hashirama/uvg266
1
0
Fork 0
mirror of https://github.com/ultravideo/uvg266.git synced 2024-11-24 18:34:06 +00:00
Code Issues Actions Projects Releases Wiki Activity
1929 commits 3 branches 6 tags 9.2 MiB
Commit graph

1712 commits

Author SHA1 Message Date
Arttu Ylä-Outinen bdd8b1c0aa Implement 2NxN partitions in inter search. 
- Try using 2NxN partitions after the usual 2Nx2N.
- Adds function kvz_search_cu_smp to search_inter module.
 2015-12-15 11:21:42 +02:00
Arttu Ylä-Outinen 410064e880 Split lcu_set_inter into two functions. 
Moves code for setting the inter modes for a single PU to a new function
lcu_set_inter_pu.
 2015-12-15 11:21:42 +02:00
Arttu Ylä-Outinen 3236428e4d Make hexbs search work with non-square blocks. 
Replaces parameter depth with parameters width and height.
 2015-12-15 11:21:42 +02:00
Arttu Ylä-Outinen 31ba8d61c3 Implement fractional chroma recon for SMP blocks. 
Adds parameter block_height to function kvz_inter_recon_frac_chroma.
 2015-12-15 11:21:42 +02:00
Arttu Ylä-Outinen 0b6cef7be5 Remove unused function kvz_inter_set_block. 2015-12-15 11:21:42 +02:00
Arttu Ylä-Outinen e63486b23f Make lcu_set_inter work with SMP blocks. 2015-12-15 11:21:41 +02:00
Arttu Ylä-Outinen 7b99eb2970 Call recon functions correctly for SMP blocks. 
Makes calls to kvz_inter_recon_lcu and kvz_inter_recon_lcu_bipred in
function search_cu work correctly when using SMP blocks.
 2015-12-15 11:21:41 +02:00
Arttu Ylä-Outinen dc4525c0e3 Implement inter recon for non-square blocks. 
Adds parameter height to functions kvz_inter_recon_lcu and
kvz_inter_recon_lcu_bipred and makes them work on non-square sizes.
Fractional reconstruction functions do not handle non-square blocks yet.
 2015-12-15 11:21:41 +02:00
Arttu Ylä-Outinen f874c8614e Add part_mode binarization table comment. 2015-12-15 11:21:41 +02:00
Arttu Ylä-Outinen c77074a7ff Implement encoding SMP blocks. 2015-12-15 11:21:41 +02:00
Arttu Ylä-Outinen 98707a1288 Move encoding intra CU to a separate function. 
Moves code for encoding a single intra coding unit from function
kvz_encode_coding_tree to a new function encode_intra_coding_unit.
 2015-12-15 11:21:40 +02:00
Arttu Ylä-Outinen c336674da3 Move encoding part mode to a separate function. 
Moves code for encoding the part mode from function
kvz_encode_coding_tree to a new function encode_part_mode.
 2015-12-15 11:21:40 +02:00
Arttu Ylä-Outinen ac952cbb44 Move encoding inter PUs to a separate function. 
Moves code for encoding a single inter prediction unit from function
kvz_encode_coding_tree to function encode_inter_prediction_unit.
 2015-12-15 11:21:40 +02:00
Arttu Ylä-Outinen 5ee9f164e8 Add macros for getting PU location and size. 
- Moves SIZE_* definitions to cu.h.
- Adds constant arrays kvz_part_mode_num_parts, kvz_part_mode_offsets
  and kvz_part_mode_sizes for storing the number of PUs, PU offsets and
  PU sizes.
- Adds macros PU_GET_X, PU_GET_Y, PU_GET_W and PU_GET_H for getting the
  location and size of a PU.
 2015-12-15 11:21:40 +02:00
Arttu Ylä-Outinen a3df13fb99 Make kvz_inter_get_merge_cand work with SMP blocks. 
- Replaces parameter depth with parameters width and height.
- Adds parameters use_a1 and use_b1 for disabling the use of merge
  candidates A1 and B1.
 2015-12-15 11:21:40 +02:00
Arttu Ylä-Outinen 1cd149fb97 Check merge/mv candidate types earlier. 
Moves checks for motion vector prediction and merge candidate block
types (inter/intra) from functions kvz_inter_get_mv_cand and
kvz_inter_get_merge_cand to kvz_inter_get_spatial_merge_candidates.
 2015-12-15 11:21:39 +02:00
Arttu Ylä-Outinen 969c91d7c4 Add a test for kvz_inter_get_spatial_merge_candidates. 2015-12-15 11:21:39 +02:00
Arttu Ylä-Outinen 02375bf7e5 Make kvz_inter_get_mv_cand work with SMP blocks. 
Replaces the depth parameter of kvz_inter_get_mv_cand with parameters
width and height.
 2015-12-15 11:21:39 +02:00
Ari Koivula 3a80c7de74 Further optimize coefficient coding 
Remove the need to count the coefficients by populating the significant
coefficient group map first and finding the last coefficient from the
last group afterward. The speedup is about 2% on ultrafast.

