Opened 11 years ago

Closed 11 years ago

#810 closed defect (fixed)

Table 9-32 : maxBinIdxCtx is incorrect for some syntax elements

Reported by: PhuongNguyen Owned by: bbross
Priority: minor Milestone: D10
Component: Text Version: D10 (L1003) v2
Keywords: maxBinIdxCtx Cc: bbross, wjhan, jct-vc@…

Description

Some syntax elements have bypass bins which are not of a prefix and suffix. The value of maxBinIdxCtx specified in table 9-32 might be wrong for some of them.

+ sao_type_idx_luma and sao_type_idx_chroma : maxBinIdxCtx should be 1 instead of 0

+ merge_idx : maxBinIdxCtx can be 1, 2 or 3 but not 0. Since table 9-37 specifies the bypass type for bins 1, 2 and 3, a coherent choice is 3 for maxBinIdxCtx.

Also, in table 9-37, we should fill binIdx 3 and 4 with na.

Best regards,
Phuong Nguyen.

Change History (12)

comment:1 Changed 11 years ago by DefaultCC Plugin

  • Cc bbross wjhan jct-vc@… added

comment:2 Changed 11 years ago by PhuongNguyen

Please read :

Also, in table 9-37, we should fill binIdx 3 and 4 with na for ref_idx_l0 and ref_idx_l1.

comment:3 Changed 11 years ago by bbross

maxBinIdxCtx equal to 0 is correct for sao_type_idx_luma, sao_type_idx_chroma and merge_idx because the bin with index 0 is the last bin that is coded using context derivation. All bins with binIdx > 0 are bypass coded.

For which Syntax Element should we fill binIdx 3 and 4 with na?

comment:4 Changed 11 years ago by PhuongNguyen

For sao_type_idx_luma and sao_type_idx_chroma, if we apply the text as it is :

From 9.3.2, bypassFlag will be not defined for binIdx = 1 because a value is assigned to ctxIdxOffset for this syntax element (so the first entry of bypassFlag is not applied) and maxBinIdxCtx = 0 (so the second entry of bypassFlag is not applied).

If bypassFlag is not defined, the decoding process flow cannot be applied as it uses bypassFlag to determine context or bypass decoding should be used for a given bin.

My suggestion is :

In 9.3.2, we add a case for bins with binIdx > maxBinIdxCtx :

Otherwise, for each possible value of binIdx up to the specified value of maxBinIdxCtx given in Table 9-32, a specific value of the variable ctxIdx is further specified in subclause 9.3.3. In such a case, bypassFlag is set equal to 0. For binIdx greater than maxBinIdxCtx, bypassFlag is set equal to 1.

Then for Table 9-32, some of the values for maxBinIdxCtx need to be modified :
+ part_mode : maxBinIdxCtx should be 2
+ cu_qp_delta_abs : maxBinIdxCtx should be 4
+ ref_idx_l* : maxBinIdxCtx should be 1

comment:5 Changed 11 years ago by bbross

  • Milestone set to D9
  • Version changed from D9 (K1003) v4 to D9 (K1003) v10

comment:6 Changed 11 years ago by bbross

  • Resolution set to fixed
  • Status changed from new to closed

comment:7 Changed 11 years ago by bbross

  • Resolution fixed deleted
  • Status changed from closed to new

comment:8 Changed 11 years ago by bbross

  • Version changed from D9 (K1003) v10 to D9 (K1003) v11

comment:9 Changed 11 years ago by bbross

  • Milestone D9 deleted
  • Version changed from D9 (K1003) v11 to D10 (L1003) v1

comment:10 Changed 11 years ago by bbross

  • Milestone set to D10

comment:11 Changed 11 years ago by bbross

  • Version changed from D10 (L1003) v1 to D10 (L1003) v2

comment:12 Changed 11 years ago by bbross

  • Resolution set to fixed
  • Status changed from new to closed

The CABAC part was cleaned up including removing a lot of redundancies in the aforementioned tables. Please check whether everything is ok in the current version v23.

Note: See TracTickets for help on using tickets.

This list contains all users that will be notified about changes made to this ticket.

These roles will be notified: Reporter, Owner, Subscriber, Participant

  • Benjamin Bross(Owner, Subscriber, Participant)
  • jct-vc@…(Subscriber)
  • karl.sharman@…(Always)
  • Karsten Suehring(Always)
  • Woo-Jin Han(Subscriber)