Opened 5 years ago
Closed 5 years ago
#842 closed defect (fixed)
parcat broken after slice header changes (JVET-Q0775)
Reported by: | ksuehring | Owned by: | |
---|---|---|---|
Priority: | major | Milestone: | VTM-8.0 |
Component: | VTM | Version: | VTM-7.3 |
Keywords: | Cc: | ksuehring, XiangLi, fbossen, jvet@… |
Description
parcat needs to be updated with any HLS changes that affect reading and patching of POC.
Since a new syntax element was added to the slice header, parcat now fails for streams generated after the merge of !1255
Change history (5)
comment:1 Changed 5 years ago by ksuehring
- Version set to VTM-7.3
comment:2 Changed 5 years ago by jonatan
comment:3 Changed 5 years ago by ksuehring
I think it's OK to fix parcat with the POC change. Asking to implement parsing of the whole PH seems unreasonable when POC is getting moved anyway.
comment:4 Changed 5 years ago by jonatan
A merge request for moving POC to PH has been submitted as !1300. The MR also includes modifications to parcat to operate with POC LSB signaled in PH.
comment:5 Changed 5 years ago by ksuehring
- Milestone set to VTM-8.0
- Resolution set to fixed
- Status changed from new to closed
I merged !1300
Note: See TracTickets for help on using tickets.
Thanks Karsten for reporting on the problem. I have started to look into how to address it.
Since PH can now be signalled in SH, parsing to the current location of POC LSB in the SH in the software would mean parsing through an entire PH in parcat. However, it was decided at the meeting to move POC LSB from SH to PH (recorded as decision on JVET-Q0115 and included in the PH text changes in JVET-Q0819).
I'm working on implementing the move of POC from SH to PH and update parcat accordingly.
Thus, my current thinking is to focus on that integration instead of implementing a temporary solution to parse through the entire PH in parcat. Please let me know if you disagree with this approach or have suggestions for a different (better) process for addressing the issue.
Best Regards Jonatan