The initial testing of the VFW binary codec was flawed,
likely due to an AviSynth bug.
Re-testing using VirtualDub and various professional editing
applications has revealed it should have been flipped.
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
* qatar/master:
rtpdec: Remove an outdated todo comment
rtpdec: Rename a static variable to normal naming conventions
sh4: dsputil: remove duplicate of ff_gmc_c()
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '30b39164256999efc8d77edc85e2e0b963c24834':
ac3dec: make downmix() take array of pointers to channel data
Conflicts:
libavcodec/ac3dsp.c
libavcodec/ac3dsp.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'b8f3ab8e6a7ce3627764da53b809628c828d4047':
ac3dec: output planar float only
svq3: make slice type value unsigned to match svq3_get_ue_golomb return type
configure: Have protocols select network code instead of depending on it
Conflicts:
libavcodec/svq3.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '91ac403b1316d59b4f43c4ea0f237e24cec2819a':
lavf: fix arithmetic overflows in avformat_seek_file()
mjpeg: initialize input padding after unescaped buffer to zero
Conflicts:
libavformat/utils.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Fixes rare race condition leading to null pointer dereferences.
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Prevent an invalid write into coeffs[scantable[-1]] if zeros_left
itself was an invalid VLC code (and thus -1).
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
* qatar/master:
golomb: use unsigned arithmetics in svq3_get_ue_golomb()
x86: float_dsp: fix loading of the len parameter on x86-32
takdec: fix initialisation of LOCAL_ALIGNED array
takdec: fix initialisation of LOCAL_ALIGNED array
Conflicts:
libavcodec/rv30.c
libavcodec/svq3.c
libavcodec/takdec.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '57231e4d5b467833fb289439cd35a92513bb55c1':
tak: demuxer, parser, and decoder
Not merged as requested by Author and Maintainer of tak in FFmpeg.
I just merged a few typo fixes and minor cosmetic improvments.
Conflicts:
Changelog
libavcodec/Makefile
libavcodec/allcodecs.c
libavcodec/avcodec.h
libavcodec/tak.c
libavcodec/tak.h
libavcodec/tak_parser.c
libavcodec/takdec.c
libavcodec/version.h
libavformat/takdec.c
libavformat/version.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '096abfa15052977eed93f0b5e01afd2d47c53c1f':
parser: fix large overreads
bitstream: add get_bits64() to support reading more than 32 bits at once
arm: detect cpu features at runtime on Linux
Conflicts:
libavcodec/parser.c
libavformat/mpegts.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '9ebd45c2d58ad9241ad09718679f0cf7fb57da52':
configure: do not bypass cpuflags section if --cpu not given
dct-test: arm: indicate required cpu features for optimised funcs
snow: fix build after 594d4d5df3
arm: fix use of uninitialised value in ff_fft_fixed_init_arm()
avpicture: Don't assume a valid pix fmt in avpicture_get_size
Conflicts:
libavcodec/avpicture.c
libavcodec/snow.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This prevents undefined behaviour of signed left shift if the coded
value is larger than 2^31. Large values are most likely invalid and
caused errors or by feeding random.
Validate every use of svq3_get_ue_golomb() and changed the place there
the return value was compared with negative numbers. dirac.c was clean,
fixed rv30 and svq3.
Also remove a duplicate function in the MPEG-TS demuxer.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
This is consistent with usual ARM nomenclature as well as with the
VFPV3 and NEON symbols which both lack the ARM prefix.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Merge get_scale/get_shift into set_sample_rate_params().
Rename tak_set_bps to set_bps_params and remove 2nd argument.
Signed-off-by: Paul B Mahol <onemda@gmail.com>
* commit 'd7d6efe42b0d2057e67999b96b9a391f533d2333':
h264: check sps.log2_max_frame_num for validity
mov: validate number of DataReferenceBox entries against box size
mov: compute avg_frame_rate only if duration is known
flac: change minimum and default of lpc_passes option to 1
Conflicts:
libavcodec/h264_ps.c
libavformat/mov.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
When initialising an FFTContext for a plain FFT, mdct_bits is not set
and can contain a garbage value. Since nbits is always valid and for
MDCT operation is mdct_bits - 2 checking this instead avoids using an
uninitialised value while having the same effect.
Signed-off-by: Mans Rullgard <mans@mansr.com>
When called from the v4l2 input device, pix_fmt can be
AV_PIX_FMT_NONE (for jpeg formats). Before 50ba57e0, this wasn't
an issue for avpicture_get_size, but after that commit, this
lead to crashes.
Signed-off-by: Martin Storsjö <martin@martin.st>
Avoid use of uninitialized and uncomputed linear least square models
during ff_lpc_calc_coefs() for FF_LPC_TYPE_CHOLESKY. Fixes running
make fate-flac-16-lpc-cholesk with valgrind --undef-value-errors=yes.
