Граф коммитов

21043 Коммитов

Автор SHA1 Сообщение Дата
Nathan E. Egge ad66b816ee Merge NEW_TOKENSET experiment into EC_MULTISYMBOL.
Change-Id: I0ed6d8f2b6a83641f65096cf41fcde94e70d94d4
2017-05-16 02:11:54 +00:00
Joe Young 12c0bc0300 [intra-edge] Filter intra edge samples
For ext-intra directional modes, filter intra edge samples
dependent on block size and prediction angle.

Congiguration:
./configure --enable-experimental --enable-ext-intra
  --enable-internal-stats --disable-cdef --disable-delta-q

Coding gain (1 frame Intra):
cif:    -0.03%
midres: -0.01
720p:   -0.65
1080p:  -0.83
4k:     -1.19

Change-Id: I372bfc7614f6a07fe1d8e7b61bd0b6660f9fb4a1
2017-05-16 02:05:58 +00:00
Yunqing Wang d097ec148a Fix debugmodes file
Fixed debugmodes.c to incorporate the mi size change in CB4X4.

Change-Id: Id4d44866636a1a0715e03a7a4c9884c51ca7f59d
2017-05-16 00:12:47 +00:00
Nathan E. Egge f059daa32e Add missing include for MAX_SEGMENTS define.
Change-Id: I69031c3f0fdd363058c159a8ee008b18469e81c1
2017-05-15 22:58:37 +00:00
Luc Trudeau b09b55dbfb [CFL] Add support for intra in inter
Change-Id: I4d8485d2bbe5348338eb909d969a0ca4b598a6b8
2017-05-15 21:36:51 +00:00
Alex Converse e46382a480 intrabc: Fix build when palette is disabled
Change-Id: I2ad15784cbc238fa4b0f993bfa64c06591ece5e9
2017-05-15 20:07:53 +00:00
Debargha Mukherjee 28d15c715a Experimental high precision convolve for Wiener
Improves coding efficiency.

Change-Id: I7bb12190cdc4581097809a020355cdc8867fc1ad
2017-05-15 19:48:41 +00:00
Ralph Giles 12a27c0150 Remove armv6 references from the build system.
We no longer support armv6 and the corresponding
optimized source files have been removed. This
commit removes support for the target and its
media instruction set extensions from the build
system as well.

Once armv6 jobs are removed from the integration
config we can start rejecting --enable/disable-media.

Also reject --enable-edsp.

Change-Id: I547dcc95adb3479d0232396952e72efe2a1343b9
2017-05-15 19:06:31 +00:00
Yue Chen 8a32e1a7d3 Add rect_tx_ext experimental flag
An experiment on 1:4/4:1 transform, which will be a sub-experiment
of ext_tx+rect_tx experiment

Change-Id: I37c41227f83c51ab11194007226a2d6bcd6ff74f
2017-05-15 19:05:18 +00:00
Yi Luo 2ab63cb74b Integrate partial IDCT to adapt_scan
- Add eob delivery in av1_first_pass().
- Fix eob threshold pointer copy.
- Change local functions in IDCT to static.

Change-Id: I1c95e8f899088773f063fe6be7f26a5f96864644
2017-05-15 18:49:21 +00:00
Yi Luo 3e629b9692 Fix build on --disable-avx2
BUG=aomedia:536

Change-Id: I92c5e43ba225968504ff1c878ca80b86449aeefe
2017-05-15 18:42:41 +00:00
Sebastien Alaiwan 817ddb73bc Fix HBD compilation of pvq
Change-Id: I3b30030b1b5f94b43b6dac7b2eac1dc1ea308e84
2017-05-15 18:21:06 +00:00
Debargha Mukherjee 34f4fd3a57 Some minor changes in wedge generation code
Makes the wedge generation code consistent with the
design doc.

