incorrect modulo value used in adjacent mode calculations

[andrew.dorrell]

In section 8.2.2 and equations ranging from 8-7 through 8-53, adjacent angular modes are calculated for the purpose of populating a most probable mode list.The equations have the general form

new_mode = 2 + ((mode + a) % b)

where b is 64 and a is a function of the distance new_mode should be offset from mode. It could for example be determined as follows:

a = delta + NUM_ANGULAR_MODES - 2     (if delta is negative)
a = delta - 2                         (if delta is positive)

The defect arises because of the modulo b=64. This is a defect because it means that the maximum value that new_mode can take is 63+2=65 (even though 66 is a legitimate and legal value). This may also mean that some values for the offset amount 'a' need to be revised in the text. However this requires a clear statement of the intention of the calculation.

The consequences of this wrt the code are documented in https://jvet.hhi.fraunhofer.de/trac/vvc/ticket/127

[bbross]

Is there a mismatch between VTM and spec?
If not this seems like a normative change that would need to be discussed at the next meeting.

[bbross]

mod 64 was a design choice and thus VTM and draft text reflect the intended design.