Commit graph

54 commits

Author SHA1 Message Date
Vincent Penquerc'h
2dc41db862 h264parse: use the framerate that was copied a few lines above 2012-01-09 11:03:24 +00:00
Vincent Penquerc'h
4f27b2b4fe h264parse: fix test and set not using the same input
This fixes H.264 output not having a set framerate, and
confusing some of the elements downstream (such as interlace).
2012-01-06 10:54:11 +00:00
Mark Nauwelaerts
c4abaf1428 h264parse: fix and tweak frame timestamping
... to run with properly init'ed variables, and to only perform interpolation
in safe cases.
2012-01-04 11:16:18 +01:00
Mark Nauwelaerts
9d0c7d054e h264parse: prefer negotiating pass-through 2012-01-03 22:10:20 +01:00
Mark Nauwelaerts
dcd2ee421a h264parse: mark byte-stream AU frame with inserted codec NALs as keyframe 2012-01-03 22:08:07 +01:00
Mark Nauwelaerts
75b8ddb6cf h264parse: properly determine skip amount when no NAL start found yet 2012-01-03 20:27:07 +01:00
Mark Nauwelaerts
c54144cc16 h264parse: some more proper parse state reset 2012-01-03 20:26:51 +01:00
Alessandro Decina
570ba0c84a h264parse: avoid pushing SPS/PPS at start only if both are found in stream
Avoid injecting SPS/PPS nals only if both are already present in the stream.
Fixes some AVC clips that happen to have only PPS nals in stream.
2011-12-19 12:36:39 +01:00
Alessandro Decina
94ae4ca833 h264parse: do wait for a keyframe when handling force key unit events
Although some decoders can start decoding with just PPS and SPS and waiting for
a keyframe is bad when the keyframe interval is large, gst-ffmpeg does need a
keyframe to start decoding.
2011-12-18 17:40:57 +01:00
Alessandro Decina
a1305b949c h264parse: handle force key unit events
Send SPS/PPS on force key unit events. Useful for example when serving HLS
without transcoding.
2011-12-06 13:16:29 +01:00
Edward Hervey
1c50dab50c h264parse: Use faster parsing variant when input is packetized
Reduces latency and cpu usage

https://bugzilla.gnome.org/show_bug.cgi?id=665584
2011-12-05 19:04:17 +01:00
Mark Nauwelaerts
9604342af3 h264parse: reflow slice nal processing 2011-12-05 13:11:39 +01:00
Mark Nauwelaerts
ef24518b60 h264parse: optimize slice nal processing
Only minimal slice info is needed here to determine real frame start,
so no need to parse extensively.

Fixes #665418.
2011-12-05 13:11:04 +01:00
Alessandro Decina
4474db4e71 h264parse: let upstream PAR override in-stream PAR 2011-12-01 09:07:18 +01:00
Alessandro Decina
21b063a78c h264parse: rework PAR parsing
And in the process refactor _update_src_caps a bit
2011-12-01 09:06:56 +01:00
Luciana Fujii Pontello
6d85d4882a h264parse: respect pixel-aspect-ratio
Parse pixel-aspect-ratio in vui information.
2011-12-01 08:23:22 +01:00
Vincent Penquerc'h
7521b597f4 various: fix pad template ref leaks
https://bugzilla.gnome.org/show_bug.cgi?id=662664
2011-11-28 13:08:27 +00:00
Mark Nauwelaerts
1f3b4ad574 h264parse: remove redundant code
... now that what should not be needed is really not needed anymore to cover
up bogus stuff that has been FIXMEd.
2011-11-18 17:20:10 +01:00
Mark Nauwelaerts
7fae33a661 h264parse: fix up AU collection processing
On the one hand, no need to collect nal if processing last one.
On the other hand, ensure AU collection processing to have sufficient
next NAL data in normal cases.

Fixes #663180.
2011-11-18 17:20:08 +01:00
Mark Nauwelaerts
66543b3dd9 h264parse: mind to-be-skipped data in subsequent re-used offsets 2011-11-18 17:20:06 +01:00
Mark Nauwelaerts
e12044c359 h264parse: avoid additional NAL parsing if no AU collection is needed 2011-11-18 17:20:03 +01:00
Mark Nauwelaerts
9db239e1ec h264parse: some more get_caps tweaking
... by having it use sink template rather than src template, and not
passing an downstream parsed(=true) to upstream.
2011-11-18 17:20:00 +01:00
Sebastian Dröge
f3f9e4b978 h264parse: Implement ::sink_get_caps to allow stream-format conversion again
Just proxying the downstream caps will prevent h264parse from
accepting a different stream-format than what is supported
downstream, although it could convert to a different stream-format.
2011-11-16 10:39:40 -08:00
Jonas Larsson
5099ff23af h264parse: outgoing byte stream prefix always has size 4
Fixes #664123.
2011-11-16 10:57:30 +01:00
Sreerenj Balachandran
f8239ff692 h264parse: correct debug message and remove some dead code
See #661113.
2011-10-17 11:17:50 +02:00
Sreerenj Balachandran
0a478db7dd h264parse: correct parsing of "numberOfPictureParameterSets"
See #661113.
2011-10-17 11:17:47 +02:00
Thibault Saunier
f94236078c h264parse: Make proper use of the h264parser lib in parse_chain
Minor debugging logs fixes
2011-10-06 12:49:19 -03:00
Mark Nauwelaerts
9c221f492d h264parse: properly interpret baseparse draining state
... which means the input data has reached the end, but need not exclude
many separate frames may still need parsing.

