Julia Robson [Mon, 1 Feb 2016 15:44:50 +0000 (15:44 +0000)]
Supertx fix for 422 colour subsampling
Fixes assertion for football_422_4sif.y4m when supertx, var_tx and
ext_tx are all enabled. Problem was after subsampling, the u and v
blocks being encoded were no longer square.
Jingning Han [Fri, 15 Jan 2016 17:05:51 +0000 (09:05 -0800)]
Enable adaptive motion vector referencing mode
This commit allows an adaptive motion vector referencing mode
approach. It checks the available reference motion vector candidate
list and decides the amount of motion vector referencing modes. The
current implementation assumes simple binary coding for the syntax.
The compression performance is improved by
derf 0.11%
hevcmr 0.38%
stdhd 0.09%
hevchr 0.23%
The coding gains due to the new reference motion vector system are
derf 1.0%
hevcmr 1.7%
stdhd 1.4%
hevchr 1.3%
James Zern [Sat, 30 Jan 2016 00:27:45 +0000 (00:27 +0000)]
Merge changes I51d42a50,Id8aaf820,I6af91959
* changes:
configure: extend armv7 hf target autodetect
configure: remove default CROSS for arm targets
configure: avoid default when CROSS is set to null
Yaowu Xu [Fri, 29 Jan 2016 00:29:29 +0000 (16:29 -0800)]
SSSE3 idct8x8 functions for highbitdpeth build
This commit changes SSSE3 optimized idct8x8 functions to work with
highbitdepth build.
With this commit and the previous one that enabled SSSE3 idct32x32
functions, tests showed virtually no difference on decoding speed for
file fdJc1_IBKJA.248.webm for the build with -enable-vp9-highbitdpeth
option and the build without the option.
Yue Chen [Wed, 27 Jan 2016 22:37:17 +0000 (14:37 -0800)]
Fixing a issue of calculating tx cost for SUPERTX+VAR_TX
Update blk_skip in update_state_supertx() and rd_supertx_sb().
Performance (SUPERTX+VAR_TX): TBD
(Eventually will merge update_state() and update_state_supertx())
James Zern [Thu, 28 Jan 2016 20:53:41 +0000 (12:53 -0800)]
vp9/10: fix encoder crash on flush
the lookahead buffer allocation is deferred to receipt of the first
frame to allow profile changes. if the encoder was flushed before
supplying any frames the encoder would crash trying to dereference the
NULL buffer. vp8 is unaffected.
fixes mozilla bug:
https://bugzilla.mozilla.org/show_bug.cgi?id=1237848
Marco [Wed, 27 Jan 2016 20:01:58 +0000 (12:01 -0800)]
vp9-aqmode=3: Limit on period for golden update.
For aqmode=3, golden period update is set based on period of cyclic refresh.
Put a limit on max golden period update, for now set to 40.
And fix comment.
Geza Lore [Tue, 26 Jan 2016 14:40:32 +0000 (14:40 +0000)]
Always recode SUPERTX blocks.
There is still an assertion failure when tokenizing transform
coefficients for a supertx block, due to the eob not being set
consistently with the coefficients, so we always recode supertx blocks
for now. Also added further PICK_MODE_CONTEXT instances to avoid
potential clash between horizontal/vertical/split partition SUPERTX
trials.
Geza Lore [Tue, 19 Jan 2016 18:15:40 +0000 (18:15 +0000)]
Update VAR_TX related context when using SUPERTX.
The encoder did not update left_txfm_context and above_txfm_context in
MACROBLOCKD (used for choosing the probability context for the vartx
split bits) when the supertx bit was set for a block. The deoder on the
other hand did update these for supertx blocks. The encoder used these
to compute the context counts, which the packer then uses to adapt it's
probabilities. This results in the packer and the decoder using
different probabilities.
This patch harmonizes the encoder and the decoder by making the encoder
update the mentioned context for supertx coded blocks.
Geza Lore [Mon, 25 Jan 2016 15:46:59 +0000 (15:46 +0000)]
Initalize mbmi->tx_size during rdopt.
This is necessary when using SUPERTX, as the bitstream packer relies on
tx_size being set correctly to decide whether to output the block using
supertx or not.
Geza Lore [Thu, 21 Jan 2016 10:46:33 +0000 (10:46 +0000)]
Set inter_tx_size for supertx coded blocks.
The loop filter relies on inter_tx_size in MB_MODE_INFO being set
properly when VAR_TX is enabled. Supertx coded blocks did not set this
previously at all, and the differing garbage values eventually resulted
in in a YUV mismatch between encoder and decoder after loop filtering.
This patch fixes this by setting inter_tx_size to the proper supertx
size in both the encoder and the decoder. This should also mean that
loop filtering is done at the proper transform boundaries, even when
supertx or vartx is being used.
Alex Converse [Mon, 25 Jan 2016 20:03:50 +0000 (12:03 -0800)]
vp9: Be more responsive to massive overshoot.
Limit oscilation detection in the case where overshoot is very very
large.
This keeps the 9-bit cost patch from breaking the DownUp reisze test.
The patch pushed us to an 11% undershoot right before a scene cut
causing a 1200% overshoot. (Whereas before we were undershooting by
only 6% before overshooting by 1200%).
Yue Chen [Wed, 20 Jan 2016 00:45:45 +0000 (16:45 -0800)]
Adding new compound modes to EXT_INTER experiment
Combinations of different mv modes for two reference frames
are allowed in compound inter modes. 9 options are enabled,
including NEAREST_NEARESTMV, NEAREST_NEARMV, NEAR_NEARESTMV,
NEAREST_NEWMV, NEW_NEARESTMV, NEAR_NEWMV, NEW_NEARMV, ZERO_ZEROMV,
and NEW_NEWMV.
This experiment is mostly deported from the nextgen branch.
It is made compatible with other experiments
Coding gain of EXT_INTER(derflr/hevcmr/hevchd): 0.533%/0.728%/0.639%
projects / libvpx / log