]> git.sesse.net Git - ffmpeg/log
ffmpeg
9 years agolclenc: Mark codec as init-thread-safe and init-cleanup
Vittorio Giovara [Mon, 6 Apr 2015 23:47:18 +0000 (01:47 +0200)]
lclenc: Mark codec as init-thread-safe and init-cleanup

9 years agohqx: Mark codec as init-thread-safe and init-cleanup
Vittorio Giovara [Mon, 6 Apr 2015 22:40:37 +0000 (00:40 +0200)]
hqx: Mark codec as init-thread-safe and init-cleanup

9 years agohuffyuvenc: Mark codec as init-thread-safe and init-cleanup
Vittorio Giovara [Mon, 6 Apr 2015 23:47:18 +0000 (01:47 +0200)]
huffyuvenc: Mark codec as init-thread-safe and init-cleanup

9 years agoasvenc: Mark codec as init-thread-safe and init-cleanup
Vittorio Giovara [Mon, 6 Apr 2015 23:48:12 +0000 (01:48 +0200)]
asvenc: Mark codec as init-thread-safe and init-cleanup

9 years agohq_hqa: Fix table data for profile 17
Vittorio Giovara [Thu, 23 Apr 2015 10:29:53 +0000 (11:29 +0100)]
hq_hqa: Fix table data for profile 17

The table had a wrong table height and was missing the first line of bytes.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agoriff: Add GMP4 fourcc for mpeg4
Vittorio Giovara [Thu, 23 Apr 2015 12:25:20 +0000 (13:25 +0100)]
riff: Add GMP4 fourcc for mpeg4

9 years agoriff: Add GLV4 fourcc for mpeg4
Vittorio Giovara [Tue, 21 Apr 2015 12:10:00 +0000 (13:10 +0100)]
riff: Add GLV4 fourcc for mpeg4

9 years agointernal: Make dlog/tlog a no-op when disabled
Vittorio Giovara [Tue, 21 Apr 2015 12:24:18 +0000 (13:24 +0100)]
internal: Make dlog/tlog a no-op when disabled

Improves Coverity analysis, avoiding "double semicolon" CIDs.

9 years agortsp: Make sure we don't write too many transport entries into a fixed-size array
Martin Storsjö [Fri, 24 Apr 2015 09:38:09 +0000 (12:38 +0300)]
rtsp: Make sure we don't write too many transport entries into a fixed-size array

CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agortpenc_jpeg: Handle case of picture dimensions not dividing by 8
Andrey Utkin [Fri, 10 Apr 2015 21:54:10 +0000 (00:54 +0300)]
rtpenc_jpeg: Handle case of picture dimensions not dividing by 8

This fixes the calculation of the number of needed blocks to make
sure that ALL pixels are represented by the result.

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agortsp: Don't warn about unparsed time ranges
Martin Storsjö [Wed, 22 Apr 2015 20:15:58 +0000 (23:15 +0300)]
rtsp: Don't warn about unparsed time ranges

This removes the error logging added in 4e54432164.

This avoids warnings about "Invalid interval start specification 'now'"
for live rtsp streams.

We only try to parse some of the many valid values for time ranges
in RTSP - the other ones are fully valid but not interesting for the
use case in rtsp.c, so we shouldn't warn about them.

(Parsing the time ranges is needed to allow seeking, but e.g. setting
the current realtime clock for the start time doesn't make sense.
av_parse_time has got a different mode for parsing absolute times
as well, which can handle the special case "now", but that doesn't
make much sense for this particular use in rtsp.c.)

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agofate: Fix test name for pixfmts tests
Timothy Gu [Mon, 20 Apr 2015 03:37:43 +0000 (20:37 -0700)]
fate: Fix test name for pixfmts tests

The last pixel format gets leaked as `$test` further down the pipeline.

See for example https://fate.libav.org/x86_32-netbsd-clang-no-inline-asm/20150420020104
Note the odd test names like “yuvj444p.”

CC: libav-stable@libav.org
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agoprores: Set the bits_per_coded_sample for alpha pix_fmt
Luca Barbato [Wed, 22 Apr 2015 18:50:10 +0000 (20:50 +0200)]
prores: Set the bits_per_coded_sample for alpha pix_fmt

Improve the compatibility with other software.

