GitLab

This gives a -0.05% change in performance over the regular
quantizer. The new_quant bins will remain uniform as we
experiment with different parameters.

Change-Id: Ic92c74fb1183cf1c2460bbcf1e223452f46d8210

Change-Id: I71faf1d2046b00ce87f85061336d840e5ffe3743

When enabled, now the code is in fact effective to collect frame-
based psnr and ssim values.

Change-Id: Idc2349392ccd6c6dfae8fd9ac4638ceaa530fb5c

Current encoding of single ref assigns bit 0 to LAST_FRAME and bit 1
to LAST2_frame, whereas the encoding of compound ref assigns bit 1 to
LAST_FRAME and bit 0 to LAST_FRAME. This patch unifies the design and
makes the binary symbol assignment for LAST/LAST2 consistent
regardless of the single/compound ref scenarios.

This patch incurs a small syntax change but should not have noticeable
coding performance impact. Once BUG 973 is fully addressed, default
cdfs for reference frame coding will be updated.

BUG=aomedia:973

Change-Id: I1f3027965c0bb2d343bc8ad873f7c0015123f151

Change-Id: I40ae0bcc7c01daf4bca1b3c2358499a5b1ae7808

Change-Id: I19db3c7a700c64fcd493ee50a30184c9529f6273

Change-Id: I43da630aa38c1b77fd76e5ef1a8e07a3b2b4ba6b

This struct does not appear to be used.

Change-Id: I964362e783a7f9ab6cf6ccc98bc7a279876075b4

Change-Id: I334281aee38be03a383ae6899970b4200ec43c01

zero_rate should not be added.

Change-Id: I372b82f69915809869cbc352c2db44c963c31cfa

Change-Id: I3acd24acbd121c93e89817f865a54327d5e8e92e

Rename aom_highbd_lpf_horizontal_edge_8() to aom_highbd_lpf_horizontal_16().
Rename aom_highbd_lpf_horizontal_edge_16() to aom_highbd_lpf_horizontal_16_dual().

based on the same change from libvpx:
7f1f35183 Unify loopfilter function names

Change-Id: I40cd587e74e0fe02bae23e6c10280c8e269df1d6

Change-Id: I4eb1800c3fca7ea76d0e44dd44f2a3edbf209518

BUG=aomedia:78

Change-Id: Id85cf22302317d33b593345fd31090a6620f4101

Change-Id: Ibca63a41bdffaab2ee603497e8a5fc66d5b7d11f

Change-Id: Ib002bb918366ba6848c983f5311aeb74ab37c1a0

Change-Id: Ic31a53a37ff9200fad178e5054cffc4b87c6cc42

This experiment has been abandonned for AV1.

Change-Id: I83fb51a17d67df6713308665d2626c232376d25a

This experiment has been abandonned for AV1.

Change-Id: Ib9a8eea6b0ad5580dc81836be8fc5e898bb9bb45

Tested on lowres for 30 frames, no compression loss; a few percent
encoding speed improvement.

Change-Id: I11b230c0082a669a494c6a9b26a291fd07ac37a4

Change-Id: Ic4719c70567b9a9faef15b3d09ebe920750aa19e

only output short usage to stderr on error, with --help use stdout

BUG=aomedia:44

Change-Id: Ia7d0316a354d2dfadb36698dd4958355f6168a60

This is for aom_memset16() used later in the code.

Change-Id: I4426354bae4f0bf2e3012a138a8e779098b35173

Change-Id: Ibfa7d22f9e304c1f8871ef0a11a63005b17dd780

Change-Id: I94200d74719c8fb646d628d14a7843ec80c1b7c4

BUG=aomedia:1074

Change-Id: Icb973b56ac60780ef4fa61bcd4114de37f6fd120

This avoids a valgrind warning of "use of uninitialized memory".

Change-Id: I4a0e6540b704a9cf532cf894101779277ec66ba4

If "--enable-dist-8x8=1" with a lossless mode, exit with system error.

BUG=aomedia:1027

Change-Id: I08b4cdfd7b94d6098702091d2aab0f94d651ceb9

Change-Id: If840d0b8d22fef8cbf5a7bfc1d9b5ee324a2f020

The dist-8x8 can only work with pixel domain distortion.
Thus, dist-8x8 will be automatically disabled if cpu-used >= 2, which
set sf->use_transform_domain_distortion as 1.

BUG=aomedia:1027

Change-Id: I3180a2d5b126318615afcca49c5c2d1b75f9fb58

Larger than 8 bit input seems to introduce different rounding methods in
different distortion calculation functions, which causes assertion failed
for dist-8x8's sanity check.

BUG=aomedia:1027

Change-Id: I1f607f32d4017d8f9abbbe0b09d9dbb01a63fe9c

Adds an option to split chroma tx size by one level if the luma
transform size is split at least one level. Otherwise use the
largest transform size for chroma.
This option is enabled with DISABLE_VARTX_FOR_CHROMA set as 2,
but is currently turned off until we finish testing.

Change-Id: Ie048c4e33193e0cc7e1bbee37e4ccc085834433c

Change-Id: Ide6069ba831fd276aae7b3cb1b8d36a984401248

Change-Id: Ie19eef5f6f588cd445b66afae7482999952e118f

Can be up to 40% faster with bit exact results

Change-Id: Ia67ba154222fdfb6173bf8942275649e511abe43

In cb4x4, dual loopfilter filters 2 * 4 = 8 pixels.
This patch does not influence encoder/decoder since
they are not applied in bit mask implementation.

Change-Id: Ifdeb8990127de39143971156db69a69ee3bd3136

This mistake was fixed in a8db0e8d
and brought back in for
the daala-tx case in 4a05a58c.

Change-Id: I9cf36b63f2dacd9d01887bfaacd8b318245f882a

Reuse the existing ones to support the 64x64, 64x32 and 32x64 configs
instead.

Change-Id: Ibadc65eb03518d19146f32657ccfd6949c0071cc

Instead of using -1 to indicate jnt_comp is used or not, use a separate
variable.

Change-Id: I9b206d63394c431741d89611a614087343a417fc

Added convolve 1d functions for 1d-only sub-pixel interpolation, including
av1_convolve_x() function for "subpel_x_q4 == 0" case and av1_convolve_y()
function for "subpel_y_q4 == 0" case. Separated into different cases for
speeding up the codec(especially the decoder). The optimized version will
be added later.

Change-Id: I08fd317d02c786af68270a971bb3b17c94bd52ca