The previous version of this patch was reverted due to a bug, which
has now been fixed.
 2015-12-11 16:47:55 +02:00
Ari Lemmetti b78460b02c Optimize another loop 2015-12-11 11:21:43 +02:00
Ari Koivula b32965925e Revert "Further optimize coefficient coding" 
This reverts commit 25462124f8.

That commit broke the bitstream. If it's not good enough to push on Friday
night, it's probably not good enough on Monday morning either.
 2015-12-07 15:12:04 +02:00
Ari Koivula 865c86fef2 Remove unused variable 2015-12-07 10:32:18 +02:00
Ari Koivula 91631a1c36 Merge branch 'coeff-optimization' 2015-12-07 10:25:46 +02:00
Ari Koivula 25462124f8 Further optimize coefficient coding 
Remove the need to count the coefficients by populating the significant
coefficient group map first and finding the last coefficient from the
last group afterward.
 2015-12-07 10:23:01 +02:00
Ari Koivula c94707e6e8 Fix bug with OWF+FME+deblocking 
Increases the MV safety margin of OWF from 2 to 3 when deblocking
is used and 4 when both deblocking and FME are used.

Fractional pixel motion estimation can move the vector one more pixel
down causing checksum error. This fixes that error by increasing the
OWF safety margin and changes the interface, so that different margin
can be used when FME or deblocking are not in use.
 2015-12-04 15:26:56 +02:00
darealshinji fe2ff12244 fix building with autotools 2015-12-03 22:41:24 +01:00
darealshinji b6d3510c2e pkg-config: move -lm to Libs.private 2015-12-03 22:39:27 +01:00
Ari Lemmetti 6fe223c4dc Nonzero calculation magic 2015-12-03 18:29:44 +02:00
Ari Lemmetti f2d8cd4d64 Merge branch 'intra-search-multi' 2015-12-03 17:25:52 +02:00
Ari Lemmetti c4e1552ef6 Replace original rough intra search 2015-12-03 17:13:11 +02:00
Ari Lemmetti ee8c2d0218 Add 4x4 dual SATD for AVX2 2015-12-03 17:13:11 +02:00
Ari Lemmetti 00736fa708 Generate larger than 8x8 dual satd functions with macro 2015-12-03 17:13:11 +02:00
Ari Lemmetti bd3e1922cd Add AVX2 8x8 dual hadamard transform 2015-12-03 17:13:11 +02:00
Ari Lemmetti d575b94357 Implement generic functions for dual sad / satd 2015-12-03 17:13:11 +02:00
Ari Lemmetti 183ee53f47 Add alternative version of rough intra search. 
Calculate two costs simultaneously to exploit larger SIMD registers.
Implementation for dual functions missing currently.
 2015-12-03 17:12:38 +02:00
darealshinji 8ff28ec974 Make dynamic linking easier 2015-12-01 14:34:08 +01:00
Arttu Ylä-Outinen 21e19067fe Extract inter search in a single ref frame. 
Moves code for doing inter search in a single reference frame from
function kvz_search_cu_inter to a new function search_cu_inter_ref.
 2015-11-18 11:16:27 +02:00
Arttu Ylä-Outinen f9f3d5929e Use macros for indexing cu_array in lcu_t. 
Replaces accesses cu_array with macro calls and adds macros
LCU_GET_TOP_RIGHT_CU and CU_GET_CU.
 2015-11-18 11:16:27 +02:00
Arttu Ylä-Outinen 8db8f3d523 Use macro SUB_SCU where possible. 
