Opened 5 years ago
Closed 5 years ago
#273 closed defect (fixed)
Wrong HMVP table size
Reported by: | lizhang | Owned by: | |
---|---|---|---|
Priority: | minor | Milestone: | VVC D5 v8 |
Component: | spec | Version: | VVC D5 v7 |
Keywords: | Cc: | ksuehring, bbross, XiangLi, fbossen, jvet@… |
Description
The HMVP table size is fixed to be 5 as in the software and JVET-M0436, but the specification uses (maximum merge number - 1) which will cause mismatch between SW and specification when maximum merge number is less than 6.
Suggested fixes are as follows:
In sub-clauses 8.6.2.6 and 8.5.2.16 for the Updating process for the history-based motion vector predictor candidate list, replace the "MaxNumMergeCand − 1" by 5.
Change history (6)
comment:1 Changed 5 years ago by ccc2384823
comment:2 Changed 5 years ago by LiZhang
Please note that using fixed table size is what has been adopted. You may refer to the following meeting notes:
Reduce HMVP number from 6 to 5
In addition, the specification is not aligned with the software. I don't think it is necessary to make a proposal to fix typos.
comment:3 Changed 5 years ago by bbross
- Version changed from VVC D5 v5 to VVC D5 v6
comment:4 Changed 5 years ago by bbross
- Version changed from VVC D5 v6 to VVC D5 v7
comment:5 Changed 5 years ago by bbross
- Milestone set to VVC D5 v8
comment:6 Changed 5 years ago by bbross
- Resolution set to fixed
- Status changed from new to closed
Fixed in D5 v8
I agree that it's reasonable to do VTM / WD alignment as your discussion
But, I think maybe a new proposal to fix it may be better.
It's suggested that maybe a new proposal will be needed ( instead of using ticket to do normative change) for this issue in next JVET meeting.
Thanks!
ChunChia from MediaTek