9 years agohq_hqa: Validate get_vlc2 return value
Luca Barbato [Tue, 21 Apr 2015 12:42:21 +0000 (14:42 +0200)]
hq_hqa: Validate get_vlc2 return value

The `hq_ac_vlc.table` is incomplete, so unaccounted symbol return -1
as value.

9 years agoavcodec: Free the default avoptions on init failure
Luca Barbato [Tue, 21 Apr 2015 16:17:38 +0000 (18:17 +0200)]
avcodec: Free the default avoptions on init failure

9 years agoavcodec: Unref the dummy buffer on the fail path
Luca Barbato [Tue, 21 Apr 2015 12:26:36 +0000 (14:26 +0200)]
avcodec: Unref the dummy buffer on the fail path

9 years agobuffersrc: Use the correct deallocation function
Federico Tomassetti [Mon, 20 Apr 2015 19:12:05 +0000 (20:12 +0100)]
buffersrc: Use the correct deallocation function

This correction was suggested to me by Michael Niedermayer

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agodrawtext: Add an alpha option
Luca Barbato [Sat, 11 Apr 2015 17:04:25 +0000 (19:04 +0200)]
drawtext: Add an alpha option

And document both `draw` and `alpha`.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agolog: Print a full backtrace along with error messages under Valgrind
Vittorio Giovara [Mon, 20 Apr 2015 13:44:32 +0000 (14:44 +0100)]
log: Print a full backtrace along with error messages under Valgrind

Useful to understand where and in what execution state a certain message
is generated. It is enabled only when optimizations are disabled, since
function names are not printed otherwise.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agohq_hqa: Fix reading hqa quantization matrix index
Vittorio Giovara [Mon, 20 Apr 2015 11:06:01 +0000 (12:06 +0100)]
hq_hqa: Fix reading hqa quantization matrix index

Value is stored as a byte.

9 years agohqxdsp.h: #include "stddef.h"
Vittorio Giovara [Mon, 20 Apr 2015 18:11:49 +0000 (19:11 +0100)]
hqxdsp.h: #include "stddef.h"

It is required by ptrdiff_t, fixes make checkheaders.

9 years agolavf: Fix variables constness in ff_iso8601_to_unix_time()
Vittorio Giovara [Mon, 20 Apr 2015 11:18:38 +0000 (12:18 +0100)]
lavf: Fix variables constness in ff_iso8601_to_unix_time()

warning: assigning to
      'char *' from 'const char *' discards qualifiers
      [-Wincompatible-pointer-types-discards-qualifiers]

9 years agoh264_ps: Return meaningful error codes and address a memory leak
Vittorio Giovara [Mon, 20 Apr 2015 12:01:59 +0000 (13:01 +0100)]
h264_ps: Return meaningful error codes and address a memory leak

Bug-Id: CID 1026763

9 years agoffv1: Pass correct pointers to av_free()
Diego Biurrun [Mon, 20 Apr 2015 11:58:45 +0000 (13:58 +0200)]
ffv1: Pass correct pointers to av_free()

Bug-Id: CID 1295124

9 years agofate: Omit the results if the test passed
Timothy Gu [Fri, 10 Apr 2015 17:36:03 +0000 (10:36 -0700)]
fate: Omit the results if the test passed

The FATE server does not report this information anyway and omitting
it makes the successful run send much less data.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agormenc: Drop the temporary buffer for ac3 byteswap
Luca Barbato [Sat, 21 Mar 2015 14:38:37 +0000 (15:38 +0100)]
rmenc: Drop the temporary buffer for ac3 byteswap

Use direcly avio_w8().

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agoogg: Forward errors further
Luca Barbato [Sat, 21 Mar 2015 15:34:39 +0000 (16:34 +0100)]
ogg: Forward errors further

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agoflac: fix realloc loop with invalid flac files
Thomas Guillem [Thu, 16 Apr 2015 16:33:31 +0000 (18:33 +0200)]
flac: fix realloc loop with invalid flac files

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agodashenc: Add a publishTime field in dynamic manifests
Martin Storsjö [Fri, 17 Apr 2015 21:01:02 +0000 (00:01 +0300)]
dashenc: Add a publishTime field in dynamic manifests

This field is mandatory in dynamic manifests.

