Since we are basically seeking the AVIOContext under the subdemuxer, we
need to flush the subdemuxer to avoid old packets from being read from
the packet queue after the seek.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Properly take stream_index into account so that a keyframe will be
looked for in the specified stream_index only.
Similarly, only check timestamp validity against the specified
stream_index.
Also remove code for stream_index == -1 case which does not actually
happen as it is handled by generic code.
This is based on an initial patch by James Deng.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Do not try to delay opening newly required playlists until a segment
switch. Applications expect that newly selected undiscarded streams are
available immediately, especially with alternative rendition streams
(selectable audio/subtitle tracks).
One might think that delaying variant stream switch until a segment
switch would allow a "seamless" switch without us having to download a
specific segment from two different variant playlists. However, that is
not the case, since the application would have to keep the previous
stream available (undiscarded) until the first packet of the newly
selected stream arrives, but by that time the demuxer would have already
downloaded the next segment of both variants.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
As per spec 3.4.3 ("A client MUST NOT assume that segments with the same
sequence number in different Media Playlists contain matching content.")
we cannot use sequence numbers for packet ordering.
This can be seen e.g. in the subtitle streams of
bipbop_16x9_variant.m3u8 that have considerably longer segments and
therefore different numbering.
Since the code now exclusively syncs using timestamps that may wrap, add
some additional checking for that.
According to the HLS spec all the timestamps should be in 33-bit MPEG
format and synced together.
v2: cleaner wrap detection
v3: further wrap detection improvements
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
While selecting a packet to return to caller in read_packet(), the code
corrects the timestamps for starting timestamps.
However, this is wrong, since for live streams the initial timestamps
might differ just because of the time delay between the retrieval of the
various Media Playlists.
Fortunately, spec 6.2.4 mandates that all variant streams must have
matching timestamps, so we do not need to correct for initial
timestamps.
Drop the correction code.
Note that ID3 timestamps were previously ignored, so this code was
previously actually needed.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Check if the playlist is still needed just before requesting the next
segment instead of after exhausting the previous segment.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Improve selection of the segment sequence number when restarting the
reception of a playlist after it was suspended due to being unneeded
(due to discard flags).
The current code assumes that each playlist contains matching data with
the same sequence number, while spec 3.4.3 specifically says that that
is not the case. Often subtitle playlists also have longer target
durations as well, causing the selection to be completely wrong.
Instead prefer using the playlist segment duration information for
non-live playlists, and other means if that is not possible.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Seeking needs to be tracked on a per-playlist basis, since the resyncing
code in hls_read_packet() has to sync each playlist to the seek
timestamp instead of stopping after the first playlist has reached it.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
HLS provides MPEG TS timestamps via ID3 tags in the beginning of each
segment of elementary audio streams.
v2: fix issues with streams that have multiple ID3 tags
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Add support for EXT-X-BYTERANGE added in HLS protocol v4.
v2: Better comment explaining ffurl_seek call and fix cur_seg_offset not
being updated.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Even if we returned AVERROR_EOF previously due to playlist no longer
being needed, we may still be called again, and we do not want to
trigger a segment download in that case.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
HLS protocol version 4 added alternative renditions to the
specification (e.g. alternative audio tracks).
The EXT-X-MEDIA tags can also contain metadata for "renditions" (i.e.
tracks) of the main Media Playlist.
Add support for those.
Note that the same rendition (AVStream) may be associated with multiple
variants (AVPrograms).
Alternative subtitle tracks will require additional work and are
therefore not enabled yet.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
Session data (cookies, user-agent) is not being sent on payload requests with
encrypted HLS streams This causes services like Akamai to give a 403 forbidden
when requesting the TS files, because they expect the same cookies
and user-agent on all requests
Not all "sub-playlists" are variant playlists (containing the same
content with a different bitrate, etc) in the current version of the HLS
specification. They can now also be alternative renditions, containing
e.g. alternative audio tracks etc.
Decouple playlists from variants to prepare for handling the new
features.