Start and end index are multiple of 2, therefore guaranteeing aligned access.
Also, this allows to generate 4 floats per loop, keeping the alignment all
along.
Timing:
- 32 bits: 326c -> 172c
- 64 bits: 323c -> 156c
Signed-off-by: Diego Biurrun <diego@biurrun.de>
Instead of having own code for calculating decoded buffer size
just use already provided functions from libavutil.
Signed-off-by: Paul B Mahol <onemda@gmail.com>
* qatar/master:
h264: slice-mt: check master context for valid current_picture_ptr
h264: slice-mt: get last_pic_dropable from master context
alacenc: add support for multi-channel encoding
Conflicts:
Changelog
libavcodec/alac.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This reverts commit 7feef7dbca.
This commit causes assertion failures due to use of parser_init from multiple
threads, for example indirectly by ffmpeg.c and more directly from the
packet read functions.
I dont know how to fix this quickly, and fixing ffmpeg.c leaves
the possibility of other applications being affected.
Crashing the applications until this is resolved is clearly no good
thus this revert, so we have time to think about the problem.
Crashes can be reproduced by using multiple input files in ffmpeg.
Will be used in future patches, together with the
variable that allows checking whether the lock is held.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
* qatar/master:
fate: workaround for slighly broken 'test' shell builtin
mimic: initialize padding of swap_buf through av_fast_padded_malloc
eamad: initialize padding of bitstream_buf through av_fast_padded_malloc()
raw demuxer: initialize end of partial packets
Conflicts:
tests/fate-run.sh
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '387bef95d28019c13c6805cfa4079e59948284e5':
lavc: factorise setting buffer type in avcodec_default_get_buffer().
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'e57c4706e969afa1f2384481b955ccd9494cddb5':
lavc: don't reuse audio buffers
This commit causes a 0.5% speedloss for mp3 and 2% for raw pcm, that is
"time ffmpeg" style tested thus includes disk IO, demux and parsing
I would not have merged it if it wasnt required for the "new" buffer API
but sadly it is.
Once the new API is in ill reimplement proper buffer reuse, which will
fix this speedloss. In case we choose not to merge the "new" buffer
API then this commit here should be reverted.
Conflicts:
libavcodec/internal.h
libavcodec/utils.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'ff953fecffd3b9a616a046723fb9d4690be032a6':
lavc: set frame properties in ff_get_buffer().
Conflicts:
libavcodec/utils.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This handles the last frame being unavailable like all the other
code in vc1dec.
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
the default is still 4 slices for any resolution, this just allows the user
to force 1 slice.
This in my quick test improves compression by 1% for a 320x240 sample
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
* commit '0b300daad2f5cb59a7c06dde5ac701685e6edf16':
h264: error out on unset current_picture_ptr for h->current_slice > 0
avprobe: report per stream bit rate if set by the decoder
aac: avoid a memcpy in sbr_qmf_analysis
Conflicts:
avprobe.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Swapping buffer indices allows saving one memcpy that accounts for 1% of the
runtime, according to oprofile.
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
Fixes null pointer dereference and potential out of array accesses.
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Prevents inconsistency and out of array accesses.
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This change requires the user to specify min and max value, and makes
possible to prevent the user to set AV_{SAMPLE,PIX}_FMT_NONE if
forbidden.
Add required ifdeffery in case of mixed libraries, when libavutil is
updated but not the other libraries.
This is a followup of 08d0969c14.
Add capability of reading multiple frames instead of only first.
Implement support for different gif frame 'disposal methods'.
Add option that allows to change background color resulting from
conversion of gif with transparency to any other format which
not support it.
Also bump lavc minor version.
Signed-off-by: Vitaliy E Sugrobov <vsugrob@hotmail.com>
Replace literals with named constants in several pieces of code
like 'return -1' and 'case 0xab'.
Change the way decoder handles absence of image data in a file:
notify gif_decode_frame() caller with got_picture set to zero
instead of returning -1.
Signed-off-by: Vitaliy E Sugrobov <vsugrob@hotmail.com>
* commit '9d46eaec7a90bd8f5cd9e45398c6d17804182320':
build: The FLAC encoder also depends on the flacdsp code
img2: K&R formatting cosmetics
h264: check context state before decoding slice data partitions
flashsv: make sure data for zlib priming is available
Conflicts:
libavcodec/Makefile
libavformat/img2.c
libavformat/img2dec.c
libavformat/img2enc.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This also fixes a long standing comment in the code.
Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Use the i64 field rather than the string value. Using a string to set a
default sample/pixel format is weird, also the new interface is more
consistent with the rest of the API.
