Opened 8 years ago

Closed 8 years ago

#1180 closed defect (fixed)

Incorrect ordering of initialisation values for sig_coeff_flag

Reported by: peterderivaz Owned by:
Priority: major Milestone: RExt D5
Component: RExt text Version: RExt D4 (N1005) v3
Keywords: Cc: davidf, joel, teruhiko, jct-vc@…

Description

Table 9-29 "Values of initValue for ctxIdx of sig_coeff_flag" has had 6 extra values added at the end.

Table 9-4 says that initType 0 uses values 0..41 and 126..127

However, judging from 9.3.4.2.5, the new context values are added by equation 9-23 at locations 27 (for luma), and 15+28=43 (for chroma). This is fine for Chroma (e.g. for I slices, the state for a ctxInc of 43 will be initialised by the new initValue at position 127), but not for Luma.

I believe there are 3 problems with the definition of the CABAC values:

  1. Table 9-39 says that ctxInc for sig_coeff_flag only takes values 0..41.
  1. The new CABAC values have been added in the opposite order to the ones in the code (the same ordering problem as reported for intra_bc_flag in https://hevc.hhi.fraunhofer.de/trac/hevc/ticket/1179)
  1. Table 9-29 has had the new luma initValues inserted at the wrong location (because equation 9-38 has changed to add on 28 instead of 27, all the initValues from position 27 upwards should have been shifted)

Change History (2)

comment:1 Changed 8 years ago by DefaultCC Plugin

  • Cc davidf joel teruhiko jct-vc@… added

comment:2 Changed 8 years ago by davidf

  • Milestone set to RExt D5
  • Resolution set to fixed
  • Status changed from new to closed

Fixed in O1005v1

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

  • David Flynn(Subscriber, Participant)
  • jct-vc@…(Subscriber)
  • Joel(Subscriber)
  • karl.sharman@…(Always)
  • Karsten Suehring(Always)
  • Peter de Rivaz(Reporter)
  • Teruhiko Suzuki(Subscriber)