Template h264dsp_internal.h with respect to pixel size, etc.
Preparatory patch for high bit depth h264 decoding support.
Signed-off-by: Michael Niedermayer <email@example.com>
Move some functions in dsputil.c into a new file dsputil_internal.h.
The functions moved are used when decoding h264.Preparatory patch for high bit depth h264 decoding support.
Move the functions in h264idct into a new file h264idct_internal.h.
Move the functions in h264pred.c into a new file h264pred_internal.h.
Move the functions in h264dsp.c into a new file h264dsp_internal.h
Add pixel formats for 9- and 10-bit yuv420p.
Also add support for these formats in libswscale.
Needed for high bit depth h264 decoding.
Choose h264 chroma dc dequant function dynamically.
H264: Fix slice threadsSigned-off-by: Michael Niedermayer <firstname.lastname@example.org>
Fix input buffer size check in adpcm_ea decoder.
Unfortunately the output buffer size check assumes that theinput buffer is never over-consumed, thus this actuallyalso allowed to write outside the output buffer if "lucky".
Add "Did you provide a clear git commit log message?" to the patch checklistSigned-off-by: Michael Niedermayer <email@example.com>
x11: fix memleak
replace movw instruction in ac3dsp_armv6.S
AS libavcodec/arm/ac3dsp_armv6.offmpeg-src/libavcodec/arm/ac3dsp_armv6.S: Assembler messages:ffmpeg-src/libavcodec/arm/ac3dsp_armv6.S:40: Error: selected processordoes not support `movw r8,#0x1fe0'make1: *** [libavcodec/arm/ac3dsp_armv6.o] Error 1...
Merge remote branch 'qatar/master'
daud: Don't write packets that are too large to have their size expressed in the bytestream.
Abort Ogg header parsing when we encounter a data packet.
Fixes ticket #15.
pthread: validate_thread_parameters() ignored slice-threading being intentionally off
The thread_type API allows you to request only FF_THREAD_FRAME (instead ofFRAME | SLICE), but it was being ignored.We don't implement both of them at the same time, so there isn't an effect...
Use consistent condition for whether to run slice-threading execute function.
Signed-off-by: Ronald S. Bultje <firstname.lastname@example.org>
avformat.h: fix a misplaced #endif
vaapi: cope with VA-API 0.32, i.e. fix VC-1 decoding on Sandy Bridge.
(cherry picked from commit a18e7b4fb77d3799cf21110b5e6dda4691a295b3)
Signed-off-by: Reinhard Tartler <email@example.com>
vaapi: filter out DELAYED_PIC_REF flag to determine field.
This fixes ticket #23.
(cherry picked from commit 621f4c98df4ee9fd604a614f31e09eef9dd7d3ca)
CrystalHD: Improve detection of h.264 content.
As previously discussed, the CrystalHD hardware returns exceptionallyuseless information about interlaced h.264 content - to the extentthat it's not possible to distinguish most MBAFF and PAFF content until...
CrystalHD: Carry picture type from input to output picture.
Now that we know the type of the input picture, we have to bringthat information to the output picture to help identify its type.We do this by adding a field to the opaque_list node.
Signed-off-by: Philip Langdale <firstname.lastname@example.org>
CrystalHD: Bring in h.264 parser to establish picture type.
As the hardware is unreliable, we will have to use the h.264 parserto identify whether an input picture is a field or a frame. Thischange loads the parser and extracts the picture type.
CrystalHD: Change opaque_list_pop to return the node.
In preparation for adding additional fields to the node, returnthe node instead of the pts value. This requires the caller tofree the node.
CrystalHD: Fix whitespace after previous change.
'git diff -w' confirmed to return nothing.
CrystalHD: Handle different h.264 MBAFF packing.
