Commit Graph

97 Commits

Author SHA1 Message Date
Christophe Gisquet
f1181e4660 fate: add 10bits YUV4:2:2 dnxhd test
It was useful to (accidentally?) spot an overflow in the column pass
of the x86 simple_idct10 implementation.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2015-10-13 04:04:02 +02:00
Christophe Gisquet
2fd14dd8eb avcodec/simple_idct10: improve precision
omse goes from 0.03060703 (which fails for dct-test) to 0.01663750.
This also actually improve the error of decoding the sample generated
by fate-vsynth3-dnxhd1080i-10bit using simple_idct10 to FAANI, which
goes (when resampled to yuv422p) from:
stddev:    0.06 PSNR: 72.28 MAXDIFF:    1
to identical.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2015-10-13 02:10:51 +02:00
Christophe Gisquet
c7e14a279f dnxhddec: use dequantization formula from specs
The current one, while correct, does not yield the best possible
results. The specificiations suggest another formula, which results
in quality gains in the decoded output from fate tests. This
justifies changing said formula.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2015-10-02 18:08:36 +02:00
Peter B
baeb8f5f71 tests: Renamed pix_fmts wording in ffv1 test target name to match pix_fmt parameter.
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2015-08-31 16:14:38 +02:00
Michael Niedermayer
edfcbf7c83 Merge commit '02b7c630875c0bc63cee5ec597aa33baf9bf4e20'
* commit '02b7c630875c0bc63cee5ec597aa33baf9bf4e20':
  h261: Signal freeze picture release for intra frames

Conflicts:
	tests/ref/vsynth/vsynth1-h261
	tests/ref/vsynth/vsynth2-h261

Merged-by: Michael Niedermayer <michaelni@gmx.at>
2015-07-09 00:49:20 +02:00
Michael Niedermayer
a137e50ee5 Merge commit 'dc1de0b958836545339611e9c050a1d4fdded263'
* commit 'dc1de0b958836545339611e9c050a1d4fdded263':
  h261: Set 'still image mode off' in picture header

Conflicts:
	tests/ref/vsynth/vsynth1-h261
	tests/ref/vsynth/vsynth2-h261

Merged-by: Michael Niedermayer <michaelni@gmx.at>
2015-07-09 00:33:52 +02:00
Stian Selnes
02b7c63087 h261: Signal freeze picture release for intra frames
Freeze picture release should be set to 1 when we're responding to a
fast update request. For simplicity we set it for all intra frames,
including those that starts a GOP.

Fixes issue where Tandberg MXP1700 does not recover from packet loss
state since it's waiting for the freeze picture relase indication.

Bug-Id: 873
CC: libav-stable@libav.org

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
2015-07-07 15:11:15 +02:00
Stian Selnes
dc1de0b958 h261: Set 'still image mode off' in picture header
Ref H.261 recommendation section 4.2.1.3, setting the still image flag
to 1 disables still image mode. Some decoders require this in order to
decode the bitstream as normal video.

Fixes H.261 calls to Cisco E20.

Also, reserved (aka spare) bits should be set to 1 unless specified
otherwise.

Bug-Id: 872
CC: libav-stable@libav.org

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
2015-07-07 15:04:55 +02:00
Luca Barbato
e95c7a6185 mov: Preserve the metadata even when bit-exactness is requested
Make sure to not write the custom `encoder` string in that case.

Bug-Id: 845
CC: libav-stable@libav.org
2015-06-30 15:34:38 +02:00
Carl Eugen Hoyos
0f5b996531 lavc/j2kenc: Make jp2 output compatible with Kakadu.
Fixes ticket #4689.
2015-06-30 15:00:38 +02:00
Michael Niedermayer
96e73fa649 avcodec/jpeg2000dec: Merge i_stepsize/2 decoder special case into dequant
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-28 15:26:31 +02:00
Tobias Rapp
2abdc6f477 fate: add some tests for ffv1 level 3 with 8/10/16 bps
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-22 19:09:50 +02:00
Michael Niedermayer
b1fdf81c6e avcodec/jpeg2000dwt: use 32x32->64 multiplies in the 9/7i DWT
This significantly improves the quality when the integer 9/7 transform
is used

