ffmpeg | branch: master | Christophe Gisquet |
Sun Jul 9 12:56:35 2017 +| [da888b790af779a7489068c25f9e7ab8ac653d41] |
committer: Paul B Mahol
avcodec/utvideodec: add vlc multi support
Faster decoding, by average 50% faster overall.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git
ffmpeg | branch: master | Christophe Gisquet |
Sat Jun 11 09:31:13 2016 +0200| [8028b8260e6d0a4a469529ec5c5d504782a5f875] |
committer: Paul B Mahol
avcodec/huffyuvdec: don't prepare unnecessary joint tables
The number of planes can be lower than 4.
> http://git.videolan.org/gi
ffmpeg | branch: master | Christophe Gisquet |
Wed Aug 6 07:43:39 2014 +| [3c504bc3599f00bfc5923adc114beef34bce11d0] |
committer: Anton Khirnov
x86: deduplicate some constants
Signed-off-by: Anton Khirnov
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Christophe Gisquet |
Tue Dec 27 09:18:51 2016 +| [e3312b3746ef364c3101381bada57f8ef75f175a] |
committer: Michael Niedermayer
MAINTAINERS: update
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=e3312b3746ef364c3101381bada57f8ef75f175a
---
MAINTAIN
ffmpeg | branch: master | Christophe Gisquet |
Tue Oct 13 21:21:41 2015 +0200| [b6e8efb082c284091e267a29b6060b97692d1740] |
committer: Michael Niedermayer
fate: add 12bpp sample
Allows testing simple_idct12 correctness/bitexactness, as the sample
was generated using faani as idct.
Signed-off
ffmpeg | branch: master | Christophe Gisquet |
Sun May 1 15:33:47 2016 +0200| [9ca16bdd3f0461b40d369080647747ae70715daf] |
committer: Michael Niedermayer
lossless audio dsp: unroll
The loops are guaranteed to be at least multiples of 8, so this
unrolling is safe but allows exploiting
ffmpeg | branch: master | Christophe Gisquet |
Thu May 5 11:06:09 2016 +0200| [01938585f4cedbcabb3c879214c24b3fd4f91dcf] |
committer: Rostislav Pehlivanov
vc2: fate tests
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=01938585f4cedbcabb3c879214c24b3fd4f91dcf
---
tests/f
ffmpeg | branch: master | Christophe Gisquet |
Sun May 1 15:33:46 2016 +0200| [9630b3fc06d8ccfab4847ecddaed96526cf4dd19] |
committer: Michael Niedermayer
x86: lossless audio: SSE4 madd 32bits
The unique user so far is wmalossless 24bits. The few samples tested show an
order of 8, so more
ffmpeg | branch: master | Christophe Gisquet |
Thu May 5 11:06:08 2016 +0200| [9c1aa14bf0b88da9f91dc114519e725cbd69180e] |
committer: Rostislav Pehlivanov
vc2enc: prevent random data
The slice prefix is 0 in the reference encoder and the decoder ignores it.
Writing 0 there seems like the
ffmpeg | branch: master | Christophe Gisquet |
Sun May 1 12:34:29 2016 +0200| [7d453aaf653d1e44c1ce62b24a2c5f85d22b43fc] |
committer: Paul B Mahol
wmalossless: allow calling madd_int16
This is done by actually handling the "prev_values" in the cascaded LMS data
as if it were int
ffmpeg | branch: master | Christophe Gisquet |
Mon Apr 18 12:53:21 2016 +0200| [79e86640ff5506e05ab4702a7e13367c8445a729] |
committer: Michael Niedermayer
fate: wma: add lossless 24bits tests
Should evaluate coefficients and raw pcm tiles.