I found another MBAFF sample where the input:output pattern isthe same as mpeg2 and vc1 (fieldpair input, individual field output).While I'm not sure how you can output individual fields from MBAFF,if I apply the mpeg2/vc1 handling to this file, it plays correctly....
flv: Only parse keyframe index when the underlaying protocol allows seeking as we need to seek to read the indexSigned-off-by: Michael Niedermayer <email@example.com>
Fix support for flvtool2 "keyframes based" generated index in FLV format decoder
Current keyframes data parser unconditionally rewind metadata to the endat the end of function. As result ALL metadata located after keyframesindex not parsed,and as metadata object can have ANY placement inside metadata it can...
Replace outdated info on the FAQ
FAQ could use some work. This updates a few little details.
Fix ticket21, temporary audio array was too small.Signed-off-by: Michael Niedermayer <firstname.lastname@example.org>
lavf: bump minor and add an APIChanges entry for avformat cleanup
lavf: get rid of ffm-specific stuff in avformat.h
avio: deprecate av_protocol_next().
avio: add a function for iterating though protocol names.
lavf: rename a parameter of av_sdp_create from buff->buf
This is more consistent with the rest of the API.
lavf: rename avf_sdp_create to av_sdp_create.
The new name is more consistent with the rest of the API.
lavf: make av_guess_image2_codec internal
It doesn't look very useful as a public function.
avio: make URLProtocol internal.
avio: make URLContext internal.
lavf: mark av_pkt_dump(_log) for remove on $next+1 bump.
Cleanup 'Fix spurious warning when stream has SPS and PPS units.'Signed-off-by: Michael Niedermayer <email@example.com>
lavf: use designated initializers for all protocols
This is more readable and makes it easier to reorder URLProtocolmembers.
applehttp: don't use deprecated url_ functions.
avio: move two ff_udp_* functions from avio_internal to url.h
asfdec: remove a forgotten declaration of nonexistent function
avio: deprecate the typedef for URLInterruptCB
There's no particular reason to pollute the namespace with a typedef forit.
Merge remote-tracking branch 'ffmpeg-mt/master'
fix another forgotten mplayerhq.hu linkSigned-off-by: Michael Niedermayer <firstname.lastname@example.org>
Fix link to ffmpeg-develSigned-off-by: Michael Niedermayer <email@example.com>
Fix deblocking with thread_slicesSigned-off-by: Michael Niedermayer <firstname.lastname@example.org>
ffplay: Set EMU_EDGE if dr1, this is an alternative fix for ticket40.Idea borrowed from mplayer.Signed-off-by: Michael Niedermayer <email@example.com>
proto: include os_support.h in network.h
Fix compilation on systems without poll()
matroskaenc: don't write an empty Cues element.
lavc: add a FF_API_REQUEST_CHANNELS deprecation macro
And hide request_channels AVOption under it.
avio: move extern url_interrupt_cb declaration from avio.h to url.h
avio: make av_register_protocol2 internal.
avio: avio_ prefix for url_set_interrupt_cb.
avio: AVIO_ prefixes for URL_ open flags.
riff: add ATRAC3+ guid
ff_get_wav_header: perform full lookup on WAVEFORMATEXTENSIBLE subformat guid
move PRI_GUID, ARG_GUID, AVCodecGuid, ff_codec_guid_get_id, MEDIASUBTYPE_BASE_GUID and audio_guids to riff.c/h (add FF_ prefix to PRI_GUID, ARG_GUID, and MEDIASUBTYPE_BASE_GUID; rename audio_guids to ff_codec_wav_guids)
move ff_asf_guid, ff_guidcmp and ff_get_guid into riff.h
pad: fix example explanation
"row" and "column" term positions in the explanation were inverted.
Signed-off-by: Stefano Sabatini <firstname.lastname@example.org>
proto: introduce listen option in tcp
This way is possible to have simple micro-server act like
ffmpeg -i file.nut -vcodec copy -acodec copy -f nut tcp://foo:1234?listen
doc: clarify configure features
This way people will not get surprises regarding the current configurebehaviour and contributors will have a guideline on what is expected.
proto: factor ff_network_wait_fd and use it on udp
Support the URL_FLAG_NONBLOCK semantic and uniform the protocol.The quick retry loop is already part of retry_transfer_wrapper.
