Opened 20 years ago
Closed 19 years ago
Last modified 19 years ago
#1591 closed defect (worksforme)
Bad seek tables again
| Reported by: | Owned by: | cpinkham | |
|---|---|---|---|
| Priority: | minor | Milestone: | 0.20 |
| Component: | mythtv | Version: | 0.19 |
| Severity: | medium | Keywords: | recordedmarkup |
| Cc: | Ticket locked: | no |
Description
DVB (Mpeg2) recordings made with 0.19 have had bad/corrupt seek tables. The offset information in recordedmarkup is a couple of bytes (2000-10000) behind the real keyframe byte position. As result other functionalities will be affected. Cutting for example doesn't match the desired positions. If I transcode the recording looseless without cutlist, the seektable is rebuild perfectly. If I run mythcommflag --rebuild the seektable is rebuild perfectly but with type 6 (MARK_GOP_START) instead of type 9 (MARK_GOP_BYFRAME).
Found a previours description of this issue in Ticked #816.
Change History (4)
Note:
See TracTickets
for help on using tickets.
I think this issue has been fixed in svn head or else we would have heard a lot more about it. If the problem is occurring in head please provide more detailed information such as a partial sample file and the seektable information from the recordedmarkup table to go with the dump. If using svn head as of around mid-July 2006, the seektable is now contained in the 'recordedseek' MySQL table.