Signed-off-by: Michael Niedermayer
> h
ffmpeg | branch: master | Christophe Gisquet |
Sun May 1 11:18:09 2016 +0200| [532e9370835d8afb7d20a86b652585b44ada627b] |
committer: Paul B Mahol
avcodec/wmalosslessdec: silence a sample request
16bits samples with CDLMS orders of 8 are currently unsupported, but have never
been encountered
ffmpeg | branch: master | Christophe Gisquet |
Wed Oct 14 19:20:49 2015 +0200| [96b165fae24b84a703728a8d34ef1ecb944471ab] |
committer: Michael Niedermayer
dnxhd: interleave AC levels and flags
This allows more efficient access to the array as the level and flags
are contiguous. Around 4
ffmpeg | branch: master | Christophe Gisquet |
Tue Oct 13 22:02:41 2015 +0200| [234369d0fd04f1ddfc10d6a73a1a0aa5cb7ea17c] |
committer: James Almer
dnxhdenc: fix access outside of image
This is the same test as for the 8bit case.
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Christophe Gisquet |
Tue Oct 13 21:21:39 2015 +0200| [74c414202f0eda01571dc362a4b4cb342dc1fb64] |
committer: Michael Niedermayer
x86: simple_idct10_template: use const
This avoid going through constants.c while still sharing them
with proresdsp.asm
Reviewed-by
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 12 19:37:50 2015 +0200| [b6594a9605debcfddc049627808d5319f1df537f] |
committer: Michael Niedermayer
x86: dct-test: add more idcts
In particular for 10 and 12 bits.
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 12 19:37:49 2015 +0200| [7ece8b50b19e140ace13eda6f1a9f45f868c2528] |
committer: Michael Niedermayer
x86: simple_idct: 12bits versions
On 12 frames of a 444p 12 bits DNxHR sequence, _put function:
C: 78902 decicycles in idct
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 12 19:37:47 2015 +0200| [4369b9dc7b2b0da594223ce46615ba8e2b4cead6] |
committer: Michael Niedermayer
x86: simple_idct(_put): 10bits versions
Modeled from the prores version. Clips to [0;1023] and is bitexact.
Bitexactness requires to add
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 12 19:37:46 2015 +0200| [e652f69b354bc6b5819012979985794cfd2805c9] |
committer: Michael Niedermayer
x86: simple_idct10_template: fix overflow in pass
When the input of a pass has 15 or 16 bits of precision (in particular
the column pass
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 12 19:37:43 2015 +0200| [f1181e466062541b6562f6c06dccc7a209ec76c5] |
committer: Michael Niedermayer
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
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 12 19:37:45 2015 +0200| [2fd14dd8eb66dc5dd14254d0b758fb80d44b3140] |
committer: Michael Niedermayer
avcodec/simple_idct10: improve precision
omse goes from 0.03060703 (which fails for dct-test) to 0.01663750.
This also actually improve
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 12 19:37:44 2015 +0200| [e9a68b0316ab127098ac4c24a6762ce68980bd23] |
committer: Michael Niedermayer
x86: prores: templatize 10 bits simple_idct
This should be reused for a generic simple_idct10 function.
Requires a bit of trickery to
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 12 19:37:42 2015 +0200| [9f3bfe30dd99b093a469e566dbd95448b32f80e9] |
committer: Michael Niedermayer
mpegvideo: dnxhdenc: permute 10bits content
Dequant or encoding were trying to reverse a scan that hadn't been
applied...
Signed-o
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 5 17:37:38 2015 +0200| [25d3013e36152635ca992b63469f6d5b890aaa83] |
committer: Michael Niedermayer
fate: add DNxHD/HR tests
Currently only 2 profiles are evaluated because they are the only 2
with distributed test sequences.