Signed-off-by: Anssi Hannula <anssi.hannula@iki.fi>
* commit '8c929098141ebc94ad3f303521c520bb3dc6d8f6':
hls: Check whether the AVIOContext contains a new redirected URL
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This allows both the main playlist itself as well as the variant
playlists to handle redirects combined with relative URLs.
Signed-off-by: Martin Storsjö <martin@martin.st>
* qatar/master:
hls: Call avformat_find_stream_info() on the chained demuxers
Conflicts:
libavformat/hls.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This allows the chained demuxer (or more precisely, the lavf
utility code) to better fill in timestamps on packets from
these, especially for cases where one stream is a raw ADTS
stream.
Signed-off-by: Martin Storsjö <martin@martin.st>
* commit '06205b5efdcf0bc4c5463bfdd02f09b5f79fc4cd':
hls: Free packets when skipping packets when seeking
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'c44191039944526dd7eb6e536990b555837961f5':
hls: Store all durations in AV_TIME_BASE
Conflicts:
libavformat/hls.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'e1d5b244761cf69db655ad7ece1dbf2c13dd4fce':
hls: Store first_timestamp in units of AV_TIME_BASE
Conflicts:
libavformat/hls.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Also parse segment durations as floating point, which is allowed
since HLS version 3.
This is based on a patch by Zhang Rui.
Signed-off-by: Martin Storsjö <martin@martin.st>
When first_timestamp was stored as-is, its actual time base
wasn't known later in the seek function.
Additionally, the logic (from 795d9594cf) for scaling it
based on stream_index is flawed - stream_index in the seek
function only specifies which stream the seek timestamp refers
to, but obviously doesn't say anything about which stream
first_timestamp belongs to.
In the cases where stream_index was >= 0 and all streams had the
same time base, this didn't matter in practice.
Seeking taking first_timestamp into account is problematic
when one variant is mpegts (with real timestamps) and one variant
is raw ADTS (with timestamps only being accumulated packet
duration), where the variants start at totally different timestamps.
Signed-off-by: Martin Storsjö <martin@martin.st>
* qatar/master:
hls: Create an AVProgram for each variant
Conflicts:
libavformat/hls.c
See: 23db5418ed
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit '9d64f236292ba28018dd9afd2d57f8f944b33f81':
hls: Respect the different stream time bases when comparing dts
Conflicts:
libavformat/hls.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'c11e33a3d9665dd1fc5dbdecdd03a4860ac6a622':
hls: Set stream offset before opening a chained demuxer
Conflicts:
libavformat/hls.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'cdd2d73d315ecaf19ff49e64c91923275f1bda68':
hls: Don't check discard flags until the parent demuxer's streams actually exist
hls: Copy the time base from the chained demuxer
Conflicts:
libavformat/hls.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* commit 'eb33ba04e03d9f36e23fffd442510c824be709c3':
hls: Return all packets from a previous variant before moving on to the next one
Conflicts:
libavformat/hls.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Without the information, an application may choose audio from one
variant and video from another variant, which leads to fetching two
variants from the network. This enables av_find_best_stream() to find
matching audio and video streams, so that only one variant is fetched.
Signed-off-by: Martin Storsjö <martin@martin.st>
Also adjust the streams timestamps according to their start
timestamp when comparing. This helps getting correctly interleaved
packets if one stream lacks timestamps (such as a plain ADTS
stream when the other variants are full mpegts) when the others
have timestamps that don't start from zero.
This probably doesn't work properly if such a stream is
temporarily disabled (via the discard flags) and then reenabled,
and such streams are hard to correctly sync against the other
streams as well - but this works better than before at least.
The segment number restriction makes sure all variants advance
roughly at the same pace as well.
Signed-off-by: Martin Storsjö <martin@martin.st>
If passing the end of one segment while initializing the
chained demuxer, the parent demuxer's streams aren't set up
yet, so we can't recheck the discard flags.
Signed-off-by: Martin Storsjö <martin@martin.st>
This serves as a safeguard; normally we want to use the dts
comparison to interleave packets from all active variants. If that
dts comparison for some reason doesn't work as intended, make sure
that all packets in all variants for a certain sequence number have
been returned before moving on to the next one.
