b

It gives a 4% speed-up on bus_cif.y4m with the following settings:
--cpu-used=4 -p 1 --end-usage=q --cq-level=40 --tile-columns=0 --tile-rows=0

Change-Id: Ic54fe4a066a2c0b5f6349d80cd13de8bb8ddcabc

8 years agoRemove encode_breakout and related speed features
Brennan Shacklett [Sat, 20 Aug 2016 00:34:11 +0000 (17:34 -0700)]
Remove encode_breakout and related speed features

Seems to be dead code

Change-Id: I17b3edc9e82d6a1da172a686522358a6b1a630e9

8 years agoFix compile error with --enable-accounting
David Barker [Tue, 25 Oct 2016 12:52:07 +0000 (13:52 +0100)]
Fix compile error with --enable-accounting

Change-Id: I4b18dbfb013c9805cb23083a68560ab212a0867a

8 years agoFix to make intra_only frames decodable out of order
Peter de Rivaz [Tue, 25 Oct 2016 09:19:57 +0000 (10:19 +0100)]
Fix to make intra_only frames decodable out of order

last_frame_type is not well defined for intra_only frames
if we are decoding them out of order.
This change removes a dependency on last_frame_type for these frames.

Change-Id: I440cac68792714de222e192a0b3e75f6e1aa5e4b

8 years agoMerge "Extend warp_frame functions to average compound predictions" into nextgenv2
Sarah Parker [Tue, 25 Oct 2016 02:00:48 +0000 (02:00 +0000)]
Merge "Extend warp_frame functions to average compound predictions" into nextgenv2

8 years agoadapt_scan experiment
Angie Chiang [Fri, 21 Oct 2016 23:44:47 +0000 (16:44 -0700)]
adapt_scan experiment

Performance improvement
        BDRate
lowres  0.921%
midres  0.730%
hdres   1.019%

Change-Id: I26208d6c0531937bff44de505b4ea355c7852802

8 years agoMark bogus palette color probabilities as zero
Alex Converse [Thu, 6 Oct 2016 22:04:35 +0000 (15:04 -0700)]
Mark bogus palette color probabilities as zero

It's clearer on inspection that the zero probabilities are unused.

Cherry-picked from aomedia/master: 8134db1

Change-Id: I56cddcb41ba256b7bb921d6a8538405165566dfb

8 years agoMerge changes I2153c57e,I0e291edd into nextgenv2
Urvang Joshi [Tue, 25 Oct 2016 01:06:28 +0000 (01:06 +0000)]
Merge changes I2153c57e,I0e291edd into nextgenv2

* changes:
  Palette: Generate encodings automatically from tree.
  Palette + Ext-Intra: shadowed declaration fix.

8 years agoget_palette_color_context: code cleanup
Urvang Joshi [Tue, 25 Oct 2016 00:11:10 +0000 (17:11 -0700)]
get_palette_color_context: code cleanup

consts, comments and other small readability improvements.

Change-Id: I40c5a652811a796fdb91dc7ca6b108e8871f72d1

8 years agoMerge "Fix bugs in SUB8X8_MC" into nextgenv2
Yue Chen [Mon, 24 Oct 2016 23:16:08 +0000 (23:16 +0000)]
Merge "Fix bugs in SUB8X8_MC" into nextgenv2

8 years agoMerge "Refactor tx_size to pixel number mapping in reconintra.c" into nextgenv2
Jingning Han [Mon, 24 Oct 2016 22:24:04 +0000 (22:24 +0000)]
Merge "Refactor tx_size to pixel number mapping in reconintra.c" into nextgenv2

8 years agoMerge "Simplify variable defs in av1_encode_block_intra" into nextgenv2
Jingning Han [Mon, 24 Oct 2016 22:23:59 +0000 (22:23 +0000)]
Merge "Simplify variable defs in av1_encode_block_intra" into nextgenv2

8 years agoMerge "Refactor av1_predict_intra_block tx_size interface" into nextgenv2
Jingning Han [Mon, 24 Oct 2016 22:23:40 +0000 (22:23 +0000)]
Merge "Refactor av1_predict_intra_block tx_size interface" into nextgenv2

8 years agoMerge "Add block size in pixels lookup table" into nextgenv2
Jingning Han [Mon, 24 Oct 2016 22:23:36 +0000 (22:23 +0000)]
Merge "Add block size in pixels lookup table" into nextgenv2

8 years agoMerge "Fix unsigned type error in aom_scale.c" into nextgenv2
Angie Chiang [Mon, 24 Oct 2016 21:41:18 +0000 (21:41 +0000)]
Merge "Fix unsigned type error in aom_scale.c" into nextgenv2

8 years agoPalette: Generate encodings automatically from tree.
Urvang Joshi [Mon, 24 Oct 2016 21:06:43 +0000 (14:06 -0700)]
Palette: Generate encodings automatically from tree.