- CID 1260
ffmpeg | branch: master | Christophe Gisquet |
Thu Oct 8 12:30:51 2015 +0200| [1896ac4abfe2528e3b22a988acf389a7652c3ef6] |
committer: Michael Niedermayer
build system: fix dependency generation
When trying to build the binary dct-test under MSYS2/Win64, the
makefile rule does not have the
ffmpeg | branch: release/2.8 | Christophe Gisquet
| Fri Sep 25 18:57:17 2015 +0200|
[f5f9c166a1b48212262974d74a5524e180d8f86f] | committer: Michael Niedermayer
dnxhddec: decode and use interlace mb flag
This bit is 1 in some samples, and seems to coincide with interlaced
mbs and CID1260. 2008
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 5 20:44:46 2015 +0200| [5911eeb033c384ba0842f7e2e11d737f1d479dd9] |
committer: Michael Niedermayer
dnxhdenc: mark CID 1260 encoder experimental
The MBAFF handling recently introduced on the decoder side shows that
the encoder does not
ffmpeg | branch: master | Christophe Gisquet |
Mon Oct 5 20:44:45 2015 +0200| [79cfb36f92bd8f6dcd456d00584d270b19004af3] |
committer: Michael Niedermayer
dnxhddata: introduce and use MBAFF flag
MBAFF-like handling of interlaced content in CID 1260 is different from
the other CIDs, and in
ffmpeg | branch: master | Christophe Gisquet |
Sun Oct 4 10:06:28 2015 +0200| [71199ee9077da7d92a8728e2a694fb1ab31488a7] |
committer: Michael Niedermayer
dnxhddec: better support for 4:4:4
Profiles 1256 & 1270 (currently) signal at the frame header and MB
levels the colorspace used, ei
ffmpeg | branch: master | Christophe Gisquet |
Sat Oct 3 18:59:20 2015 +0200| [66db504f038afa02aaecb94a81f549e431211a47] |
committer: Michael Niedermayer
dnxhd: add CID 1270
This a 4:4:4 10 bits profile, where image size is not fixed by the
profile, and which strays a bit outside the old
ffmpeg | branch: master | Christophe Gisquet |
Sat Oct 3 18:59:17 2015 +0200| [1f4ff53aea7c5090f31cd1323d95f7c407c9b2bb] |
committer: Michael Niedermayer
dnxhddec: Introduce DNXHD_VARIABLE
Currently not used, but will be used to indicate that a CIDEntry field
is not set, because it is
ffmpeg | branch: master | Christophe Gisquet |
Sat Oct 3 18:59:18 2015 +0200| [3a6a432b58fab042cfd90bf4a77e30bfe7de6a07] |
committer: Michael Niedermayer
isom: add support for DNxHR codec family
This is actually similar to DNxHD.
Signed-off-by: Michael Niedermayer
> h
ffmpeg | branch: master | Christophe Gisquet |
Sat Oct 3 18:59:15 2015 +0200| [4e7cba1d8a9c62016fb70fd13aca9df44d564481] |
committer: Michael Niedermayer
dnxhd: profile flags
Move the 'interlaced' flag to this element (arbitrarily set to 16bits).
This should allow better detection
ffmpeg | branch: master | Christophe Gisquet |
Fri Oct 2 21:00:44 2015 +0200| [9139a7e84f759edbff4f1ba0cd145dc07e411379] |
committer: Michael Niedermayer
dnxhddec: cleanup frame header parsing
Rely more on the actual syntax from the specs (also seen in the
encoder code).
Signed-off-by
ffmpeg | branch: master | Christophe Gisquet |
Sat Oct 3 18:59:16 2015 +0200| [52256860d6e3eb003294659ae015cd1f1b3ffdfc] |
committer: Michael Niedermayer
dnxhdenc: do not select 4:4:4 profiles
The encoder can only deal with 4:2:2.
Signed-off-by: Michael Niedermayer
> h
ffmpeg | branch: master | Christophe Gisquet |
Mon Sep 28 14:03:36 2015 +0200| [f827a170052e16445366cd34ca0cb6c12fd000a6] |
committer: Michael Niedermayer
blockdsp: reindent after parameter removal
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=com
ffmpeg | branch: master | Christophe Gisquet |
Thu Oct 1 16:54:57 2015 +0200| [e54d7e4e8ef7d778e2ddc5a60bf7307ed55d6333] |
committer: Vittorio Giovara
dnxhddata: Deduplicate cid 1256 chroma table
CID 1256 is specified as using the same table for luma and chroma,
which is the same as CID 1235
ffmpeg | branch: master | Christophe Gisquet |
Thu Oct 1 22:01:27 2015 +0200| [74ef5449a6fb64f4c4be4bfd470dc17de5ecb561] |
committer: Michael Niedermayer
dnxhddec: init scantable once permutation is set
Otherwise, the dsp may change without its permuation being applied.
