The Matroska muxer now allows WebVTT subtitle tracks to be written
while in WebM muxing mode.
WebVTT subtitle tracks have four kinds: "subtitles", "captions",
"descriptions", and "metadata". Each text track kind has a distinct
Mastroska CodecID and track type, as described in the temporal
metadata guidelines here:
http://wiki.webmproject.org/webm-metadata/temporal-metadata/webvtt-in-webm
When the stream has codec id AV_CODEC_ID_WEBVTT, the stream packet is
serialized per the temporal metadata guidelines cited above. The
WebVTT cue is written as a Matroska block group. The block frame
comprises the WebVTT cue id, followed by the cue settings, followed by
the cue text. (The block timestamp is synthesized from the cue
timestamp.)
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This allows to read a live isml movie and segment it using the
smoothstreaming muxer, which requires the bitrates to be known for each stream.
Signed-off-by: Alexandre Sicard <alexandre.sicard@smartjog.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The RTP timestamps can be decreasing for codecs with B-frames. For
these cases, make sure the timestamps in the MP4 file track itself
are nondecreasing, and add an offset to the RTP packet hint instead
to produce the intended RTP timestamp.
Signed-off-by: Martin Storsjö <martin@martin.st>
This fixes speex in rtmp
Fixes Ticket2409
the nellymoser in flv case actually needs larger analyzeduration. The code
previously just failed to calculate the duration
If this causes any problems, like premature analyze/probe end, please report!
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This fixes crashes when playing back certain RealRTSP streams.
When invoked from the RTP depacketizer, the full realmedia
demuxer isn't invoked, but only certain functions from it, where
a separate AVIOContext is passed in as parameter (for the buffer
containing the data to parse). The functions called from within
those entry points should only be using that parameter, not
s->pb. In the depacketizer case, s is the RTSP context, where ->pb
is null.
Cc: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes sure the ffurl_read_complete function actually
returns the number of bytes read, as the documentation of the
function says, even if the underlying protocol uses AVERROR_EOF
instead of 0.
Signed-off-by: Martin Storsjö <martin@martin.st>
A sid 0 would be mismatched to the attachment.
Prevent NULL pointer dereference.
Reported-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
CC: libav-stable@libav.org
start_granule should be applied to the stream referenced in the fisbone packet, not to the
Skeleton stream.
This was broken in d1f05dd183 and produced bogus warnings about
multiple fisbone in the same stream on files with more than one stream.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The toc is inexact and not using it can thus make sense.
Using it is faster though, thus the opposite can similarly makes sense
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This should be closer to how tcp behaved longer ago and should
fix the issue with idle connections timing out.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The method guess_ni_flag needs to divide timestamps in the index
by sample_size if it is set in order to compare different streams correctly.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Data is appended in fill_buffer() when there is sufficient space left
and the data pointer only reset when needed.
Previously the data pointer was more often reset, loosing more seekback
space than otherwise needed.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The sample from ticket #2691 currently does not trigger "Consider increasing
the value for analzeduration and probesize" because the audio streams are
only added after calling estimate_timings(). Attached patch moves the message
below this function call.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>