Commit graph

1460 commits

Author SHA1 Message Date
Benjamin M. Schwartz
a09d9fdece theoradec: PARs of 0:x, x:0 and 0:0 are all allowed and map to 1:1
Fixes #609252.
2010-02-08 11:21:35 +01:00
Vincent Penquerc'h
6b8175d96a oggdemux: use the default granpos functions for kate streams
Set timestamps on kate packets. See bug #600929.
2010-02-07 10:16:28 +01:00
Sebastian Dröge
4914aeda2f oggdemux: Don't leak allocated buffers
This can happen if the combined flow return is not OK although the
allocation succeeded or if the packet in question is a BOS and we're
not going to push headers.

Fixes bug #608699.
2010-02-02 10:33:07 +01:00
Vincent Penquerc'h
ed2e09133e oggdemux: sparse streams aren't timed by end time, and their duration isn't implicit
Fixes timestamps and durations on Kate subtitle streams.

See http://www.xiph.org/ogg/doc/ogg-multiplex.html section 'start-time and
end-time positioning' for some more details, and bug #600929.
2010-01-25 23:50:09 +00:00
Vincent Penquerc'h
25873a050f oggdemux: properly set up the media type for kate streams
See #600929.
2010-01-25 23:49:53 +00:00
Wim Taymans
e4b68a3658 oggdemux: use right type for the serialno
Use a consistent type for the serialno to avoid problems when comparing between
signed and unsigned variants.

Fixes #607926
2010-01-25 15:14:56 +01:00
Wim Taymans
95511c8821 oggdemux: don't push headers twice
Don't push the stream headers twice but only in the activation of a chain.

Fixes #607929
2010-01-25 14:00:52 +01:00
Wim Taymans
62f8c3c672 oggdemux: rename a variable
Rename the 'seekable' variable to 'pullmode'. We might be able to seek in push
mode too eventually.
2010-01-25 12:31:24 +01:00
Tim-Philipp Müller
e1c319949d oggdemux: fix crash when freeing headers
Use _ogg_packet_free() instead of gst_mini_object_unref in one more
place now that the header list contains ogg packets and not buffers.

file: Stephen_Fry-Happy_Birthday_GNU-nq_600px_425kbit.ogv
2010-01-24 13:32:20 +00:00
Sebastian Dröge
1652005f26 oggdemux: Strip trailing \0 for subtitle OGM streams
Fixes bug #607870.
2010-01-24 08:57:13 +01:00
Sebastian Dröge
a66ce94adc oggdemux: Correctly set DELTA_UNIT flag for OGM streams 2010-01-23 22:10:09 +01:00
Sebastian Dröge
b38dcf5277 oggdemux: Don't strip all 0-bytes from the end of OGM packets
This fixes broken packets pushed downstream by oggdemux for
MPEG4 streams for example.
2010-01-23 22:10:09 +01:00
Sebastian Dröge
592bc87dc9 oggdemux: Extract tags from OGM text streams and don't push them downstream 2010-01-23 22:10:08 +01:00
Sebastian Dröge
541da50c53 oggdemux: Store header/queued packets as ogg_packet and use normal peer chaining functions to pass them downstream 2010-01-23 22:10:08 +01:00
Sebastian Dröge
d78cffb2e8 oggdemux: ...and set caps on queued packet buffers too 2010-01-23 13:20:46 +01:00
Sebastian Dröge
b63a582bf9 oggdemux: Set caps on header buffers 2010-01-23 13:19:44 +01:00
Wim Taymans
4b9666aedb oggdemux: keep track of added pads
Keep track of the pads we added and removed.
Remove some unused fields.
Don't add pads for which we don't have caps.
2010-01-21 17:39:52 +01:00
Wim Taymans
06e2d2791b oggstream: don't call NULL setup functions
If we find a known mapper but it doesn't have a setup function, simply skip it
instead of crashing.
2010-01-21 17:39:51 +01:00
Wim Taymans
4ed58b5882 oggstream: avoid division by 0 on bad annodex streams 2010-01-21 17:39:51 +01:00
Thiago Santos
4b771bff7a oggdemux: No need to subtract begin time
Last stop is already based on the chain start and there is no need
to subtract the chain start as it may lead to a negative overflow.
This was causing seeking issues when the target chain was not
the first one (that has chain start = 0)