Replaces expressions like (x & 0x3f) with SUB_SCU(x).
 2015-11-18 11:16:26 +02:00
Arttu Ylä-Outinen 9532d79adb Add macros for indexing cu array in lcu_t. 
- Adds macros LCU_GET_CU and LCU_GET_CU_AT_PX to cu.h.
- Replaces accesses to the cu array of lcu_t by calls to these macros.
 2015-11-18 11:16:26 +02:00
Arttu Ylä-Outinen 39302b0328 Refactor inter_clear_cu_unused. 
Replaces duplicated code with a for loop.
 2015-11-18 11:16:26 +02:00
Arttu Ylä-Outinen 33208ac9fb Add a comment explaining the cu array in lcu_t. 2015-11-18 11:16:25 +02:00
Arttu Ylä-Outinen e0b02599a5 Refactor filter_deblock_edge_ functions. 
Replaces repetitive calls to kvz_filter_deblock_luma and
kvz_filter_deblock_chroma with loops in functions
filter_deblock_edge_luma and filter_deblock_edge_chroma.
 2015-11-18 11:12:32 +02:00
Arttu Ylä-Outinen 43fc6ac419 Mark deblock functions static. 
Marks the following functions static and removes them from filter.h
since they are not used outside the filter module.
- kvz_filter_deblock_luma
- kvz_filter_deblock_chroma
- kvz_filter_deblock_edge_luma
- kvz_filter_deblock_edge_chroma
- kvz_filter_deblock_cu
 2015-11-18 11:12:31 +02:00
Arttu Ylä-Outinen c93a190940 Refactor deblocking filter functions. 
- Replace parameter depth in kvz_filter_deblock_edge_{luma,chroma} with
  length.
- Move checking whether an edge needs to be filtered from functions
  kvz_filter_deblock_edge_{luma,chroma} to functions
  kvz_filter_deblock_{cu,lcu}.
- Use pixel coordinates instead of 8-pixel block coordinates in
  kvz_filter_deblock_cu.
- Add comments.

These changes should make it easier to modify the deblocking filter to
handle SMP and AMP blocks.
 2015-11-18 11:12:31 +02:00
Arttu Ylä-Outinen 863bd1c55d Replace EDGE_ macros with an enum in filter. 
Replaces macros EDGE_HOR and EDGE_VER with enum edge_dir.
 2015-11-18 11:12:30 +02:00
Ari Koivula b6e443f3ce Fix bug in lp-gop parsing 
Unnecessary mod operation resulted in 0 as the reference delta.
 2015-11-16 11:58:57 +02:00
Ari Koivula cfe834bb53 Merge branch 'lowdelay_GOP' 
Conflicts:
	README.md
 2015-11-14 00:05:13 +02:00
Ari Koivula 5ae97b46c6 Remove redundant LP-GOP structures 
These structures can now be defined with the LP-GOP syntax.
The syntax for lb is g4d3r4t1 and for ultralow it is g8d3r1t1.
 2015-11-14 00:01:29 +02:00
Ari Koivula d2de2aa6aa Add "lp-g8d4r2t2" style GOP selection 
This is my own syntax that I've been using when testing this feature.
It allows for defining some simple type of hierarchical GOP structures.
 2015-11-14 00:01:28 +02:00
Ari Koivula b10866cb1e Fix SPS ref pic counts for lowdelay GOP 2015-11-13 23:11:11 +02:00
Ari Koivula a6a713ac02 Use P-slices for lowdelay GOPs 2015-11-13 23:11:11 +02:00
Ari Koivula 0722f461c5 Fix compiling tests on mac 
The mac version of KVZ_GET_TIME macro has many statements, which
prevented it being used inside a for loop statement. Added brackets
to all versions to prevent this issue arising in the future.