Signed-off-by
ffmpeg | branch: master | Christophe Gisquet |
Thu Oct 1 22:01:26 2015 +0200| [c7e14a279fa7348db10ec824bb2d67858cb1c1ca] |
committer: Michael Niedermayer
dnxhddec: use dequantization formula from specs
The current one, while correct, does not yield the best possible
results. The
ffmpeg | branch: master | Christophe Gisquet |
Wed Sep 30 10:14:59 2015 +0200| [7e4070d2e70bbd0e70dd83c06089e465a0eda704] |
committer: Michael Niedermayer
dnxhddec: initialize with mb-aligned dimensions
The coded size is a multiple of the macroblock size, which is 16.
Signed-off-by: Michael
ffmpeg | branch: master | Christophe Gisquet |
Mon Sep 28 13:59:23 2015 +0200| [562ba4a827ceb9ed5b7d056484a9c2312a5458c5] |
committer: Michael Niedermayer
blockdsp: remove high bitdepth parameter
It is only (mis-)used to set the dsp fucntions clear_block(s). But
these functions always work on
ffmpeg | branch: master | Christophe Gisquet |
Mon Sep 28 17:30:55 2015 +0200| [e3242c021980e0a2688c27cb50674e7cd14c282a] |
committer: Michael Niedermayer
dnxhddata: deduplicate table
CID 1256 is specified as using the same table for luma and chroma,
which is the same as CID 1235 luma table
ffmpeg | branch: master | Christophe Gisquet |
Sat Sep 26 17:38:22 2015 +0200| [c49cbecbae5a42f4ca004197b0118cc50aaaca2e] |
committer: Vittorio Giovara
dnxhddec: Decode and use interlace mb flag
This bit is 1 in some samples, and seems to coincide with interlaced
mbs and CID1260. 2008 specs
ffmpeg | branch: master | Christophe Gisquet |
Sun Sep 27 10:09:08 2015 +0200| [578f721f5dcf5577e0b06447082497319ccb7a39] |
committer: Michael Niedermayer
dnxhddec: add my contributions
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Christophe Gisquet |
Sun Sep 27 10:09:06 2015 +0200| [6110a55b7d0d22b1d23e195681b4cae99df07e34] |
committer: Michael Niedermayer
dnxhddec: reindent/cosmetics
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Christophe Gisquet |
Sun Sep 27 10:09:05 2015 +0200| [235381e674bf0921cbb7871946e2ff08bd84d6ba] |
committer: Michael Niedermayer
dnxhddec: use unsafe bitstream reader
Each line is padded by the format, and errors are now reported and
stop the decoding. Around 5
ffmpeg | branch: master | Christophe Gisquet |
Sun Sep 27 10:09:04 2015 +0200| [b8b8e82ea14016b2cb04b49ecea57f836e6ee7f8] |
committer: Michael Niedermayer
dnxhddec: check and report bitstream errors
This only occur when an overrun in coefficient decoding is
detected.
Signed-off-by: Michael
ffmpeg | branch: master | Christophe Gisquet |
Sun Sep 27 10:09:03 2015 +0200| [5c6e3a019c5cbcaeb17f3ae0da55070867d6f63d] |
committer: Michael Niedermayer
dnxhddec: simplify block parsing calls
Fewer arguments, less duplicated code.
Signed-off-by: Michael Niedermayer
> h
ffmpeg | branch: master | Christophe Gisquet |
Sun Sep 27 10:09:02 2015 +0200| [8e8ed57ea7d33fa2b872884ead5ec3033637ce92] |
committer: Michael Niedermayer
dnxhddec: remove unused qscale parameter
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Christophe Gisquet |
Fri Sep 25 08:56:43 2015 +0200| [08a7510fcad552bfcda644e4ba047605de15fab5] |
committer: Michael Niedermayer
dnxhddec: implement slice multithreading
Around 3x speedup with 4 threads. Maybe more mb lines should be
batched per thread, but that
ffmpeg | branch: master | Christophe Gisquet |
Fri Sep 25 17:07:38 2015 +0200| [dc218bd08305656d97ad51c57ee4ab2040cd2425] |
committer: Michael Niedermayer
dnxhddec: parse and print adaptive color transform
Indicates a YCbCr->RGB transform at the block level. Although nothing
explicitly sta
ffmpeg | branch: master | Christophe Gisquet |
Sat Sep 26 14:31:55 2015 +0200| [e4b7c3a9f21b9be84fe3ba9aed953c60720c6459] |
committer: Michael Niedermayer
dnxhddec: proper rule for interlaced mb flag
It currently only applies to CID 1260, but this flag is dependent on
a higher-level flag
ffmpeg | branch: master | Christophe Gisquet |
Fri Sep 25 23:25:07 2015 +0200| [e19ea0218aeec60d5dc2c55b6db310242b7757e7] |
committer: Michael Niedermayer
dnxhddec: indicate colorspace
It is supposed to only old BT.709 colorspaces.