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agojack: Check memory allocation
Federico Tomassetti [Sun, 19 Apr 2015 11:28:58 +0000 (12:28 +0100)]
jack: Check memory allocation

Bug-Id: CID 1265785
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agobuffersrc: Fix resource leak on error
Federico Tomassetti [Mon, 23 Feb 2015 17:55:53 +0000 (17:55 +0000)]
buffersrc: Fix resource leak on error

Bug-Id: CID 1267902
CC: libav-stable@libav.org
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agoavplay: Free frame and graph when memory allocation fails
Federico Tomassetti [Sun, 19 Apr 2015 10:43:14 +0000 (11:43 +0100)]
avplay: Free frame and graph when memory allocation fails

Bug-Id: CID 1291564
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agog2meet: Support G2M5 codec variant
Eric Zimmerman [Fri, 17 Apr 2015 12:39:01 +0000 (13:39 +0100)]
g2meet: Support G2M5 codec variant

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agofiltfmts: Avoid null pointer dereference
Himangi Saraogi [Fri, 3 Apr 2015 14:12:11 +0000 (19:42 +0530)]
filtfmts: Avoid null pointer dereference

9 years agoavutil: remove pointless bmi1 define
James Almer [Sat, 18 Apr 2015 23:12:54 +0000 (20:12 -0300)]
avutil: remove pointless bmi1 define

Signed-off-by: James Almer <jamrial@gmail.com>
9 years agoaacdec: consistently use avctx for logging in decode_eld_specific_config
Andreas Cadhalpun [Thu, 16 Apr 2015 14:58:32 +0000 (16:58 +0200)]
aacdec: consistently use avctx for logging in decode_eld_specific_config

ac may be NULL and then accessing ac->avctx results in a segmentation fault.

Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>
9 years agompegvideo_enc: Simplify picture allocation
Vittorio Giovara [Wed, 15 Apr 2015 14:16:13 +0000 (15:16 +0100)]
mpegvideo_enc: Simplify picture allocation

9 years agomotion_est: Fix warning from zero_cmp() assignment
Vittorio Giovara [Tue, 14 Apr 2015 11:32:06 +0000 (13:32 +0200)]
motion_est: Fix warning from zero_cmp() assignment

warning: incompatible
      pointer types assigning to 'me_cmp_func' (aka 'int (*)(struct
      MpegEncContext *, uint8_t *, uint8_t *, ptrdiff_t, int)') from 'int
      (MpegEncContext *, uint8_t *, uint8_t *, int, int)'
      [-Wincompatible-pointer-types]

9 years agovc1_pred: Always initialize px and py in ff_vc1_pred_mv_intfr()
Vittorio Giovara [Sun, 19 Apr 2015 22:10:07 +0000 (23:10 +0100)]
vc1_pred: Always initialize px and py in ff_vc1_pred_mv_intfr()

Fix a rather lengthy initialization warning from clang.

9 years agomov: Double-check that alias path is not an absolute path
Vittorio Giovara [Tue, 7 Apr 2015 13:06:05 +0000 (15:06 +0200)]
mov: Double-check that alias path is not an absolute path

nlvl_to and nlvl_from can be set to 1 if both alias and target files
are in the same directory, so actually check the first character of the
string. We can do this because MacOS filepaths (alis type 2) are always
converted to UNIX filepaths (alis type 18).

9 years agomov: Read alternate absolute path in dref alis
Vittorio Giovara [Tue, 7 Apr 2015 13:06:04 +0000 (15:06 +0200)]
mov: Read alternate absolute path in dref alis

Absolute paths can be stored in alis type 2 and 18 according to my research:
the first is the canonical MacOS filepath, with path level separated by
colons, and the volume name within the filepath, while the second should be the
absolute filesystem path from the mount point.

