Opened 18 years ago
Closed 18 years ago
#1788 closed enhancement (fixed)
H264 tools patch
Reported by: | Owned by: | danielk | |
---|---|---|---|
Priority: | minor | Milestone: | 0.20 |
Component: | mythtv | Version: | head |
Severity: | medium | Keywords: | h264 |
Cc: | Ticket locked: | no |
Description
This patch contains a set of tools needed to enable H264 support in MythTV. It contains a minimalist state machine class used to detect H.264 keyframes. A usage example is included in the IP Network Recorder patch.
Attachments (1)
Change History (4)
Changed 18 years ago by
Attachment: | mythtv_crc_h264_r9878.tar.2.gz added |
---|
comment:1 Changed 18 years ago by
Milestone: | unknown → 0.20 |
---|---|
Status: | new → assigned |
comment:2 Changed 18 years ago by
The throw()'s are only there to formally show that the code doesn't throw any exceptions so it has nothing to do with the error checking of any C/C++ libs. In fact, in the current code there are no used C/C++ libs. It is sometime a standard practice to include the throw statement in every function even if they are empty but they can be removed if not in the coding standards of MythTV.
comment:3 Changed 18 years ago by
Resolution: | → fixed |
---|---|
Status: | assigned → closed |
(In [9919]) Fixes #1788. Refs #1787. Adds much of the code needed for H.264 (aka MPEG-4 AVC) TS decoding in MythTV.
In particular this adds it to the MythCodecID and other enums, and adds a H.264 keyframe/frame finder to avformatdecoder. You would still need to get a H.264 video into MythTV with a keyframe map for proper playback with skipping (see #1787).
The patch was submitted by Jean Michel Bouffard on behalf of the Queen of Canada and the Canadian Communications Research Centre.
Very cool.
NALUnitType should be formatted so that the major bits are indented 4 chars instead of 8, and we line things like:
to be formatted as so:
or better yet:
But those I could fix. What I'm concerned about is the "throws()" I didn't see you throwing anything, but does this means you are not checking for common error values from the C/C++ libs?