Signed-off-by: Michael Niedermayer
> h
ffmpeg | branch: master | Christophe Gisquet |
Fri Sep 25 18:57:17 2015 +0200| [2801a1352dc8682b028e53880f9847fcb2116947] |
committer: Michael Niedermayer
dnxhddec: decode and use interlace mb flag
This bit is 1 in some samples, and seems to coincide with interlaced
mbs and CID1260. 2008
ffmpeg | branch: master | Christophe Gisquet |
Wed Sep 23 15:06:02 2015 +0200| [f94af8d32ae3c2df19d3e0d27ac230e9bd742a91] |
committer: Michael Niedermayer
fate: add chroma position scale test
The sample position is made weird and non-nominal to force catching
such issues as default values or
ffmpeg | branch: master | Christophe Gisquet |
Wed Sep 23 15:06:01 2015 +0200| [552faecf4b1c60ac7fe805653fa3189a15bc21f0] |
committer: Michael Niedermayer
vf_scale: conditionally override chroma position
For yuv420p, the chroma position is unilaterally overriden, even
if ffmpeg's command
ffmpeg | branch: master | Christophe Gisquet |
Sat Aug 1 17:27:34 2015 +0200| [f5e486f6f8c242bb2be01ad3ae952b5733ba1113] |
committer: Anton Khirnov
x86inc: Fix instantiation of YMM registers
Signed-off-by: Henrik Gramner
Signed-off-by: Anton Khirnov
> http://git.videolan.org/gitweb.
ffmpeg | branch: release/2.4 | Christophe Gisquet
| Sat Mar 14 11:48:59 2015 +|
[b9b56ded7dd4987fc3395bc58320c42493c41a62] | committer: Michael Niedermayer
eac3dec: fix scaling
This is the remaining error, the output on the SPX samples,
respectively csi_miami_stereo_128_spx.eac3 and
ffmpeg | branch: release/2.4 | Christophe Gisquet
| Sat Mar 14 11:48:58 2015 +|
[691f9dee0060458e5d3c36346af8c41cf9e74500] | committer: Michael Niedermayer
ac3_fixed: fix computation of spx_noise_blend
It was set to 1 instead of sqrt(3)
Signed-off-by: Michael Niedermayer
(cherry picked
ffmpeg | branch: release/2.4 | Christophe Gisquet
| Sat Mar 14 11:48:57 2015 +|
[40dab450503125f22a82cebf03453b2bdd65d748] | committer: Michael Niedermayer
ac3_fixed: fix out-of-bound read
Should also improve decoding, but actually doesn't...
Signed-off-by: Michael Niedermayer
(c
ffmpeg | branch: release/2.5 | Christophe Gisquet
| Sat Mar 14 11:48:59 2015 +|
[808d5444c4072dd5c3f6e7f7341d0c94f54133e5] | committer: Michael Niedermayer
eac3dec: fix scaling
This is the remaining error, the output on the SPX samples,
respectively csi_miami_stereo_128_spx.eac3 and
ffmpeg | branch: release/2.5 | Christophe Gisquet
| Sat Mar 14 11:48:57 2015 +|
[c657b08fd76a5d09bb36db13dedca839b5bc1ede] | committer: Michael Niedermayer
ac3_fixed: fix out-of-bound read
Should also improve decoding, but actually doesn't...