9 years agoCanopus HQ/HQA decoder
Vittorio Giovara [Thu, 9 Apr 2015 16:12:30 +0000 (18:12 +0200)]
Canopus HQ/HQA decoder

Based on work by Kostya Shishkov <kostya.shishkov@gmail.com>.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agoImplement shared parsing of INFO tag in Canopus family
Vittorio Giovara [Thu, 9 Apr 2015 16:11:10 +0000 (18:11 +0200)]
Implement shared parsing of INFO tag in Canopus family

Add some bounds checking to CLLC; reduce HQX variable scoping,
add an error message.

9 years agofate: Prefix cllc tests with canopus
Vittorio Giovara [Thu, 9 Apr 2015 16:09:01 +0000 (18:09 +0200)]
fate: Prefix cllc tests with canopus

They belong to the same family of codecs.

9 years agohqx: Implement slice-threaded decoding
Vittorio Giovara [Thu, 9 Apr 2015 17:10:45 +0000 (19:10 +0200)]
hqx: Implement slice-threaded decoding

Inspired by a patch from Ferdinand Oeinck <ferdo@demon.nl>.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agohqx: Store shareable data in main decoder context
Vittorio Giovara [Thu, 9 Apr 2015 17:06:33 +0000 (19:06 +0200)]
hqx: Store shareable data in main decoder context

In preparation for multithreaded decoding.

9 years agohqx: Move DSP related code to a separate file
Vittorio Giovara [Thu, 9 Apr 2015 16:09:32 +0000 (18:09 +0200)]
hqx: Move DSP related code to a separate file

9 years agohqx: Merge invalid format check within switch block
Vittorio Giovara [Thu, 9 Apr 2015 16:09:31 +0000 (18:09 +0200)]
hqx: Merge invalid format check within switch block

9 years agolavc: Replace av_dlog and tprintf with internal macros
Vittorio Giovara [Mon, 16 Mar 2015 08:57:36 +0000 (08:57 +0000)]
lavc: Replace av_dlog and tprintf with internal macros

9 years agoReplace av_dlog with normal av_log at trace level
Vittorio Giovara [Mon, 16 Mar 2015 08:57:35 +0000 (08:57 +0000)]
Replace av_dlog with normal av_log at trace level

This applies to every library where performance is not critical.

9 years agolog: Introduce a more verbose debug level
Vittorio Giovara [Mon, 16 Mar 2015 08:57:34 +0000 (08:57 +0000)]
log: Introduce a more verbose debug level

And deprecate av_dlog macro.

9 years agomatroskadec: fix crash when parsing invalid mkv
Thomas Guillem [Fri, 10 Apr 2015 17:04:51 +0000 (19:04 +0200)]
matroskadec: fix crash when parsing invalid mkv

CC: libav-stable@libav.org
Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years agoconfigure: Simplify avisynth check
Diego Biurrun [Tue, 17 Mar 2015 12:31:41 +0000 (13:31 +0100)]
configure: Simplify avisynth check

9 years agoavisynth: Simplify shared library name construction
Diego Biurrun [Tue, 17 Mar 2015 12:32:06 +0000 (13:32 +0100)]
avisynth: Simplify shared library name construction

9 years agox265: Map color parameters
Luca Barbato [Mon, 13 Apr 2015 08:45:51 +0000 (10:45 +0200)]
x265: Map color parameters

9 years agox264: Map color parameters
Luca Barbato [Mon, 13 Apr 2015 08:45:51 +0000 (10:45 +0200)]
x264: Map color parameters

9 years agoavisynth: Bump minimum required version to interface version 6
Stephen Hutchinson [Thu, 2 Apr 2015 19:39:26 +0000 (15:39 -0400)]
avisynth: Bump minimum required version to interface version 6

The AVSC_API changes in the new headers mean that the 2.6 alphas
are just as incompatible as 2.5 is.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years agoavisynth: Fix detection of AviSynth 2.5
Stephen Hutchinson [Thu, 2 Apr 2015 19:39:25 +0000 (15:39 -0400)]
avisynth: Fix detection of AviSynth 2.5