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-22 02:27:50 +02:00
Michael Niedermayer
6c7b1597c7 avcodec/jpeg2000: Move H band scaling from wavelet into quantization code
This reduces the number of operations done and is equivalent except for
rounding

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-22 02:00:57 +02:00
Michael Niedermayer
1c495b0bf6 avcodec/jpeg2000: Move gainb handling into the quantization code
thats how the specification defines it, this also improves numerical
accuracy of the integer wavelet implementation. It otherwise should
be equivalent, in case of overflows this can be reverted.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-22 01:36:05 +02:00
Michael Niedermayer
3bcf61f495 avcodec/jpeg2000dwt: Fix value of F_LFTG_X and I_LFTG_X
Even if the jpeg2000 spec uses a wrong value this does not
make mathematics work this way, also this has been corrected in the 2004
version AFAIK

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-21 19:25:35 +02:00
Michael Niedermayer
75651cd63f avcodec/jpeg2000dwt: Fix order of operations in integer decomposition wavelets
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-20 00:53:56 +02:00
Michael Niedermayer
7e2bdea764 avcodec/j2kenc: Use JP2 format by default
Without the JP2 fields its not possible to identify the colorspace

Fixes Ticket538

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-19 21:56:57 +02:00
Michael Niedermayer
4624656797 avcodec/j2kenc: Properly flush the end of the truncated AC stream
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-17 20:39:03 +02:00
Michael Niedermayer
7c3f7e6ed1 avcodec/jpeg2000dwt: Support 9/7 singularity cases on the encoder side
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-17 15:24:14 +02:00
Michael Niedermayer
9841024402 avcodec/jpeg2000dwt: More special cases to handle singularities in 9/7 decode
Fixes Ticket4634

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-17 15:07:50 +02:00
Michael Niedermayer
7c9fcdfabd avcodec/jpeg2000dec: Fix some 5/3 bitexactness issues
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-06-09 02:17:53 +02:00
Michael Niedermayer
729108515d avcodec/snowenc: set dia_size for the first pass of motion estimation too
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-05-10 14:19:56 +02:00
Michael Niedermayer
58201c6e9e avcodec/snowenc: correct fullpel search pattern
No real difference in quality, its a bit slower for the same dia_size as more
vectors are searched for the same dia_size

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-05-10 02:38:05 +02:00
Michael Niedermayer
db6d0a5bd2 avcodec/snowenc: Fix used reference MV in fullpel iterative ME
This very very slightly improves quality, no measurable speed change

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-05-10 01:05:42 +02:00
Michael Niedermayer
4b8a819496 avcodec/ffv1enc: set bits_per_raw_sample for 8bit based formats
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-04-26 21:23:14 +02:00
James Almer
d7f62f033c fate: add cinepak encoder vsynth tests
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-04-23 04:44:45 +02:00
Michael Niedermayer
adfde9a646 avformat/avienc: Fix wrongly placed comment
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-03-19 17:30:28 +01:00
Kevin Wheatley
31c7c0e156 avformat/movenc: Move avid DNxHD padding to the correct spot
Outputting DNxHD into .mov containers 'corrupts' following atoms until end of stsd

ffmpeg and qtdump could not decode pasp/colr atoms in the files made by ffmpeg,
when outputting DNxHD due to the incorrect padding placement. Now we add the
padding in the correct place

Tidy up FATE changes due to padding changes.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-02-20 12:17:14 +01:00
Michael Niedermayer
7d24ccaf00 fate: mjpeg trellis test
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-02-11 22:44:06 +01:00
Kevin Wheatley
7fc33dca39 libavformat: DNxHD in .mov, switch unspecified color_range to mpeg
Avid prefers mpeg range [16-235] by default this change brings
ffmpeg into line with that. To obtain the old behaviour use
'-color_range jpeg' on the command line prior to the ouput
filename.