The polling routine is common to the network protocols:udp, tcp and, once merged, sctp.
Align stack variables manually in yadif x86 code, should fix ticket25Signed-off-by: Michael Niedermayer <email@example.com>
move image check down in ffplay.cSigned-off-by: Michael Niedermayer <firstname.lastname@example.org>
Disable dr1 in ffplay for svq1 (fixes ticket40)Signed-off-by: Michael Niedermayer <email@example.com>
Fix parser: mark av_parser_parse() for removal on next major bump
Fix the breakage introduced with 9073ca6c6ebb9e3e9b60d968b7aa16cc8b6d558d
Check sizes a bit in ffplay, cant hurt.Signed-off-by: Michael Niedermayer <firstname.lastname@example.org>
Fix ticket34Signed-off-by: Michael Niedermayer <email@example.com>
Drop checkheaders from check, its uselessSigned-off-by: Michael Niedermayer <firstname.lastname@example.org>
Fix ticket20Signed-off-by: Michael Niedermayer <email@example.com>
swscale: postpone sws_getContext removal until next major bump.
The replacement isn't ready yet.
fate: add AAC LATM test
Add PlayStation Portable PMP format demuxer
Not yet complete, for demuxing AAC the AAC header must be generatedmanually.Possibly the decoder could accept the header as extradata to simplifythis.
mmst: get rid of deprecated AVERRORs
lxfdec: use AVERROR instead of deprecated AVERROR_NOMEM.
Reemove remaining uses of deprecated AVERROR_NOTSUPP.
lavf: if id3v2 tag is present and all else fails, guess by file extension
Revert "ffmpeg: fix aspect ratio setting"
This reverts commit d1eb50bb29caad9745631759265f475177df99b9.Reason, it breaks fate
Fix ffmpeg-mt fixme in h264.Signed-off-by: Michael Niedermayer <firstname.lastname@example.org>
Fix REBASE_PICTURE with h.264Signed-off-by: Michael Niedermayer <email@example.com>
Fix crash in ffmpeg.c with PIX_FMT_NONESigned-off-by: Michael Niedermayer <firstname.lastname@example.org>
Revert "Set the correct target for mingw64 dlltool" Revert due to dlltool/lib revertThis reverts commit 417516f63f8a9ee214ee9fa6fcb483962ea0f096.
Fix gazeta-20090408-1456-aa.tsregression introduced by 15957b197a63d08c33c34f08ef8b161335736914Signed-off-by: Michael Niedermayer <email@example.com>
psymodel: extend API to include PE and bit allocation.
Signed-off-by: Janne Grunau <firstname.lastname@example.org>
avio: always compile dyn_buf functions
Fixes build with --disable-muxers --disable-network.
Thanks to Hendrik Leppkes for noticing.
Do not offer unsupported HWAccel PIX_FMT.
Only fixes MPEG-1/MPEG-2 as a first step.
Use CONFIG_VDPAU as condition for using ff_vdpau_mpeg_picture_complete.
CONFIG_VDPAU is the condition on which ff_vdpau_mpeg_picture_completeis compiled in, so it is more appropriate, particularly since theseparate VDPAU decoder should be removed in the longer term.
ac3dec: avoid pointless alloc and indirection for input_buffer
Since we now always allocate it, it can simply be made part of the contextinstead.
Remove unnecessary parameter from ff_thread_init() and fix behavior
thread_count passed to ff_thread_init() is only used to set AVCodecContext.thread_count, and can be removed. Instead move it to the legacy implementationof avcodec_thread_init().
This also fixes the problem that calling avcodec_thread_init() with pthreads...
documentation: fix @url syntax in developer.texi