Opened 9 years ago

Closed 6 years ago

#1384 closed enhancement (fixed)

orphaned symbol in the spec

Reported by: kolya Owned by:
Priority: minor Milestone: v5
Component: Text Version: v4 (12/2016)
Keywords: Cc: bbross, wjhan, jct-vc@…

Description (last modified by ksuehring)

The paragraph just before E.3.2 section says:

log2_max_mv_length_horizontal and log2_max_mv_length_vertical indicate the maximum absolute value of a decoded horizontal and vertical motion vector component, respectively, in quarter luma sample units, for all pictures in the CVS. A value of n asserts that no value of a motion vector component is outside the range of −2n to 2n − 1, inclusive, in units of quarter luma sample displacement. The value of log2_max_mv_length_horizontal shall be in the range of 0 to 16, inclusive. The value of log2_max_mv_length_vertical shall be in the range of 0 to 15, inclusive. When log2_max_mv_length_horizontal is not present, the values of log2_max_mv_length_horizontal and log2_max_mv_length_vertical is inferred to be equal to 15.

What "n" here stands for?

Change History (8)

comment:1 Changed 9 years ago by DefaultCC Plugin

  • Cc bbross wjhan jct-vc@… added

comment:2 Changed 9 years ago by ksuehring

(via reflector)

To me that statement doesn't seem very likely to be misinterpreted, but I agree that a different phrasing would be desirable.

"n" refers to the value of log2_max_mv_length_horizontal and log2_max_mv_length_vertical for the horizontal and vertical component of the MV, respectively.

Best regards,
Gary Sullivan

I was about to write a similar reply.

comment:3 Changed 9 years ago by ksuehring

  • Description modified (diff)

comment:4 Changed 9 years ago by ksuehring

I just noticed the different allowed range:

0..16 for log2_max_mv_length_horizontal
0..15 for log2_max_mv_length_vertical

Maybe the difference is there because for video in typical landscape format we can expect longer horizontal movements. But shouldn't both values be inferred to their respective maximum values and not 15?

Or is that difference a typo?

comment:5 Changed 9 years ago by ksuehring

Updating bug report from reflector discussion:

Gary:

See the equations at the end of clause 8.5.3.2.1. Doesn't that make a value bigger than 15 meaningless?

Karsten:

Yes, as the note states:

NOTE – The resulting values of mvLX[ 0 ] and mvLX[ 1 ] as specified above will always be in the range of −215 to 215 − 1, inclusive.

Does that mean, we can fix the typo?

Gary:

Yes I think we can just change it to 15 for both.

comment:6 Changed 6 years ago by bbross

  • Milestone set to v5
  • Version set to v4 (12/2016)

Agreed by the authors of the AHG2 report JCTVC-AD0002.
Fix two parts:

  1. Incorporate the definition of ‘n’ as suggested in one comment.

In E.3.1, the semantics of log2_max_mv_length_horizontal and log2_max_mv_length_vertical, change the following sentence:

A value of n asserts that no value of a motion vector component is outside the range of −2n to 2n − 1, inclusive, in units of quarter luma sample displacement.

to

A value of n asserts that no value of a motion vector component is outside the range of −2n to 2n − 1, inclusive, in units of quarter luma sample displacement, where n refers to the value of log2_max_mv_length_horizontal and log2_max_mv_length_vertical for the horizontal and vertical component of the MV, respectively.

  1. Set the range constraint on log2_max_mv_length_horizontal to 0..15 as suggested in another comment.

The suggested range constraint is present in the "Last Call" text for ITU-T HEVC text version 5.

comment:7 Changed 6 years ago by bbross

Ticket #1114 has been marked as a duplicate of this ticket.

comment:8 Changed 6 years ago by bbross

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

fixed in v5 (2018/02)

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(Subscriber, Participant)
  • jct-vc@…(Subscriber)
  • karl.sharman@…(Always)
  • Karsten Suehring(Participant, Always)
  • Nikolay Shlyakhov(Reporter)
  • Woo-Jin Han(Subscriber)