Change-Id: Ifabc63c950de005309c44802c725f45cb78e21b1
2017-05-15 18:20:41 +00:00
Debargha Mukherjee e39e2ee1e6 Add finer search for wiener filter parameters
Improves coding efficiency a little

Change-Id: I1d5b3e3925e7fb640ea4da76b529672c4183a71c
2017-05-15 18:04:02 +00:00
Yue Chen 7bfa86773c Fix filter_intra overflow
(1) clipping output of intra filter right away to stop
unreasonable pixel values from propogating.
(2) fix wrong filter coefficients

Bug=aomedia:418

Change-Id: I2620034ad03b499afb64da6cadb17fc1aa221c44
2017-05-15 17:39:52 +00:00
Fergus Simpson 3502d08b71 frame-superres: Pre-encode downscaling
Adds handling needed for the pre-encode downscaling for superres
encoding. Makes use of the existing resizing functionality, just
adjusting the target resolution after the resize logic and before the
resize has actually occurred.

Since the upscaling portion of superres is not yet implemented, superres
scaling is not yet allowed to trigger. The numerator is set to be equal
to the denominator and the superres_pending flag is not set.

Change-Id: Ia5dac752d653c3bda319f0c8196e95bac2c8c554
2017-05-15 17:28:40 +00:00
Alex Converse 44c2bad5c0 intrabc: Add a DV prediction scheme derived from ref_mv
wikipedia_420.y4m
PSNR: -1.5035 PSNR-HVS: -1.2607 SSIM: -1.1581

Change-Id: I6e6c4bd9671427b99299df39ff0f42af7daf8ce7
2017-05-15 16:40:37 +00:00
Zoe Liu 40bc8b8a34 Clean up on getting the compound mode context
Change-Id: I231523ad60cc90559dfbdd51fd3e8c07c10d63a4
2017-05-15 16:33:43 +00:00
Thomas Daede 0743857c1a Only seed motion search with mvrefs if they are available.
Fixes an encoder-side crash.

BUG=aomedia:519

Change-Id: I420404e21fe23bc753bfb08896f2b8b2e014af4e
2017-05-15 16:19:07 +00:00
Ryan 25d58829c4 Fix invalid assertion
BUG=aomedia:535

Change-Id: I514b2b0503e39023ae19b38a9b533bcd9abeee77
2017-05-15 16:11:15 +00:00
James Zern 5bedd5dc8f idct16x16_sse4_1: correct fn sig for visual studio
fixes:
formal parameter with __declspec(align('16')) won't be aligned

Change-Id: Ie073a53dd99913ba8c9f05964590dd7dcdf7d9fe
2017-05-15 15:58:09 +00:00
Ralph Giles be111b3838 Remove armv6 media-extension assembly.
Libvpx dropped armv6 support sometime after the aom fork.

We don't intend to support this platform, which is likely
too slow in any case. Remove the assembly and intrinsics
optimized routines, their tests, cpu feature detection,
and rtcd specialization for this instruction set extension.

Change-Id: If44ec28e5ddafc6af179c5d1982ac7e81fe54d5e
2017-05-15 15:55:47 +00:00
Yaowu Xu 70d9acc1c5 Avoid left shift of negative numbers
Slience warnings by converting the shifts to mulitplies.

Change-Id: Icde8f2df650f740b8e90691ba706a0853be84984
2017-05-14 20:34:04 -07:00
Yaowu Xu 40131320f4 Silence a warning of conversion to int16_t
Change-Id: I89c559ba4959b7674e4823c7e3c9b7c509b62172
2017-05-14 20:23:11 -07:00
Tom Finegan 782cdc125f Add aom.pc (pkg-config file) generation to CMake.
BUG=aomedia:76,aomedia:471

Change-Id: I7919e00d88ea448465dfb4c51b61a65a37b06060
2017-05-15 03:11:15 +00:00
Tom Finegan e9d70c9b96 Add a test target with sharding support to the CMake build.
Two major things going on here:
- One target is created for each file in test-data.sha1 (testdata_N),
  and the testdata target now depends on all testdata_N targets. The
  testdata build rule can now run with as many jobs as there are
  input files to speed up test data download.
- GTest sharding support has been added to the tests via a runtests
  custom build target. First, the number of processors is detected,
  and then a custom target is created for each processor (test_N).
  Once each test_N target is created, the runtests custom target
  is created, and then made to depend on each test_N target. When
  CMake is unable to detect the number of processors 10 test targets
  are created. Each custom target then sets the GTEST_SHARD_INDEX and
  GTEST_TOTAL_SHARDS environment variables, allowing GTest to handle
  sharding internally.