Fixes #115.
 2015-11-13 22:57:29 +02:00
Ari Koivula 93637f4683 Move macros in threads.h to KVZ_ namespace 2015-11-13 22:46:32 +02:00
Arttu Ylä-Outinen 95fb2ed9ed Fix unexpected behavior of --deblock option. 
Giving a single number as an argument to --deblock option would enable
deblocking and set both beta and tc to that value. This commit changes
a single number argument to be interpreted as a boolean specifying
whether to enable deblocking or not. As a result, "--deblock 0" can be
now used to disable deblocking.

This fixes deblocking being enabled in all presets.
 2015-11-13 14:22:42 +02:00
Arttu Ylä-Outinen 60ad19d0c8 Fix --deblock option. 
Fixes --deblock option so that it takes a "beta:tc" argument as
advertised in the README and command line help.
 2015-11-13 14:22:42 +02:00
Arttu Ylä-Outinen 8960ce369e Reject extra command line arguments. 
Changes the command line program to print an error and exit instead of
silently ignoring non-option arguments.
 2015-11-13 13:57:00 +02:00
Arttu Ylä-Outinen e42f1351f9 Call config_parse through the api struct in cli. 
Replaces a call to kvz_config_parse with api->config_parse.
 2015-11-09 14:31:04 +02:00
Arttu Ylä-Outinen 87ca9e1856 Drop an unnecessary call to kvz_threadqueue_flush. 
Removes threadqueue dependency from the command line program.
 2015-11-09 14:31:04 +02:00
Arttu Ylä-Outinen b1abe65e83 Move kvz_get_padding to encmain. 2015-11-09 14:31:03 +02:00
Arttu Ylä-Outinen 0eb1e710e6 Move PSNR computation from videoframe to encmain. 
Moves function kvz_videoframe_compute_psnr to encmain and renames it to
compute_psnr. Removes videoframe dependency from the command line
program.
 2015-11-09 13:50:43 +02:00
Arttu Ylä-Outinen 940ada4c0d Mark AVX2 intra filter functions as static. 
Marks functions filter_4x4_avx2, filter_16x16_avx2 and filter_NxN_avx2
static as they are not used outside strategies/avx2/intra-avx2.
 2015-11-09 12:48:20 +02:00
darealshinji f51e3847b6 Fix cross-building on Linux 2015-11-06 21:53:44 +01:00
Marko Viitanen 94bec1b444 Cleanup of mv-rdo, removed unused functions 2015-11-05 14:40:06 +02:00
Marko Viitanen 0cb57961b0 Use dynamically selected get_mvd_cost function for MV candidate selection 2015-11-05 14:31:37 +02:00
Marko Viitanen bb4f50aded Added mv-rdo commandline parameter and use it in presets 2015-11-05 13:59:30 +02:00
Marko Viitanen 4e7e9eefbf Enable usage of MV RDO with a config parameter (in hexbs, tz, frac, bipred) 2015-11-05 12:24:03 +02:00
Marko Viitanen 9a535e1c56 Added missing kvz_ prefixes and fixed some warnings 2015-11-05 09:07:59 +02:00
Marko Viitanen 822c174377 Set cabac to only count bits 2015-11-05 09:07:59 +02:00
Marko Viitanen 1ed0d85020 Added a function for cabac mvd coding cost get_mvd_coding_cost_cabac() 
Conflicts:
	src/rdo.c
 2015-11-05 09:07:59 +02:00
Marko Viitanen 6a2658cc74 Added calc_mvd_cost_cabac() to calculate real bits used for motion vectors 
Conflicts:
	src/rdo.h
	src/search_inter.c

