Commit graph

1280 commits

Author SHA1 Message Date
Marko Viitanen b5756821dd Tuned cost calculations for "final cost"
By adding chroma cost and using 1.5x coeff cost instead of 0.5x, additional 3% (LP) and 1.8% (AI) drop in BD-rate noticed.
2014-03-12 10:35:56 +02:00
Ari Koivula 409b094acf Try and get NxN to work again.
- Moved NxN search to be done on the same level as other searches, as it's
  really not any different from 2Nx2N.
- Produces working bitstream but reconstruction is different.
2014-03-11 19:19:20 +02:00
Marko Viitanen 75dd1f48ef Bugfix: calculate "final cost" only when block is inter or intra 2014-03-11 17:09:08 +02:00
Marko Viitanen 3b59d7c3a3 Implemented "final cost" calculation for best CU on each depth in search
Take SAD between reconstruction and original and add cost from coding and cost for coding coefficients.
2014-03-11 15:01:50 +02:00
Marko Viitanen aa0fb6ccad Store intra coding bit costs to CU structure 2014-03-11 12:15:50 +02:00
Marko Viitanen bdf8166e5c Store inter coding bit costs to CU structure
To be used for calculating better coding cost later
2014-03-11 11:40:39 +02:00
Marko Viitanen d2d877933a bugfix: only fclose recout if it was opened 2014-03-11 09:09:43 +02:00
Marko Viitanen eaf4434ae9 Removed unused parameter cur_cu from inter_get_merge_cand() 2014-03-11 09:04:17 +02:00
Marko Viitanen f2e722fed8 Implemented early search termination when skip mode is used 2014-03-11 08:55:38 +02:00
Ari Koivula 1b2705a042 Add video dimensions to the filename of encoder reconstruction. 2014-03-10 18:10:50 +02:00
Ari Koivula 965d8c377a Make encoder reconstruction obey --debug. 2014-03-10 18:10:49 +02:00
Ari Koivula 7e507fd615 Add --input-res CLI parameter.
- Also set default width and height to 0 because it makes no sense to assume
  dimensions. It's kind of like defaulting to the most common altitude when
  the altimeter is broken.
- Add config_validate to have a place for checks on overall cfg state.
2014-03-10 16:09:33 +02:00
Ari Koivula a644848b36 Add --seek CLI parameter.
Related to issue #12.

Motivation for adding this is that the syntax for frame accurate seeking in
ffmpeg is a little cumbersome. Also it's kind of a standard feature I guess.
2014-03-10 15:14:56 +02:00
Marko Viitanen a935aa5d96 Untabified encmain.c and nal.c 2014-03-10 12:10:40 +02:00
Marko Viitanen a7dade5069 Tuned inter MV coding costs to achieve BD-rate reduction
Compared to version with no inter costs, around -2% BD-rate in LP configuration
2014-03-10 12:08:09 +02:00
Marko Viitanen 1bd79c719e Fixed MV candidate derivation in search 2014-03-10 10:16:30 +02:00
Marko Viitanen 9a4b0d2cc7 Added newline after printing of real video frame size 2014-03-10 10:16:30 +02:00
Marko Viitanen c7e4861dbf Added coding cost calculations to MV search 2014-03-10 10:16:30 +02:00
Ari Koivula 9dde96f25e Refactor search_cu_intra.
- Formatted.
- Renamed and consolidated variables.
- Removed unused x and y pos from intra_prediction function.
2014-03-07 16:25:25 +02:00
Ari Koivula 470f36ce34 Remove left-over commented out code. 2014-03-07 12:44:45 +02:00
Ari Koivula 5d9aa3409b Merge branch 'new-search-wip'
Conflicts:
	src/encoder.c
	src/intra.c
	src/intra.h
	src/search.c
2014-03-06 19:00:53 +02:00
Ari Koivula 47af5207c5 Remove dead code and fix white space. 2014-03-06 18:35:17 +02:00
Marko Viitanen 877e156cb8 Fixed the case where transform split is used (always on depth 0)
After this fix, inter coding should be working.
2014-03-06 14:52:58 +02:00
Marko Viitanen c9d35aeacf Removed unused block data setting functions from picture.c/.h 2014-03-05 17:22:36 +02:00
Marko Viitanen c3e47c6a8d Removed redundant coeff_y/u/v variables from CU 2014-03-05 17:18:03 +02:00
Marko Viitanen b00f4c8dbf Fixed transform offset value and set tr_depth to correct value when depth 0 2014-03-05 17:10:17 +02:00
Ari Koivula 1cf03cfb37 Add --no-rdoq parameter. 2014-03-05 17:02:21 +02:00
Ari Koivula ca51f660d5 Fix fluctuation of coefficients on identical frames when RDOQ is on.
Moves CABAC context initialization to take place before search. This fixes
an issue with RDOQ returning different coefficients for identical adjacent
frames.