Signed-off-by: Martin Storsjö <martin@martin.st>
As far as I can tell the code should not change behaviour
depending on locale in any of these places.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
As the name indicates we can't just assume what the
"opaque" field contains.
This fixes a crash in third-party applications see e.g.:
http://bugzilla.mplayerhq.hu/show_bug.cgi?id=2126
This fixes also FFmpeg trac #2293, which is a different
third-party application.
Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Without the information, application may choose audio from one variant
and video from another variant, which leads to fetch two variants from
network. This enables av_find_best_stream() to find matching audio and
video streams, so that only one variant is fetched from network.
Signed-off-by: LYF <yefei.li@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Some HLS servers return 403 when the Range header is present. Disabling http
seekability probing prevents the header from being added.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
* qatar/master:
avformat: Drop pointless "format" from container long names
swscale: bury one more piece of inline asm under HAVE_INLINE_ASM.
wv: K&R formatting cosmetics
configure: Add missing descriptions to help output
h264_ps: declare array of colorspace strings on its own line.
fate: amix: specify f32 sample format for comparison
tiny_psnr: support 32-bit float samples
eamad/eatgq/eatqi: call special EA IDCT directly
eamad: remove use of MpegEncContext
mpegvideo: remove unnecessary inclusions of faandct.h
af_asyncts: avoid overflow in out_size with large delta values
af_asyncts: add first_pts option
Conflicts:
configure
libavcodec/eamad.c
libavcodec/h264_ps.c
libavformat/crcenc.c
libavformat/ffmdec.c
libavformat/ffmenc.c
libavformat/framecrcenc.c
libavformat/md5enc.c
libavformat/nutdec.c
libavformat/rawenc.c
libavformat/yuv4mpeg.c
tests/tiny_psnr.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* qatar/master: (35 commits)
h264_idct_10bit: port x86 assembly to cpuflags.
x86inc: clip num_args to 7 on x86-32.
x86inc: sync to latest version from x264.
fft: rename "z" to "zc" to prevent name collision.
wv: return meaningful error codes.
wv: return AVERROR_EOF on EOF, not EIO.
mp3dec: forward errors for av_get_packet().
mp3dec: remove a pointless local variable.
mp3dec: remove commented out cruft.
lavfi: bump minor to mark stabilizing the ABI.
FATE: add tests for yadif.
FATE: add a test for delogo video filter.
FATE: add a test for amix audio filter.
audiogen: allow specifying random seed as a commandline parameter.
vc1dec: Override invalid macroblock quantizer
vc1: avoid reading beyond the last line in vc1_draw_sprites()
vc1dec: check that coded slice positions and interlacing match.
vc1dec: Do not ignore ff_vc1_parse_frame_header_adv return value
configure: Move parts that should not be user-selectable to CONFIG_EXTRA
lavf: remove commented out cruft in avformat_find_stream_info()
...
Conflicts:
Makefile
configure
libavcodec/vc1dec.c
libavcodec/x86/h264_deblock.asm
libavcodec/x86/h264_deblock_10bit.asm
libavcodec/x86/h264dsp_mmx.c
libavfilter/version.h
libavformat/mp3dec.c
libavformat/utils.c
libavformat/wv.c
libavutil/x86/x86inc.asm
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Previously, we returned any error code except AVERROR_EOF to the
caller - only if AVERROR_EOF or 0 was returned, we proceeded to
the next segment.
With some setups of web servers, using Connection: close in https
and GnuTLS, we don't get a clean error code at the end of segments.
In those cases, just proceed to the next segment.
Tested-by: Antti Seppälä <a.seppala@gmail.com>
Signed-off-by: Martin Storsjö <martin@martin.st>
* qatar/master:
libspeexenc: add supported sample rates and channel layouts.