BUG=aomedia:76,aomedia:469

Change-Id: Ib6b7974932396fbf44b735d37155fa57561027ab
2017-05-15 03:11:15 +00:00
Tom Finegan a9c4e511ef Add some missing sources to fix the CMake build.
CMake build out of sync/broken since
ee6743235d landed.

Change-Id: Ica3186db2e8b80750069f224f5b5858491e5be0b
2017-05-15 03:11:15 +00:00
Debargha Mukherjee 0a981101ec Fix allocation for least squares
Change-Id: I663a63a1b97b529142ea92bcf7f9270a1df838f5
2017-05-15 01:40:14 +00:00
Urvang Joshi 0ba850ea6d Palette_delta_encoding: compile fix.
s/uint16/uint16_t/

Issue introduced in https://aomedia-review.googlesource.com/c/11148/

Change-Id: I77cb5865828adef98976863fafa2ecbb907290a5
2017-05-14 19:40:39 +00:00
James Zern 52a2c83870 idct_intrin_sse2,fliplr*: use *in instead of in[]
fixes visual studio warning:
formal parameter with requested alignment of 16 won't be aligned

Change-Id: I49bb4eeeda4d0f0829b2777a0ae1d66d20821883
2017-05-13 00:15:48 +00:00
James Zern d0545b38a9 iidtx4_sse2: silence tran_high conv warning
cast Sqrt to short, _mm_set1_epi16 expects this as input

Change-Id: I667cd0bfcc86acfa5ccd7069f02e4dd1ec7b6bae
2017-05-13 00:15:48 +00:00
Urvang Joshi 378644ca82 warped motion: fix some UBsan errors.
Fixes the following UBSan errors due to left shifting
negative values.

Change-Id: I578ca3f56417a94ffe83f634171903897117aa6b
2017-05-12 23:56:20 +00:00
hui su 33567b21aa Palette: use color cache to compress base colors
Get a list of palette base colors that are used in the above and
left blocks, referred to as "color cache". For each cache color,
signal if it is present in current block's palette, so that we
don't need to transmit their raw values.

When palette-delta-encoding is enabled, compression is improved
by 2% on keyframe and 1% overall for the screen_content testset.

Change-Id: I4cb027f1904aa9d0ab1c8f00ea9ee34bf5f16234
2017-05-12 19:17:32 +00:00
Yi Luo d1fb415ff2 Fix uninitialzed __m256i vector warning
BUG=aomedia:529

Change-Id: I9fccd0d29d100c92152d33a74dc3df8b7d256bcb
2017-05-12 19:13:40 +00:00
Fergus Simpson 090bb60f68 frame_superres: Move av1_scale_if_requrired to common
This commit moves the av1_scale_if_required functions from
av1/encoder/encoder to av1/common/resize. They take the
scale_and_extend_frame_nonnormative function with them.

This will be useful for resizing received frames in the decoder for the
FRAME_SUPERRES experiment.

Change-Id: I83afc3d97ff8ebce5ecb2e1205ba6c271156cb31
2017-05-12 19:05:06 +00:00
Yaowu Xu eec78cc986 Prevent call to non-existing error_handler
BUG=aomedia:121

Change-Id: Iec2e98c9f80ef3a01476234e637c135ff0513efd
2017-05-12 18:08:01 +00:00
emilkeyder@google.com 6e3557c9b0 Split av1_warp_erroradv into av1_{frame,warp}_error.
This avoids repeatedly computing the frame error in
refine_integerized_param.