- This actually probably worsens BD-rate a little for all frames except the
  first one because we were using last frames final CABAC context for every
  LCU and now we are using initialized CABAC contexts. The fix is to encode
  the LCU before we start compressing the next LCU so we can update CABAC
  contexts.
2014-03-05 17:02:13 +02:00
Ari Koivula 8f3d5045b0 Add --no-rdoq parameter. 2014-03-05 17:00:08 +02:00
Ari Koivula 139e00f064 Fix fluctuation of coefficients on identical frames when RDOQ is on.
Moves CABAC context initialization to take place before search. This fixes
an issue with RDOQ returning different coefficients for identical adjacent
frames.

- This actually probably worsens BD-rate a little for all frames except the
  first one because we were using last frames final CABAC context for every
  LCU and now we are using initialized CABAC contexts. The fix is to encode
  the LCU before we start compressing the next LCU so we can update CABAC
  contexts.
2014-03-05 17:00:08 +02:00
Marko Viitanen c5842e1769 More fixes to B0 mv candidate selection
After this fix, inter coding works in depth != 0
2014-03-05 12:53:35 +02:00
Marko Viitanen 278042aaac Fixed top-right CU copy to lcu structure 2014-03-04 16:32:31 +02:00
Marko Viitanen 91e83db3c6 Bugfix for inter_get_spatial_merge_candidates() b2 block checking
Wrong variable was used to check for x-border
2014-03-04 15:34:32 +02:00
Marko Viitanen df669ca7df Added skip mode checking to the new lcu search 2014-03-04 12:48:19 +02:00
Marko Viitanen bddadbb4ed Inter coding fixed and added call to encode_transform_tree() 2014-03-03 17:42:44 +02:00
Marko Viitanen 09d4b47395 Fixed CU array calculations in inter_get_spatial_merge_candidates() 2014-03-03 16:25:33 +02:00
Marko Viitanen 17b256b5e3 Converted inter mv candidate functions to use lcu structure
Implemented merge and mvd candidate search
Added initialization of the whole work_tree
2014-03-03 16:08:35 +02:00
Marko Viitanen 1f82239367 Implemented inter search and reconstruction with lcu structure 2014-03-03 14:51:36 +02:00
Marko Viitanen b5fca8b206 Added missing chroma ref pixel init and fixed pic width/height in intra ref border
intra_build_reference_border() was given wrong picture width and height on chroma.
Chroma top-left reference pixels were not copied on init_lcu_t()
2014-03-03 12:21:55 +02:00
Ari Koivula 448285a56a Fix bug in intra mode coding. 2014-02-28 17:36:04 +02:00
Marko Viitanen f1882a0d82 Fixed CU type initialization and PU index calculations 2014-02-28 15:56:27 +02:00
Ari Koivula c05e91472d Fix intra reference border construction on LCU boundary. 2014-02-28 15:43:59 +02:00
Marko Viitanen 83e2079587 Fixed chroma stride in intra_build_reference_border() 2014-02-28 15:24:47 +02:00
Ari Koivula 19f67d4466 Fix LCU reference pixel indexing in reference border construction. 2014-02-28 14:22:13 +02:00
Ari Koivula 768d54ff4c Fix wrong array offset in intra search. 2014-02-28 14:22:13 +02:00
Marko Viitanen 0a6d9bda41 Fixed stride values in transform_chroma() 2014-02-27 17:47:00 +02:00
Marko Viitanen 4027de6e1f Fixed reconstruct_chroma() stride values 2014-02-27 16:48:51 +02:00
Ari Koivula 6e11eda133 Fix coeff flag initialization.
They should be initialized, even if none exist.
2014-02-27 16:43:20 +02:00
Marko Viitanen 9968fd46b0 Fixed chroma array index calculations in transform_tree() and intra_recon_lcu() 2014-02-27 16:42:32 +02:00
Ari Koivula 71d9ab7e26 Fix copying-up of coefficients during search. 2014-02-27 15:58:47 +02:00