Fixes #606382
2010-01-19 08:39:14 -03:00
Thiago Santos
125f7dfdb0 oggdemux: granulepos is relative to its chain
When performing seeks, the granulepos should be offset by
its chain start time to avoid using wrong values to
update segment's last_stop. A sample file is indicated on
bug #606382
2010-01-18 15:39:55 -03:00
Sebastian Dröge
82f3adaa9b theora: Rename source files to have the same name as the headers 2010-01-14 10:10:23 +01:00
Sebastian Dröge
1de4169c1c vorbis: Rename source files to have the same name as the headers 2010-01-14 10:07:22 +01:00
Sebastian Dröge
d4e9225ed4 vorbis: Move channel layout definitions into a single separate file
...instead of having two copies.
2010-01-14 10:05:35 +01:00
Sebastian Dröge
68c6ddc2fc vorbis: Add official 6.1 and 7.1 channel mappings
These are in the Vorbis spec since 2010-01-13. Fixes bug #606926.
2010-01-14 08:20:38 +01:00
Edward Hervey
7d2c90423a oggdemux: Fix unitialized variable.
If the package isn't handled, gracefully return GST_FLOW_OK.
2010-01-12 16:35:50 +01:00
Wim Taymans
0201326db1 oggdemux: push headers when activating chains
Keep a list of headers for each stream of a chain. When a chain is activated,
push the headers before pushing the data so that decoders can sync.
Fix seeking in chains, take the chain start time into account when comparing
timestamps.

See #606382
2010-01-08 16:57:40 +01:00
Mark Nauwelaerts
3f4820d6c2 oggdemux: decide flac header packet by content rather than count 2010-01-06 14:03:44 +01:00
Mark Nauwelaerts
66545eb7a2 oggdemux: reset header packet count at bos page 2010-01-06 14:03:41 +01:00
Olivier Crête
15fb8f9bf7 textoverlay: Ignore zero framerate
https://bugzilla.gnome.org/show_bug.cgi?id=606163
2010-01-06 07:49:27 +01:00
Mark Nauwelaerts
8fb183c7c9 oggdemux: enhance flac packet duration calculation 2010-01-05 16:44:58 +01:00
David Schleef
b3c15a340a theoraenc: Add encoder controls for libtheora 1.1
Added drop-frames, cap-overflow, cap-underflow, and rate-buffer.
2009-12-20 17:34:46 -08:00
Wim Taymans
e897373a0a ogg: ogm video has constant packet duration 2009-12-11 12:20:13 +01:00
David Schleef
118b62a764 oggdemux: implement old fLaC mapping 2009-12-10 22:48:23 -08:00
Wim Taymans
d18118418e oggdemux: remove redundant fields 2009-12-10 16:09:09 +01:00
Wim Taymans
43bdb13ec2 vorbisdec: adapt to new oggdemux
Remove all granulepos hacks and simply use the timestamps from the new oggdemux
like any other decoder.
2009-12-10 13:22:00 +01:00
Wim Taymans
79b25e634c vorbisdec: fix peer query 2009-12-09 19:04:18 +01:00
Wim Taymans
1a7c8c7919 theoradec: fix query 2009-12-09 17:24:01 +01:00
Wim Taymans
f1fdf5d3e4 theoradec: small cleanups 2009-12-09 16:55:52 +01:00
Wim Taymans
08c16c0db7 vorbisdec: use gst_pad_peer_query() 2009-12-09 16:38:43 +01:00
Sebastian Dröge
01ca87e785 textoverlay: Only flush downstream on seeks for flushing seeks 2009-12-09 09:47:30 +01:00
Sebastian Dröge
d5205bb0fa textoverlay: Proxy buffer allocation on the video sinkpad to the srcpad 2009-12-09 09:36:26 +01:00
Wim Taymans
1818d795ee theora: remove granulepos hacks
Remove the granulepos hacking now that oggdemux outputs timestamps like any
other demuxer.
2009-12-08 17:37:56 +01:00
Wim Taymans
1ad0e4342e oggdemux: improve keyframe seeking
Improve keyframe seeking.
Fix reverse playback.
2009-12-07 18:49:43 +01:00
Wim Taymans
c53cd385e3 oggdemux: implement keyframe seeking
Implement keyframe seeking in oggdemux by doing the double seek trick. First
seek to the required position, then read pages for all streams to grab the
granulepos (to know the timing of the keyframe) of each stream, then seek back
to the first keyframe.
2009-12-07 15:42:05 +01:00
Wim Taymans
56d3856620 oggdemux: fix timestamps after seek
After a seek, discard all packets before the packet with the granulepos on it so
that the output buffers contain valid timestamps.

Reorder some code so that we check the timestamps before allocating and pushing
an output buffer.

Do more checks on valid packets in ogm mode.
2009-12-04 16:35:09 +01:00
Wim Taymans
eb4fc976e5 oggdemux: add comment 2009-12-04 15:39:59 +01:00
Wim Taymans
9e222a385c oggdemux: don't do math with invalid granulepos
When the current granulepos is unknown and set to -1, don't try to add durations
to it.
2009-12-04 14:01:11 +01:00
Wim Taymans
96e1c4dbe6 oggdemux: guard against wrong granulepos
Clamp the initial granulepos to 0 instead of going negative for some badly muxed
ogg files.
2009-12-04 13:14:57 +01:00
Wim Taymans
05f4605180 theoradec: don't fail on bogus granulepos
Do some additional checks on the granulpos timestamp before using it for
calculating the duration because oggdemux generates wrong granulepos now.
Fixes seeking somewhat again.
2009-12-04 12:26:47 +01:00