This is technically an API break, but hopefully there are no applications
using this feature outside of FFmpeg. In order to save backward
compatibility with mixed libraries in case libavutil is updated but not
the other libraries, some ifdeffery hacks are added.
Note that the version check is only performed when class->version != 0,
since if it is not defined then we assume that no version was defined and
the class is not affected by the change.
We will luckily get rid of the hack at the next major bump.
It would be called while registering the codec, which means
it needlessly wastes memory when it is never used.
Instead do the work when first opening the codec.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
* commit 'af7d13ee4a4bf8d708f9b0598abb8f6e22b76de1':
asink_nullsink: plug a memory leak.
x86: h264_idct: port to cpuflags
x86: cpu: Drop unused HAVE_RWEFLAGS condition
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'f5fa03660db16f9d78abc5a626438b4d0b54f563':
vble: Do not abort decoding when version is not 1
lavr: do not pass consumed samples as a parameter to ff_audio_resample()
lavr: correct the documentation for the ff_audio_resample() return value
lavr: do not pass sample count as a parameter to ff_audio_convert()
x86: h264_weight: port to cpuflags
configure: Enable avconv filter dependencies automatically
Conflicts:
configure
libavcodec/x86/h264_weight.asm
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Some combinations of OS, VirtualDub, and VBLE can accidentally
set the version to a value other than 1. Since no other version
of VBLE was ever released, simply warn about it.
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
* commit '6e5cdf26281945ddea3aaf5eca4d127791f23ca8':
h264: check ref_count validity for num_ref_idx_active_override_flag
h264: add missing new line to log message
dcadec: skip QMF on unused channels
wavenc: write fact chunk sample count at the correct file position
riff: do not add empty metadata tags in INFO chunk
Conflicts:
libavcodec/dcadec.c
libavcodec/h264.c
libavformat/riff.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '284ea790d89441fa1e6b2d72d3c1ed6d61972f0b':
dsputil: move vector_fmul_scalar() to AVFloatDSPContext in libavutil
aacenc: use the correct output buffer
aacdec: fix signed overflows in lcg_random()
base64: fix signed overflow in shift
Conflicts:
libavcodec/dsputil.c
libavutil/base64.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This does not seem to have an effect currently. Fate-h264 passes with
THREADS=1..16 and both threading types as before. It fixes however a
segfault during error resilience with my adaptive-frame-mt patchset.
A picture in use during error resilience gets realloced in another
thread in the fuzzed sample sample_varPAR.avi_s226019.
When the extra rear channel is present but unused, the
s->channel_order_tab[] value for that channel is -1. The QMF can be
skipped for the extra channel, and doing so avoids an out-of-array read
on s->samples_chanptr[].
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
This fixes segfault caused by 3d3cf6745e
when SingleChannelElement.ret was renamed to SingleChannelElement.ret_buf.
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
* commit '5ae72f54532960cb9eae82a1c9e8d505106c022b':
flashsv: check for keyframe before using differential coding
h264: enable low delay only if no delayed frames were seen
x86: fix build without inline asm
Conflicts:
libavcodec/h264.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '3d3cf6745e2a5dc9c377244454c3186d75b177fa':
aacdec: use float planar sample format for output
Conflicts:
libavcodec/aacdec.c
libavcodec/aacsbr.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '8e134e5104e99a69cd4cea10540a7ce9c3682a2c':
lavc: clarify get_buffer() documentation
mpegaudiodec: use planar sample format for output unless packed is requested
x86: h264 qpel: use the correct number of utilized xmm regs in cglobal
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Dropping frames is undesirable but that is the only way by which the
decoder could return to low delay mode. Instead emit a warning and
continue with delayed frames.
Fixes a crash in fuzzed sample nasa-8s2.ts_s20033 caused by a larger
than expected has_b_frames value. Low delay keeps getting re-enabled
from a presumely broken SPS.
CC: libav-stable@libav.org
The qpel functions referenced here are not related to h264 and should
thus never have been under CONFIG_H264QPEL.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Signed-off-by: Diego Biurrun <diego@biurrun.de>
Make the functions av_samples_fill_arrays, av_samples_alloc, and
avcodec_fill_audio_frame return a buffer size rather than 0 in case of
success. This will be enabled at the next libavutil major bump, in order
to preserve backward compatibility.
Returning the size allows to simplify the code, avoiding a few function
calls.
The only CPUs that have 3dnow and don't have mmxext are 12 years old.
Moreover, AMD has dropped 3dnow extensions from newer CPUs.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
The palette is organised into 'hold' and 'modify' pairs, hence why
the first palette entry is pal[1].
Fixes ticket #1701.
Signed-off-by: Peter Ross <pross@xvid.org>
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>