Signed-off-by: Kevin Wheatley <kevin.j.wheatley@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-02-10 16:46:12 +01:00
Carl Eugen Hoyos
3ea97767e4 Add an ARES atom to extradata when encoding avui.
Allows playback with QuickTime and the original codec.

Tested by Olaf Matthes, olaf matthes gmx de
2015-01-29 01:07:08 +01:00
Michael Niedermayer
aa508a9c5f fate: Fix bitexactness for vsynth3-dnxhd-1080i-colr
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-01-27 03:45:14 +01:00
Kevin Wheatley
7b6f419176 avformat/movenc: Add simplistic 'colr' tag writing support to mov container
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2015-01-26 17:49:50 +01:00
Kieran Kunhya
9a738c27dc v210enc: Add SIMD optimised 8-bit and 10-bit encoders
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
2014-12-05 13:03:49 +00:00
Michael Niedermayer
42801505cd tests: Add vsynth2 which uses the new reference.pnm
The file is already present in git and by using it we can perform more tests
without the need of fate samples

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-12-02 03:39:56 +01:00
Michael Niedermayer
ed49b142bb tests: Rename vsynth2 to vsynth_lena
This avoids confusion with a growing number of vsynth tests

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-12-02 03:02:30 +01:00
Reinhard Tartler
8895bf7b78 Replace lena.pnm
The new reference.pnm is a freely licensed replacement. The photo has
been taken by Reinhard Tartler on August 28 2014, and is licensed under
the expat license as stated at http://www.jclark.com/xml/copying.txt
2014-11-28 17:55:27 -05:00
Kieran Kunhya
36091742d1 v210enc: Add SIMD optimised 8-bit and 10-bit encoders
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-11-26 20:30:47 +01:00
Reimar Döffinger
6369a7b742 xface: Fix encoder crashes due to too small on-stack array.
Also add a FATE test.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
2014-11-23 06:51:18 +01:00
Christophe Gisquet
1086f09da3 dv: more precise weight table for 8x8
It is derived from the actual equations of the specs. In
particular, it is closer to the inverse of what the encoder uses.

fate tests accordingly updated.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-10-25 19:01:45 +02:00
Michael Niedermayer
b0f7de3e7c avcodec/asvenc: fix AAN scaling
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-09-23 17:04:02 +02:00
Christophe Gisquet
f75baa6c9b huffyuvdec: decode the last odd sample
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-08-25 23:24:07 +02:00
Christophe Gisquet
6ee7681723 huffyuvenc: write last odd sample
If width is odd, last sample wouldn't be written.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-08-25 23:12:30 +02:00
Christophe Gisquet
50a35f0d2e fate: add test for old (v1) huffyuv and rgba
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-08-21 02:20:07 +02:00
Eric Lasota
8be23d424f avcodec/roqvideo: use JPEG color range
Signed-off-by: ejlasota <ejlasota@gmail.com>

Also update fate test checksums
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-07-06 23:34:58 +02:00
Christophe Gisquet
e7fc5d53a0 huffyuv: change statistics initialization
Normally, a Laplace distribution is more typical of the residuals
encoded, but for noisy input, it's both better and simpler to be
safe and use a 1/d^2 distribution. Second hunk could use some
renormalization but it has effectively little impact.

Output size of ffvhuff on various 4:2:0 sequences:
context=0,1/d:    851974   27226   1137281
context=0,1/d²:   619081   25069   1051500
context=0,1/d³:   501983   30454   1290561
context=0,lapl:   500650   31754   1304082

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-06-19 06:03:51 +02:00
Josh Allmann
74bc945805 flvenc: Don't over-count metadata.
Over-counting occurs if framerate is not set.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
2014-06-17 21:38:21 +02:00
Christophe Gisquet
02bffc560f huffyuv: fate: add a BGR24 test
It was previously untested.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-06-16 00:38:03 +02:00
Michael Niedermayer
60ab6e2457 avcodec/mpegvideo_enc: fix padding for odd dimensions and interlaced video
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
2014-06-11 05:02:46 +02:00