Ran some manual sanity checks:
- Verified that the automatically generated encodings match the
  hand-written encodings before the patch.
- Verified that the encoded bitstream before/after this patch is
  identical.

Change-Id: I2153c57e463cff09c1d03d619b432fb1015199c3

8 years agoMove small fixes and refactoring for obmc pred from AV1
Yue Chen [Fri, 21 Oct 2016 23:50:52 +0000 (16:50 -0700)]
Move small fixes and refactoring for obmc pred from AV1

Covering commits 1c263e0 and 79d8a07 from AOM codebase

Change-Id: I6400e5f99bbb2ef6584ef232d465e520230c06e0

8 years agoPalette + Ext-Intra: shadowed declaration fix.
Urvang Joshi [Mon, 24 Oct 2016 21:13:55 +0000 (14:13 -0700)]
Palette + Ext-Intra: shadowed declaration fix.

This shadowed declaration warning was generated when both experiments
are on.

Change-Id: I0e291eddeefabd68c5c3a0e5f8ac87706a82d55a

8 years agoPrevent potential token buffer overflow in format 444
Jingning Han [Fri, 29 Jul 2016 17:48:34 +0000 (10:48 -0700)]
Prevent potential token buffer overflow in format 444

For a 16x16 pixel block, one needs to allocate 16x16 coefficient
tokens, plus up to 16 eob tokens, per plane. This commit increases
the token allocation size to cover the case where all the transform
blocks are of size 4x4 in format 444.

Change-Id: I5755e6a53771053d51163d01ec1d62e670c5009e

8 years agoAutomatically set internal bit depth to at least the input bit depth.
Thomas Daede [Fri, 26 Feb 2016 01:23:17 +0000 (17:23 -0800)]
Automatically set internal bit depth to at least the input bit depth.

Upgrade profile if required.

Change-Id: Ieb2b77d2446290a8fc749739247a01e8f0600c55

8 years agoMerge "Add MAX_VARTX_DEPTH macro" into nextgenv2
Jingning Han [Mon, 24 Oct 2016 21:01:29 +0000 (21:01 +0000)]
Merge "Add MAX_VARTX_DEPTH macro" into nextgenv2

8 years agoMerge "Simplify the recursive transform block decoding" into nextgenv2
Jingning Han [Mon, 24 Oct 2016 21:01:17 +0000 (21:01 +0000)]
Merge "Simplify the recursive transform block decoding" into nextgenv2

8 years agoMerge "Correct data size estimation for odd size video" into nextgenv2
Yaowu Xu [Mon, 24 Oct 2016 20:57:55 +0000 (20:57 +0000)]
Merge "Correct data size estimation for odd size video" into nextgenv2

8 years agoMerge "Fix a bug when combining new-quant + supertx" into nextgenv2
Debargha Mukherjee [Mon, 24 Oct 2016 19:53:41 +0000 (19:53 +0000)]
Merge "Fix a bug when combining new-quant + supertx" into nextgenv2

8 years agoFix bugs in SUB8X8_MC
Yue Chen [Mon, 24 Oct 2016 19:21:48 +0000 (12:21 -0700)]
Fix bugs in SUB8X8_MC

Change-Id: Ia544974f83c6b7f9cdb148eeb13a6d0c6eb4ed24

8 years agoMerge "Increase min size of compressed data" into nextgenv2
Yaowu Xu [Mon, 24 Oct 2016 19:21:44 +0000 (19:21 +0000)]
Merge "Increase min size of compressed data" into nextgenv2

8 years agoMerge "Avoid the use of uninitialized value in ActiveMap encoding route" into nextgenv2
Yaowu Xu [Mon, 24 Oct 2016 19:21:29 +0000 (19:21 +0000)]
Merge "Avoid the use of uninitialized value in ActiveMap encoding route" into nextgenv2