Fixes merging NALs into AU.
2011-10-06 12:26:23 +02:00
Sreerenj Balachandran
3129ea29b1 h264parse: correct the size argument of gst_h264_parser_identify_nalu_avc() 2011-10-06 11:59:41 +02:00
Sreerenj Balachandran
bbc30f3644 h264parse: minimum size of parsable avcC data should be 8 2011-10-06 11:59:38 +02:00
Mark Nauwelaerts
66f644cfcc h264parse: bring debug statements up to desired specs 2011-09-19 14:55:09 +02:00
Thibault Saunier
404c18ab18 h264parse: Make use of the h264_nal_parser_free function 2011-09-09 15:02:35 -03:00
George Kiagiadakis
ddf32e2c49 h264parse: Fix IDR position marking when the IDR is split into several NALs 2011-09-07 13:36:12 +03:00
Sebastian Dröge
d4147a2798 h264parse: Put stream-format and alignment into the src template caps
Also remove the parsed field from the sink template caps.
2011-09-07 11:29:37 +02:00
Sebastian Dröge
cfb986b6bf h264parse: Add Converter to classification
This parser can convert between different alignments and stream-formats.
2011-09-06 13:21:28 +02:00
Thibault Saunier
b968f82b18 h264parse: Port to the new h.264 parsing library 2011-09-02 15:46:02 +02:00
Alessandro Decina
f12598fa64 h264parse: remove a check made useless by the last commit 2011-07-18 11:51:23 +02:00
Alessandro Decina
5f6a805dfd h264parse: assume byte-stream if stream-format and codec_data are missing from input caps 2011-07-18 11:50:41 +02:00
Alessandro Decina
7bcc0c71b8 h264parse: ignore codec_data if stream-format=byte-stream
Some encoders set codec_data even when outputting byte-stream. Before this patch
h264parse would incorrectly detect such streams as AVC.
2011-07-18 11:50:41 +02:00
Mark Nauwelaerts
10bec5c08f h264parse: avc input must either pass-through or be split into bytestream chunks
... since parsing code is not meant to receive avc data.

Fixes #653087.
2011-06-21 12:46:37 +02:00
Mark Nauwelaerts
160207abb9 h264parse: avoid bogus frame parsing state 2011-06-21 12:46:35 +02:00
Mark Nauwelaerts
e5d5ff4394 h264parse: plug some more buffer leaks 2011-05-25 21:51:55 +02:00
Mark Nauwelaerts
1c4b6fee9b h264parse: compensate for skipped bytes in parsing state offsets 2011-05-25 21:51:55 +02:00
Mark Nauwelaerts
62d9ef42d4 h264parse: fix clearing adapter forming avc output frame
... which needs to be reset upon a new frame and otherwise (only)
for avc input in passthrough mode (rather than on every chain call).
2011-05-25 21:51:55 +02:00
Alessandro Decina
6a8d66728f h264parse: allow full negotiation for packetized input
... by defaulting to allow splitting packetized input and having
negotiation with downstream deciding whether or not this applies.

Also enable pass-through parsing mode if input and output format
(stream-format and alignment) match.

API: GstH264Parse:split-packetized (removed)

Fixes #650228.
2011-05-25 21:51:55 +02:00
Mark Nauwelaerts
f92f5b70fb h264parse: avoid stale parsing state 2011-05-23 16:57:25 +02:00
Mark Nauwelaerts
217e389689 h264parse: simplify minimum frame size handling 2011-05-23 16:57:25 +02:00
Mark Nauwelaerts
409a51bcea h264parse: gracefully handle truncated input NAL units
Rather than assert'ing in such case, emit warning if the length of a NAL unit
is less than expected 2 and discard it.

Based on patch by Benjamin M. Schwartz <bens@alum.mit.edu>

Fixes #650416.
2011-05-23 16:57:25 +02:00
Mark Nauwelaerts
c203fd52e8 h264parse: fix some buffer leaks
Fixes #650323.
2011-05-17 12:34:37 +02:00
Mark Nauwelaerts
02e392ca83 h264parse: fixup boilerplate and debug code style 2011-05-17 12:34:30 +02:00