Signed-off-by: Michael Niedermayer
(c
ffmpeg | branch: release/2.5 | Christophe Gisquet
| Sat Mar 14 11:48:58 2015 +|
[6915dd49c7c6bf1f2b10f880cf1bb5a11c1d790d] | committer: Michael Niedermayer
ac3_fixed: fix computation of spx_noise_blend
It was set to 1 instead of sqrt(3)
Signed-off-by: Michael Niedermayer
(cherry picked
ffmpeg | branch: release/2.6 | Christophe Gisquet
| Sat Mar 14 11:48:58 2015 +|
[7edd380668457782e06093320c41c4f060635861] | committer: Michael Niedermayer
ac3_fixed: fix computation of spx_noise_blend
It was set to 1 instead of sqrt(3)
Signed-off-by: Michael Niedermayer
(cherry picked
ffmpeg | branch: release/2.6 | Christophe Gisquet
| Sat Mar 14 11:48:57 2015 +|
[26bed98d64705348fdc684c2d62c305045dedaec] | committer: Michael Niedermayer
ac3_fixed: fix out-of-bound read
Should also improve decoding, but actually doesn't...
Signed-off-by: Michael Niedermayer
(c
ffmpeg | branch: release/2.6 | Christophe Gisquet
| Sat Mar 14 11:48:59 2015 +|
[7ff46a20d5cc2d746d7396cc80120cb77e5fd90e] | committer: Michael Niedermayer
eac3dec: fix scaling
This is the remaining error, the output on the SPX samples,
respectively csi_miami_stereo_128_spx.eac3 and
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 14:26:16 2015 +0100| [3814f92fd77f8f09fa6f88333b8d06512b6ced82] |
committer: Michael Niedermayer
lavc: use LOCAL_ALIGNED instead of DECLARE_ALIGNED
The later may yield incorrect code for on-stack variables.
Signed-off-by: Michael
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 14:30:18 2015 +0100| [238db7cc56be4c4e71ff98758b7b8ef077ebd4ca] |
committer: Michael Niedermayer
x86: lavc: use LOCAL_ALIGNED instead of DECLARE_ALIGNED
The later may yield incorrect code for on-stack variables.
Signed-off-by: Michael
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 15:30:27 2015 +| [98cccdd91c9aab9e031753f6feb81be0c44a9c8e] |
committer: Michael Niedermayer
ppc: lavc: use LOCAL_ALIGNED instead of DECLARE_ALIGNED
The later may yield incorrect code for on-stack variables.
Signed-off-by: Michael
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 15:30:26 2015 +| [5d38c628b0a84fafc220f070d9d4cf8fbfe96447] |
committer: Michael Niedermayer
ppc: libswscale: use LOCAL_ALIGNED instead of DECLARE_ALIGNED
The later may yield incorrect code for on-stack variables.
Signed-off-by
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 11:49:00 2015 +| [73ebbfdff0bb2c22e3b5c14b11485e5798a317fd] |
committer: Michael Niedermayer
ac3dec: cosmetics
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 11:48:58 2015 +| [c4bf3833f4663fd484441907f73c5bc4700021a4] |
committer: Michael Niedermayer
ac3_fixed: fix computation of spx_noise_blend
It was set to 1 instead of sqrt(3)
Signed-off-by: Michael Niedermayer
> h
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 11:48:59 2015 +| [0c3339f4bd7aceebfd74deb437ba2e5c04ef3d0e] |
committer: Michael Niedermayer
eac3dec: fix scaling
This is the remaining error, the output on the SPX samples,
respectively csi_miami_stereo_128_spx.eac3 and
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 11:48:57 2015 +| [b0834400608b3980c06bf6d2cf747116e60d10c7] |
committer: Michael Niedermayer
ac3_fixed: fix out-of-bound read
Should also improve decoding, but actually doesn't...
