- Jul 19, 2018
-
-
-
Monty Montgomery authored
* cdef: WIP * cdef: now calling cdef_find_dir_c (but only on the first block) * cdef: call cdef_find_dir on every block (only safe for size % 64 == 0) * cdef: apply CDEF to luma with pri_strength 7 and sec_strength 0. * cdef: only apply to full 64X64 filter blocks for now * cdef: signal cdef strength * cdef: comment out printing * cdef: first attempt at applying CDEF to chroma planes * Corrections to first CDEF attempts This is still not quite first cuts, but it is basically functional. It still hard-codes specific CDEF strengths. There is still an occasional block mismatch. * Implement CDEF paramters, remove hardwired settings Add CDEF settings to the FrameInvariants, rather than using hardwires strength values duplicated in multiple locations. * Remove last CDEF dependency on libaom * Remove unrelated change from CDEF work remove -r rec_file.y4m from rav1e command line
-
Michael Bebenita authored
-
-
-
- Jul 18, 2018
-
-
-
-
-
It turns out that there's already some code to emit the skip_mode_present flag, it's just dead code because skip_mode_allowed is currently always false, added an assertion to revisit this code when it becomes live because it's incomplete.
-
Michael Bebenita authored
-
-
-
-
-
- Jul 17, 2018
-
-
-
-
* Drop trailing spaces * Add a QuantizationContext to cache tx_scale, quantization and offset Speedups compared to the baseline TX_4X4, 20 1.32% TX_4X4, 55 1.55% TX_8X8, 20 -0.08% TX_8X8, 55 0.02% * Convert integer divide into equivalent multiply and add operations Speedup compared to the baseline TX_4X4, 20 0.52% TX_4X4, 55 0.37% TX_8X8, 20 2.90% TX_8X8, 55 3.20% * Pre-generate the multiply-add values for the integer division Speedup from the baseline: TX_4X4, 20 0.66% TX_4X4, 55 0.85% TX_8X8, 20 3.51% TX_8X8, 55 3.66%
-
Clean up CDFs.
-
This is "Plan B".
-
Which includes both tile group header and tile data.
-
TODO: Need to fix writing the size of OBU_FRMAE payload in obu header, which is the sum of obu_tile_gropu and a tile.
-
So that it can be used for both show_existing case and tile (group) case. TODO: Add write_tile_group_header() for each tile group.
-
-
-
-
Both OBU_FRAME or OBU_FRAME_HEADER obu types call this function.
-
Write the length of OBU payload between a OBU header and a OBU payload, which introduces two byte buffers and BitWriters at the same time. TODO: - Do similar thing for Frame OBU - Can better factor out the code which does post-writing of the length of OBU payload data. - Multi buffer approach, that is rust dependent, can be improved.
-
The uleb size is the size of length field in bytes. https://aomediacodec.github.io/av1-spec/#leb128 TODO: - Include uncounted sequence syntax data to seqence payload size. - Implement uleb128()
-
Also, add trailing bits for sequence OBU.
-
For each frame, write a OBU_TEMPORAL_DELIMITER, even before OBU_SEQUENCE_HEADER or OBU_FRAME_HEADER.
-
Enable OBU Headers, also use the most recent code of libaom, so that rav1e generated bitstream can be decodable by current version of aomdec. - Add OBU_Type enum - write_obu_header() funcion - Copy reference av1_pack_bitstream() C code fro libaom, where sequence and frame level OBUs are written. [OBU] Add Sequence Header and revise its functions [OBU] More on sequence header functions and types - Revise the type of Sequence struct members. - Add write_sequence_header2(), which is modified version of write_sequence_header() for OBU. - Add new color_config(), i.e. updated color info. TODO: - The structure of OBU is three parts: 1) OBU header 2) Size of OBU (written in leb128() format of AV1) 3) OBU (with specific obu_type signlaed in OBU header) So, need to figure out how to post-write the size of obu (i.e, payload size) after writing obu. In libaom, this is implemented by calling obu_memmove() on byte buffer, which moves the obu data by payload size in bytes to create a gap between obu header and obu then write a obu size.
-
- Jul 16, 2018
-
-
-
-
-
Fixes ./build.sh since README.md now includes more references to cmake.
-
Before: real 1m3.772s user 1m3.454s sys 0m0.178s After: real 0m37.467s user 0m37.163s sys 0m0.142s
-
Fix doc examples. Remove cfg_attr.
-
-
Nathan E. Egge authored
-
-