In order to safely exit when the user tries to use AviSynth 2.5, the
continue_on_fail value for 2.6's functions need to be set to 1.
Otherwise, the library loader fails before the 'upgrade to 2.6'
log message appears.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years agompeg4videodec: Remove useless messages
wm4 [Mon, 13 Apr 2015 17:35:54 +0000 (19:35 +0200)]
mpeg4videodec: Remove useless messages

They seem to sometimes trigger with old AVI files. They are just
confusing, do not help anyone, and use sloppy language.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years agommal: Reference MMAL VC lib explicitly
wm4 [Mon, 13 Apr 2015 17:18:00 +0000 (19:18 +0200)]
mmal: Reference MMAL VC lib explicitly

This is optional, but ensures that linking with -Wl,--as-needed does not
drop the library containing the MMAL VC driver. The driver normally
"registers" itself in the library constructor, but since no symbols are
explicitly referenced, the linker could remove it with --as-needed
enabled.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years agommal: Move system headers before local headers
wm4 [Mon, 13 Apr 2015 17:17:59 +0000 (19:17 +0200)]
mmal: Move system headers before local headers

Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years agoffv1: Check memory allocations
Vittorio Giovara [Thu, 9 Apr 2015 21:37:59 +0000 (23:37 +0200)]
ffv1: Check memory allocations

Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years agortsp: Fix unchecked return value
Himangi Saraogi [Thu, 9 Apr 2015 18:34:01 +0000 (00:04 +0530)]
rtsp: Fix unchecked return value

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agortsp: Check a malloc return value
Himangi Saraogi [Sat, 4 Apr 2015 10:51:40 +0000 (16:21 +0530)]
rtsp: Check a malloc return value

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agolibx265: print supported presets and tunes on error
James Almer [Fri, 10 Apr 2015 19:47:03 +0000 (16:47 -0300)]
libx265: print supported presets and tunes on error

Based on code from libavcodec/libx264.c

Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
9 years agolibdc1394: Unbreak build after c201069fa
Michael Kostylev [Sat, 11 Apr 2015 08:37:49 +0000 (10:37 +0200)]
libdc1394: Unbreak build after c201069fa

9 years agovfwcap: Unbreak building after c201069fa
Martin Storsjö [Fri, 10 Apr 2015 09:22:33 +0000 (12:22 +0300)]
vfwcap: Unbreak building after c201069fa

These headers can't be included in any arbitrary order.

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agoavdevice: Apply a more consistent file naming scheme
Diego Biurrun [Fri, 27 Mar 2015 11:40:23 +0000 (12:40 +0100)]
avdevice: Apply a more consistent file naming scheme

9 years agoavdevice: Add missing header for NULL_IF_CONFIG_SMALL
Diego Biurrun [Wed, 8 Apr 2015 15:27:42 +0000 (17:27 +0200)]
avdevice: Add missing header for NULL_IF_CONFIG_SMALL

Also reshuffle headers into canonical order where appropriate.

9 years agolibavformat: Handle error return from ff_listen_bind
Anders Nystrom [Thu, 9 Apr 2015 00:38:41 +0000 (02:38 +0200)]
libavformat: Handle error return from ff_listen_bind

Handle error return from ff_listen_bind without leaking file descriptors.

Signed-off-by: Anders Nystrom <anders.nystrom@southpole.se>
Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agohqx: Fix clipping error in idct put function
Ferdinand Oeinck [Tue, 7 Apr 2015 10:42:17 +0000 (12:42 +0200)]
hqx: Fix clipping error in idct put function

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agoriff: Add 'M705' FourCC for mpeg2video
Vittorio Giovara [Tue, 7 Apr 2015 10:30:16 +0000 (12:30 +0200)]
riff: Add 'M705' FourCC for mpeg2video

9 years agodnxhd: Log the selected profile id
Vittorio Giovara [Thu, 2 Apr 2015 20:56:39 +0000 (22:56 +0200)]
dnxhd: Log the selected profile id

9 years agomov: Rely on box type rather than file type for colr atom
Vittorio Giovara [Mon, 23 Mar 2015 19:43:27 +0000 (19:43 +0000)]
mov: Rely on box type rather than file type for colr atom

Although it's not allowed to use only allows 'nclc' in ISOM files, there
are samples that do not always respect this rule. This change prevents
atom overread and a spurious color range initialization.

Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
9 years agov210: Fix typo in header guards
Vittorio Giovara [Mon, 23 Mar 2015 20:13:43 +0000 (20:13 +0000)]
v210: Fix typo in header guards

9 years agog2meet: K&R formatting cosmetics
Diego Biurrun [Fri, 10 Oct 2014 08:52:03 +0000 (10:52 +0200)]
g2meet: K&R formatting cosmetics

9 years agomatroskadec: export cover art correctly
wm4 [Fri, 3 Apr 2015 14:11:53 +0000 (16:11 +0200)]
matroskadec: export cover art correctly

Generally, libavformat exports cover art pictures as video streams with
1 packet and AV_DISPOSITION_ATTACHED_PIC set. Only matroskadec exported
it as attachment with codec_id set to AV_CODEC_ID_MJPEG.

Obviously, this should be consistent, so change the Matroska demuxer to
export a AV_DISPOSITION_ATTACHED_PIC pseudo video stream.

Matroska muxing is probably incorrect too. I know that it can create
broken files with an audio track and just 1 video frame when e.g.
remuxing mp3 with APIC to mkv. But for now this commit does not change
anything about muxing, and also continues to write attachments with
AV_CODEC_ID_MJPEG should the muxer application have special knowledge
that the Matroska is broken in this way.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years agoparseutils: Make av_small_strptime public
Luca Barbato [Sun, 5 Apr 2015 23:45:37 +0000 (01:45 +0200)]
parseutils: Make av_small_strptime public

And use it in libavformat.

Based on a similar patch by Stefano Sabatini <stefasab@gmail.com>.

9 years agoparseutils: Extend small_strptime to be used in avformat
Luca Barbato [Sun, 5 Apr 2015 23:25:52 +0000 (01:25 +0200)]
parseutils: Extend small_strptime to be used in avformat

The strptime implementation is supposed to support whitespace and %T.

9 years agosvq3: initialize some required H264Context fields.
Anton Khirnov [Sun, 5 Apr 2015 19:02:56 +0000 (21:02 +0200)]
svq3: initialize some required H264Context fields.

They are no longer initialized in ff_h264_decode_init() since 43fd3dd,
so svq3 needs to initialize the manually.

Fixes svq3 decoding, broken since 43fd3dd.

9 years agolibavcodec: Clarify the documentation of the internal codec capability flags
Martin Storsjö [Thu, 2 Apr 2015 12:17:01 +0000 (15:17 +0300)]
libavcodec: Clarify the documentation of the internal codec capability flags

The previous documentation was very vague and almost misleading.

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agoavcodec/libx265: export chosen picture types
Michael Niedermayer [Mon, 30 Mar 2015 22:37:46 +0000 (00:37 +0200)]
avcodec/libx265: export chosen picture types

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
9 years agoh264: reset the private data in init_thread_copy()
Anton Khirnov [Sun, 29 Mar 2015 10:26:45 +0000 (12:26 +0200)]
h264: reset the private data in init_thread_copy()

The generic code copies the main context's private data to all the
others. However that is quite dangerous, as it might end up copying some
pointers that are or will become invalid.

Since everything we actually need will be copied later in
update_thread_context(), it's safest to zero the private data in
init_thread_copy(), so it works the same way as init for the main
context.

9 years agoh264: remove a commented out nonsense line
Anton Khirnov [Sun, 29 Mar 2015 10:19:24 +0000 (12:19 +0200)]
h264: remove a commented out nonsense line

9 years agoh264: drop redundant initialization of the scaling matrices
Anton Khirnov [Sun, 29 Mar 2015 10:18:36 +0000 (12:18 +0200)]
h264: drop redundant initialization of the scaling matrices

It will always be initialized when actually parsing the PPS.