8 years agoFix unsigned type error in aom_scale.c
Angie Chiang [Mon, 24 Oct 2016 16:34:53 +0000 (09:34 -0700)]
Fix unsigned type error in aom_scale.c

Avoid unary minus operator applied to unsigned type

Change-Id: I6986cd2b0ea236e0129ee94c02275593c287a87d

8 years agoMerge "Use the actual inter prediction filter buffer in DRL" into nextgenv2
Yaowu Xu [Mon, 24 Oct 2016 18:34:29 +0000 (18:34 +0000)]
Merge "Use the actual inter prediction filter buffer in DRL" into nextgenv2

8 years agoMerge "fdct4x4_test: fix unsigned overflow" into nextgenv2
Yaowu Xu [Mon, 24 Oct 2016 18:25:55 +0000 (18:25 +0000)]
Merge "fdct4x4_test: fix unsigned overflow" into nextgenv2

8 years agoMerge "Add bit accounting information for deringing" into nextgenv2
Debargha Mukherjee [Mon, 24 Oct 2016 18:14:31 +0000 (18:14 +0000)]
Merge "Add bit accounting information for deringing" into nextgenv2

8 years agoMerge "Use remove some magic numbers in aom_rans_merge_prob8_pdf." into nextgenv2
Yaowu Xu [Mon, 24 Oct 2016 18:10:32 +0000 (18:10 +0000)]
Merge "Use remove some magic numbers in aom_rans_merge_prob8_pdf." into nextgenv2

8 years agoRefactor tx_size to pixel number mapping in reconintra.c
Jingning Han [Mon, 24 Oct 2016 18:04:47 +0000 (11:04 -0700)]
Refactor tx_size to pixel number mapping in reconintra.c

Change-Id: Id66a14a869df8317c5bbb693d14262326fe84206

8 years agoSimplify variable defs in av1_encode_block_intra
Jingning Han [Mon, 24 Oct 2016 17:32:37 +0000 (10:32 -0700)]
Simplify variable defs in av1_encode_block_intra

Use direct table access to fetch the block size and transform size
in pixels.

Change-Id: Ia0093d5aed912be24996a06b0567bb2d873ec068

8 years agoRefactor av1_predict_intra_block tx_size interface
Jingning Han [Mon, 24 Oct 2016 17:27:28 +0000 (10:27 -0700)]
Refactor av1_predict_intra_block tx_size interface

Simplify the input arguments. Make direct use of the block size
in the unit of pixels.

Change-Id: Ifec9d90b4b4fa9605f93b4f93b8242f76f898b5f

8 years agoCorrect data size estimation for odd size video
Yaowu Xu [Tue, 12 Jul 2016 18:57:51 +0000 (11:57 -0700)]
Correct data size estimation for odd size video

Given the largest transform size is 32x32, this commmit changes size
estiiation based on the size rounding up to 32 multiples to avoid
insufficient buffer allocations.

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=36

Change-Id: I6eab09dc6acdc0f5a6bcadb918d62c4852aae21f

8 years agoAdd MAX_VARTX_DEPTH macro
Jingning Han [Mon, 24 Oct 2016 17:38:43 +0000 (10:38 -0700)]
Add MAX_VARTX_DEPTH macro

Change-Id: I85532cf88f91f0f0cb4d9cb4b2dbda8a181297bf

8 years agoIncrease min size of compressed data
Yaowu Xu [Mon, 11 Jul 2016 18:38:53 +0000 (11:38 -0700)]
Increase min size of compressed data

This commit increases the minimum size for allocated buffer for
compressed data. The old size underestimated the size needed for
small images with width or height less than 64 pixels.

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=31

Change-Id: Ia12507edc2be1e737ec49c32f64fd2ebf1eab41f

8 years agoFix a bug when combining new-quant + supertx
David Barker [Mon, 24 Oct 2016 09:55:35 +0000 (10:55 +0100)]
Fix a bug when combining new-quant + supertx

Previously, we assumed that av1_init_plane_quantizers is always called with
segment_id == xd->mi[0]->mbmi.segment_id (and use the latter to derive the value
of 'qindex' to use in the quantizer). But this is no longer true when supertx
is enabled. This patch instead remembers the value of 'qindex' derived from
the latest call to av1_init_plane_quantizers and uses that directly.

Change-Id: Ifa1c5bf74cad29942ff79b88ca92c231bc07f336