Conflicts:
	src/rdo.c
 2015-11-05 09:07:59 +02:00
Ari Lemmetti fbd0596114 Merge branch 'avx2-pixels-blit' 2015-11-04 11:06:10 +02:00
Ari Lemmetti 57ea7d223b Pass SIMD registers to functions as pointers to fix 32-bit compilation in visual studio 2015-11-04 10:51:26 +02:00
Ari Lemmetti a3855652e9 Add AVX2 version with separate handling of basic blocks and strideless copy. 2015-11-04 10:07:25 +02:00
Ari Lemmetti 0816fbea2c Create generic strategy of blit function 2015-11-04 10:07:25 +02:00
Ari Koivula 5c1ff57f9f Add corresponding option for every "--no-X" option 
Needed in order to turn back options turned off by presets.
 2015-11-04 00:12:26 +02:00
Ari Koivula 8d9e8aad73 Fix lambda calculation to match HM 
The lambda was not being increased for non-key frames and was different
in other ways too. The new implementation matches HM.
 2015-11-03 16:49:42 +02:00
Ari Koivula ba47b3cdb1 Make --preset accept numbers 
Ultrafast corresponds to 0 and placebo to 9.
 2015-11-03 15:46:23 +02:00
Ari Koivula 74ee2f3b27 Redefine presets and include them in README. 2015-11-03 15:26:34 +02:00
Marko Viitanen 27e743a507 Added a commandline option for using a preset 
- Defined presets: ultrafast, superfast, veryfast, faster, fast, medium,
                    slow, slower, veryslow, placebo
 2015-11-03 12:25:06 +02:00
Marko Viitanen 641c204277 Use lowdelay flag in GOP for not using input picture caching 
- Reduced layers to 3 in LB
 2015-11-02 12:36:41 +02:00
Marko Viitanen 9a99f7972f New GOP structure for ultralow delay 2015-11-02 11:33:16 +02:00
Marko Viitanen 388986399f Added a definition for low-delay B GOP structure 2015-11-02 10:53:06 +02:00
Marko Viitanen 821d5c478b Added missing parameter to kvz_strategy_register_picture_generic() 2015-11-02 08:55:54 +02:00
Ari Lemmetti 6dce1f1e33 Update versions for a new release 2015-10-30 17:31:55 +02:00
Ari Lemmetti d71f1b5bd0 Disable incompatible optimizations for 32-bit version 2015-10-24 15:32:27 +03:00
Ari Lemmetti df995d85e8 Utilize AVX2 for dequantization. 2015-10-23 20:17:08 +03:00
Ari Lemmetti cf347e33c4 Move dequant to strategies. Copy generic to AVX2 as well. 2015-10-23 19:53:50 +03:00
Ari Lemmetti 47082738aa ...and the same tricks for quantized reconstruction 2015-10-23 19:44:38 +03:00
Ari Lemmetti 7961ba80d8 Add functions for bigger block sizes to calculate more residual simultaneously and reduce memory accesses 2015-10-23 19:11:56 +03:00
Ari Lemmetti 15edd5060d Load and store multiple elements simultaneously. Use 128-bit wide zero 
test. *wip*
 2015-10-23 17:03:16 +03:00
Ari Lemmetti b37cca87c8 Copy generic to avx2 2015-10-23 17:03:15 +03:00
Ari Lemmetti cad2ea9d6e Move quantize_residual to quant strategies. 2015-10-23 17:03:15 +03:00
Ari Lemmetti c013e58f0c Merge branch 'avx2-faster-angular' 2015-10-23 16:54:35 +03:00
Ari Lemmetti 0c63041ba7 Add filtering functions for different block sizes. Simplify logic a bit to reduce branching. Sorry for the large commit! 2015-10-23 16:54:15 +03:00
Arttu Ylä-Outinen f7b6365db8 Merge pull request #109 from lu-zero/master 
version: Bump
 2015-10-23 12:26:01 +03:00
Luca Barbato 7ecd9c7284 version: Bump 
d5f3778f72 provided a new interface
 2015-10-23 10:02:28 +02:00
Arttu Ylä-Outinen 1cf55f066f Fix memory leak in encoder_headers. 
The header data was not freed when data_out was NULL.
 2015-10-23 09:55:08 +03:00
Arttu Ylä-Outinen a1272e98f8 Prevent disabling VPS from command line. 
Disabling VPS when using the command line encoder would result in an
invalid bitstream.
 2015-10-19 11:25:29 +03:00
Arttu Ylä-Outinen 024fedff57 Disable writing VPS when vps_period is negative. 
Turns vps_period in struct encoder_control_t into a signed value.
Negative values are interpreted as "never send parameter sets."
 2015-10-19 11:25:18 +03:00