Signed-off-by: Michael Niedermayer
&
ffmpeg | branch: master | Christophe Gisquet |
Tue Mar 10 23:11:54 2015 +| [15ce160183c61fbd98915e07879f1225fcf1c080] |
committer: Michael Niedermayer
x86: xvid_idct: SSE2 merged add version
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 07:57:42 2015 +| [4eb4451be191a0144ee6af0650f3798b526b9b2c] |
committer: Michael Niedermayer
x86: dct-test: fix compilation for prores
When the decoder is deactivated, the x86-optimized versions are
not compiled, resulting in a
ffmpeg | branch: master | Christophe Gisquet |
Tue Mar 10 23:11:53 2015 +| [decd5193e1d409a4d8dd4c55ca46467d51c86976] |
committer: Michael Niedermayer
x86: xvid_idct: merged idct_put SSE2 versions
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git
ffmpeg | branch: master | Christophe Gisquet |
Sat Mar 14 07:57:43 2015 +| [8200575d849473516cfe0a76cfceecdf9099a218] |
committer: Michael Niedermayer
x86: dct-test: evaluate prores idct avx version
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git
ffmpeg | branch: master | Christophe Gisquet |
Tue Mar 10 23:11:52 2015 +| [c3bf52713a8485f5bcba4c37ae7373c6b67cd1eb] |
committer: Michael Niedermayer
x86: xvid_idct: port MMX iDCT to yasm
Also reduce the table duplication with SSE2 code, remove duplicated
macro parameters.
Signed-off-by
ffmpeg | branch: master | Christophe Gisquet |
Tue Mar 10 23:11:51 2015 +| [2999bd7da2cf6343592d4eb6da3cd0748354775d] |
committer: Michael Niedermayer
x86: xvid_idct: port SSE2 iDCT to yasm
The main difference consists in renaming properly labels, and
letting yasm select the gprs for
ffmpeg | branch: master | Christophe Gisquet |
Sun Mar 8 17:54:25 2015 +0100| [2e5605f6fa0a2a633afaa096bd60befb5e4ab431] |
committer: Michael Niedermayer
x86: Makefile: fix DBG parameter evaluation
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=com
ffmpeg | branch: master | Christophe Gisquet |
Mon Mar 2 18:17:28 2015 +0100| [efd3f407e50ec04eb9b90825bf8436f221e65de8] |
committer: Michael Niedermayer
hevc/sao: use unaligned copy
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Christophe Gisquet |
Fri Feb 27 16:10:38 2015 +0100| [60bdc868b9c291e4fdd58004d0c839719c23902e] |
committer: Michael Niedermayer
lavf/mov: correct printf specifier
The type is int64_t.
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Christophe Gisquet |
Thu Feb 5 16:00:11 2015 +0100| [e87e20613b013aa56b183c7e956ddaeaa76efa51] |
committer: Michael Niedermayer
lavc/pthread_slice: release entries
When calling ff_alloc_entries, a number of entries are created.
They are never freed, as running fate
ffmpeg | branch: master | Christophe Gisquet |
Fri May 10 11:32:58 2013 +0100| [e0c7ba40cf0bb521430f4a808ed1f1a408093f9c] |
committer: James Almer
takdec: pad filter coeff buffer for DSP functions
Currently, a part of the loop has to be handled in plain C, because of DSP
function requirements
ffmpeg | branch: master | Christophe Gisquet |
Wed Feb 18 10:51:17 2015 +0100| [3fdcb8ccdc2cf5668e02134a815e5a7acc39f763] |
committer: Michael Niedermayer
Makefile: improve DBG option for asm
This improves the cleanup, dependency generation and DBG command issuing.
Signed-off-by: Michael
ffmpeg | branch: master | Christophe Gisquet |
Mon Feb 16 20:12:04 2015 +0100| [398f531915602c0b5d55b3d0f23a01fc1a3e6e5a] |
committer: Michael Niedermayer
x86: hevc_mc: fewer xmm regs used in epel h/v
11 xmm regs seem only required for avx2.