Replace usleep() calls with av_usleep()
lavu: add av_usleep() function
utvideo: mark interlaced frames as such
utvideo: Fix interlaced prediction for RGB utvideo.
cosmetics: do not use full path for local headers
lavu/file: include unistd.h only when available
configure: check for unistd.h
log: include unistd.h only when needed
lavf: include libavutil/time.h instead of redeclaring av_gettime()
Conflicts:
configure
doc/APIchanges
ffmpeg.c
ffplay.c
libavcodec/utvideo.c
libavutil/avutil.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* qatar/master: (36 commits)
adpcmenc: Use correct frame_size for Yamaha ADPCM.
avcodec: add ff_samples_to_time_base() convenience function to internal.h
adx parser: set duration
mlp parser: set duration instead of frame_size
gsm parser: set duration
mpegaudio parser: set duration instead of frame_size
(e)ac3 parser: set duration instead of frame_size
flac parser: set duration instead of frame_size
avcodec: add duration field to AVCodecParserContext
avutil: add av_rescale_q_rnd() to allow different rounding
pnmdec: remove useless .pix_fmts
libmp3lame: support float and s32 sample formats
libmp3lame: renaming, rearrangement, alignment, and comments
libmp3lame: use the LAME default bit rate
libmp3lame: use avpriv_mpegaudio_decode_header() for output frame parsing
libmp3lame: cosmetics: remove some pointless comments
libmp3lame: convert some debugging code to av_dlog()
libmp3lame: remove outdated comment.
libmp3lame: do not set coded_frame->key_frame.
libmp3lame: improve error handling in MP3lame_encode_init()
...
Conflicts:
doc/APIchanges
libavcodec/libmp3lame.c
libavcodec/pcxenc.c
libavcodec/pnmdec.c
libavcodec/pnmenc.c
libavcodec/sgienc.c
libavcodec/utils.c
libavformat/hls.c
libavutil/avutil.h
libswscale/x86/swscale_mmx.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
This avoids reading any old data in the AVIOContext buffer after
the seek, and indicates to the mpegts demuxer that we've seeked,
avoiding continuity check errors.
Signed-off-by: Martin Storsjö <martin@martin.st>
Enhance seeking by demuxing until the requested timestamp is
reached within the segment selected by the seek code using the
playlist info.
Some mpegts streams don't have dts set for all packets though,
this seeking method doesn't work well for that case.
Signed-off-by: Martin Storsjö <martin@martin.st>
* qatar/master: (21 commits)
CDXL demuxer and decoder
hls: Re-add legacy applehttp name to preserve interface compatibility.
hlsproto: Rename the functions and context
hlsproto: Encourage users to try the hls demuxer instead of the proto
doc: Move the hls protocol section into the right place
libavformat: Rename the applehttp protocol to hls
hls: Rename the functions and context
libavformat: Rename the applehttp demuxer to hls
rtpdec: Support H263 in RFC 2190 format
rv30: check block type validity
ttadec: CRC checking
movenc: Support muxing VC1
avconv: Don't split out inline sequence headers when stream copying VC1
rv34: handle size changes during frame multithreading
rv40: prevent undefined signed overflow in rv40_loop_filter()
rv34: use AVERROR return values in ff_rv34_decode_frame()
rv34: use uint16_t for RV34DecContext.deblock_coefs
librtmp: Add "lib" prefix to librtmp URLProtocol declarations.
movenc: Use defines instead of hardcoded numbers for RTCP types
smjpegdec: implement seeking
...
Conflicts:
Changelog
doc/general.texi
libavcodec/avcodec.h
libavcodec/rv30.c
libavcodec/tta.c
libavcodec/version.h
libavformat/Makefile
libavformat/allformats.c
libavformat/version.h
libswscale/x86/swscale_mmx.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
When this demuxer was created, there didn't seem to be any
consensus of a common short name for this protocol. Now
the consensus seems to be to call it hls.
Signed-off-by: Martin Storsjö <martin@martin.st>
Anssi Hannula
Mikael Finstad
Michael Niedermayer
Martin Storsjö
kowalsky
LYF
Zhang Rui
Timothy Gu
Reimar Döffinger
Micah Galizia
Diego Biurrun
Clément Bœsch
Duncan Salerno
Lou Logan
Stefano Sabatini
Mans Rullgard
Panagiotis H.M. Issaris
Alex Converse