Opened 13 years ago Closed 10 years ago #255 closed defect (fixed)Potential Bug in HM (both in 4.0 and 5.0)
Description
When we did expeiments on HM4.0, we observed that the following section of the code does not seem to behave in the correct way.In the function "encodeResAndCalcRdInterCU" in the file "TEncSearch.cpp".
Change History (5)comment:1 Changed 13 years ago by DefaultCC Plugin
comment:2 Changed 13 years ago by ksuehring
comment:3 Changed 13 years ago by davidf
comment:4 Changed 11 years ago by fbossencomment:5 Changed 10 years ago by karlsharman
As identified, the transformIdx data structure can contain undefined values for a 2Nx2N skipped block (when all CBFs will be 0).
Fortunately, since the CBFs are all 0, the other places in the code that transformIdx are used are not affected by the undefined values (although the loop filter may test unnecessary TU edges before discovering that they belong to the same PU).
However, to avoid future confusion, this has been tidied in r4242. 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
|
In r3772, line 4551 and following in TEncSearch.cpp