Opened 4 years ago
Closed 4 years ago
#1217 closed defect (fixed)
Unit of input location (8.8.3.5)
| Reported by: | fbossen | Owned by: | |
|---|---|---|---|
| Priority: | minor | Milestone: | |
| Component: | spec | Version: | VVC D10 vD |
| Keywords: | Cc: | ksuehring, bbross, XiangLi, fbossen, jvet@… |
Description
The location (xCb, yCb) input to 8.8.3.5 appears to be intended to be in sample units of the colour component cIdx. In 8.8.3.2 it is described as a luma location when invoking 8.8.3.5. This seems incorrect. It is also inconsistent with other usage of (xCb, yCb) in 8.8.3.2.
Furthermore, in 8.8.3.5, these locations are compared to virtual boundary locations without appropriate scaling related to chroma subsampling.
Change history (4)
comment:1 Changed 4 years ago by bbross
- Resolution set to fixed
- Status changed from new to closed
comment:2 Changed 4 years ago by fbossen
Note that scaleWidth and scaleHeight should be locally defined. Currently they are only locally defined in the SAO process.
comment:3 Changed 4 years ago by fbossen
- Resolution fixed deleted
- Status changed from closed to reopened
Reopening since scaleWidth and scaleHeight do not appear to be appropriately defined in JVET-S2001-vE
comment:4 Changed 4 years ago by bbross
- Resolution set to fixed
- Status changed from reopened to closed
Thanks, both will be defined in deblocking as well in JVET-S2001-vF.
Good catch, thanks! Fixed in JVET-S2001-vE by removing "luma" in the invocation and scaling the virtual boundary locations according to scaleWidth and scaleHeight.