9 years agoh264: factor out common code from init() and init_thread_copy()
Anton Khirnov [Mon, 23 Mar 2015 04:46:49 +0000 (05:46 +0100)]
h264: factor out common code from init() and init_thread_copy()

9 years agoh264: drop redundant initialization in init()
Anton Khirnov [Sat, 21 Mar 2015 15:12:44 +0000 (16:12 +0100)]
h264: drop redundant initialization in init()

There is no real advantage to initializing any of those in init,
assuming yuv420, before the real stream parameters are known.

9 years agoh264: initialize cur_chroma_format_idc
Anton Khirnov [Sat, 21 Mar 2015 15:08:18 +0000 (16:08 +0100)]
h264: initialize cur_chroma_format_idc

This makes sure the various DSP contexts get properly initialized in
ff_h264_set_parameter_from_sps() whatever the value of
raw_bits_per_sample.

9 years agoh264: use the correct SPS during PPS parsing
Anton Khirnov [Fri, 6 Mar 2015 15:03:51 +0000 (16:03 +0100)]
h264: use the correct SPS during PPS parsing

There is in general no reason for the currently active SPS to be the one
referenced by the PPS being parsed.

9 years agoavconv: Avoid theoretical NULL dereferences
Himangi Saraogi [Wed, 1 Apr 2015 19:33:04 +0000 (01:03 +0530)]
avconv: Avoid theoretical NULL dereferences

Bug-Id: CID 1292519

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agoh264: disable ER by default
Anton Khirnov [Tue, 31 Mar 2015 08:35:15 +0000 (10:35 +0200)]
h264: disable ER by default

The way it is currently designed is fundamentally unsafe and cannot be
reasonably fixed without completely rewriting it.

9 years agoavconv: do not overwrite the stream codec context for streamcopy
Anton Khirnov [Thu, 2 Apr 2015 04:09:05 +0000 (06:09 +0200)]
avconv: do not overwrite the stream codec context for streamcopy

Since we are not doing encoding, there is no point in ever touching the
separate encoding context. Always use the stream codec context.

Fixes writing attachments.

CC:libav-devel@libav.org

9 years agoFATE: add a test for parameter changes in HEVC
Anton Khirnov [Sun, 29 Mar 2015 14:07:36 +0000 (16:07 +0200)]
FATE: add a test for parameter changes in HEVC

9 years agohevc: make sure no dangling pointers remain around on VPS/SPS change
Anton Khirnov [Sun, 29 Mar 2015 13:24:46 +0000 (15:24 +0200)]
hevc: make sure no dangling pointers remain around on VPS/SPS change

9 years agohevc: do not change the VPS if its contents are the same
Anton Khirnov [Sun, 29 Mar 2015 13:20:30 +0000 (15:20 +0200)]
hevc: do not change the VPS if its contents are the same

This is the same as is done for SPS.

9 years agohevc: export stream parameters from extradata
Anton Khirnov [Sun, 29 Mar 2015 12:15:41 +0000 (14:15 +0200)]
hevc: export stream parameters from extradata

This allows the callers to have a hint of the probable stream parameters
without actually decoding anything.

9 years agohevc: split out setting AVCodecContext parameters
Anton Khirnov [Sun, 29 Mar 2015 12:09:51 +0000 (14:09 +0200)]
hevc: split out setting AVCodecContext parameters

Additionally always set the software pixel format, so it's available
even if ff_get_format() is not called later. This will be useful for
exporting stream parameters from init().

9 years agompegtsenc: Take max_delay into account when buffering multiple audio packets into...
Martin Storsjö [Thu, 2 Apr 2015 08:09:31 +0000 (11:09 +0300)]
mpegtsenc: Take max_delay into account when buffering multiple audio packets into one PES packet

Make sure we don't buffer up more than max_delay worth of data
before writing a PES packet, even if pes_payload_size is set to
a larger value.

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agortsp: Add a buffer_size option
Luca Barbato [Sun, 22 Mar 2015 20:16:55 +0000 (21:16 +0100)]
rtsp: Add a buffer_size option

And forward it to rtp and udp.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>
9 years agortp: Add an option to set the send/receive buffer size
Luca Barbato [Sun, 29 Mar 2015 04:26:53 +0000 (06:26 +0200)]
rtp: Add an option to set the send/receive buffer size

It gets forwarded down to UDP.