Change-Id: If4af77b68a7e61a700bafa3f60940b9017e12b40
2017-05-12 17:26:13 +00:00
Sebastien Alaiwan 1dcb707517 Fix HBD compilation of intrabc
Change-Id: Ie5718de1eb70fd753c2bfd7df9581aa92e13984c
2017-05-12 17:06:33 +00:00
Luc Trudeau 64ed2ce7f8 [CFL] Fix undeclared 'cm' when EC_ADAPT disabled
Change-Id: I9a9d7960b46dd80012fc8efdefb640939e1f1e48
2017-05-12 16:34:31 +00:00
Sebastien Alaiwan 14af5b9591 Simplify conditions on bitdepth
This avoids use of preprocessor when possible, and removes a
duplicated legacy check on codec name.

Change-Id: I8f0d25cf69b51d3379f25c1a69f88c8d404c4e38
2017-05-12 16:22:45 +00:00
Sebastien Alaiwan 18605a3d31 Fix output YUV bitdepth
When the input bitstream is 8 bit, the output YUV should default to
8 bit, even if the internal storage format is 16-bit.

Change-Id: I77eea8a78a60411cfd707c7a4605365abc508200
2017-05-12 16:00:51 +00:00
Luc Trudeau e8dfbcf348 [CFL] Remove explicit dependency on EC_ADAPT
Change-Id: I3e3d458aec7e26abc366992ac20f9fa838fe35bb
2017-05-12 15:47:13 +00:00
Luc Trudeau 041201908b [CFL] move cfl_idx_to_alpha to header
Move cfl_idx_to_alpha in the header to facilitate inlining.
Remove the forward MB_MODE_INFO forward declaration

Change-Id: Id33fb0228d88b6285252843e2345a0d3ae875cd2
2017-05-12 13:58:48 +00:00
Debargha Mukherjee bcfb0e1528 Build fix with interintra alone
Change-Id: I9d771edccd90c72ada72942e11cf3fbabb8146a9
2017-05-11 20:09:16 -07:00
Luc Trudeau 9bab28b61f [CFL] Reduce arrays used in inspection
Merge cfl_alpha_sign_u and cfl_alpha_sign_v into cfl_alpha_sign,
where:
    cfl_alpha_sign = (cfl_alpha_sign_v << 1) + cfl_alpha_sign_u;

Change-Id: I8c491e37d23f6ef85a33a8b2f8770b7b606d8e0a
2017-05-12 00:31:09 +00:00
Luc Trudeau f436b12e24 [CFL] Check for uninitialized signs
To avoid writting out uninitialized signs for CfL alpha, checks are
added in the write_cfl_alphas function.

Change-Id: I832875e496a6318132c82bc1e5406dc95ae9c055
2017-05-12 00:27:11 +00:00
Yaowu Xu 367da993a3 Change the initializaton of test parameters
BUG=aomedia:533

Change-Id: Iffd43f859f958ce2ccae80e3616c368ffebfc8db
2017-05-12 00:12:32 +00:00
Yue Chen 4d26acb4e4 Add CONFIG_INTERINTRA flags and reduce # of interintra modes to 4
Use CONFIG_INTERINTRA to properly separate interintra from the basic
ext_inter experiment.
When macro REDUCE_INTERINTRA_MODES is 1, only 4 ii modes are enabled
so as to reduce the complexity overhead.
(Right now the flag is off)

Change-Id: Iec6f36a1000f181adbb822ad66c7d5b2625e9cc7
2017-05-11 22:58:13 +00:00
Urvang Joshi 4c6d4369aa Handle tx64x64 correctly in get_entropy_contexts.
This unhandled case was causing an assertion failure.

BUG=aomedia:532

Change-Id: I46665c380ad171e2b9ddf4d41b6e9fff838f8b3e
2017-05-11 21:56:34 +00:00
Luc Trudeau a9bd85f709 [CFL] Rename cfl_alpha_ind to cfl_alpha_idx
Change-Id: Id14d83bdce8321bef81972a4f5530a50589d9578
2017-05-11 20:12:38 +00:00