Reviewed-by: Mickaƫl Raulet
> h
ffmpeg | branch: master | Christophe Gisquet |
Sun Feb 8 12:18:27 2015 +0100| [bc8e044d8c3277958c920091412b77d4e7411d96] |
committer: Michael Niedermayer
x86/doc/Makefile: DBG=1 to preprocess external asm
The macro hell sometimes make it difficult to trace the source of
an error, so it is
ffmpeg | branch: master | Christophe Gisquet |
Tue Feb 3 20:40:41 2015 +0100| [d9293c776ed9129acaf3d41fb727380c293dc943] |
committer: Michael Niedermayer
x86inc: Correctly warn on use of SSE2 instructions in SSE functions
SSE2 instructions that are XMM-implementations of pre-existing MMX
ffmpeg | branch: master | Christophe Gisquet |
Sat Feb 7 18:49:38 2015 +| [89cb4995fa0cc9375d9bb246512a82037ffeedc1] |
committer: Michael Niedermayer
x86: hevc_mc: save 1 gpr in epel filter loading
The 3*stride value stored in r3src can be loaded much later,
so use r3src instead of a
ffmpeg | branch: master | Christophe Gisquet |
Thu Feb 5 12:37:52 2015 +0100| [b533949813cc8eaa2eb68142b0fa384c231ab6c9] |
committer: Michael Niedermayer
x86: hevc: remove a parameter to WP internals
The second stride is always the internal buffer one, MAX_PB_SIZE (times 2 to
get the value
ffmpeg | branch: master | Christophe Gisquet |
Sat Feb 7 18:49:37 2015 +| [b61b9e4919dcdcfc88c47692020be6bada2a2c9d] |
committer: Michael Niedermayer
x86: hevc_mc: remove lea in EPEL_LOAD
The second parameter to the macro is always an immediate address,
so no lea is needed.
Signed-off
ffmpeg | branch: master | Christophe Gisquet |
Sat Feb 7 18:49:36 2015 +| [4919b384214ed91c1d81b6a0541f84f0b08a76b2] |
committer: Michael Niedermayer
x86: hevc_mc: fewer gpr autoloads for _v filters
In that case, it's just to load my, but mx/r3src is not used.
Signed-off-by: Mi
ffmpeg | branch: master | Christophe Gisquet |
Sat Feb 7 13:19:58 2015 +0100| [626d6184ce74186cf588f2c8bef1a44bd0a595d0] |
committer: Michael Niedermayer
x86: lavc/hevc_mc: fix comments
The width parameter is now completely at the back, and actually
never used. This helps understanding the
ffmpeg | branch: master | Christophe Gisquet |
Tue Feb 3 14:06:39 2015 +0100| [97996eff4f2631fa5fdb6bc76a32b58d8b9f7611] |
committer: Michael Niedermayer
hevc/sao: do in-place band filtering when possible
The copies are only needed when data must be restored, so skip them
when it must not be
ffmpeg | branch: master | Christophe Gisquet |
Wed Feb 4 11:41:45 2015 +0100| [ed450d4acfc9fea0dae2df2b0a543ec0d602d31a] |
committer: Michael Niedermayer
x86: lavc: share more constant through defines
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.cgi/ffmpeg.git
ffmpeg | branch: master | Christophe Gisquet |
Sat Feb 7 09:20:09 2015 +| [46e2afa4dcb5ed6b4a6aa7aacbce2df74ed09c03] |
committer: Michael Niedermayer
wmalossless: reset lms_update
It may contain garbage at the end, and zeroing allows using DSP
with longer loops.
Signed-off-by: Michael
ffmpeg | branch: master | Christophe Gisquet |
Sat Feb 7 09:20:10 2015 +| [691b7f5e9e29698841c6bf1654605760020c0e8e] |
committer: Michael Niedermayer
lavc/lossless_audiodsp: revert various commits
Their intent was to make the DSP work with wmalossless pro.
The later was fixed to work
ffmpeg | branch: master | Christophe Gisquet |
Thu Feb 5 19:20:43 2015 +| [9dc45d1f42801036443e41a3cc424c7635936539] |
committer: Michael Niedermayer
x86: lavc: share more constants
Reviewed-by: "Ronald S. Bultje"
Signed-off-by: Michael Niedermayer
> http://git.videolan.o
ffmpeg | branch: master | Christophe Gisquet |
Thu Feb 5 19:20:40 2015 +| [5eedd36df1089aa0605ddb4c751b310b62f08ac4] |
committer: Michael Niedermayer
x86: hevc_mc: use epel_hv 16-wide function
The epel_hv functions were still relying on only epel_hv 8-wide
being the maximum width
1 - 100 of 242 matches
Mail list logo