Function: xCalculateInterlacedAddPSNR

Issue: Condition to call this function is not correct.

Test case: Gop period is 32. Interlaced frame related PSNR data collection should happen after POC 17 but it happens after POC 16 itself.

Some condition should be added based on "IRAPtoReorder" flag

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". In the section "Residual coding", after calculating the best RD cost. if ( pcCU->isSkipped(0) )

{

pcCU->setCbfSubParts( 0, 0, 0, 0, pcCU->getDepth( 0 ) );

}

To my understanding, this is similar to the case when the value of bSkipRes is true. Therefore, it seems that the above code should be changed to if ( pcCU->isSkipped(0) )

{

pcCU->setCbfSubParts( 0, 0, 0, 0, pcCU->getDepth( 0 ) ); pcCU->setTrIdxSubParts( 0, 0, pcCU->getDepth(0) );

}

As described in section 8.5.2.1.5, the 2nd step of the ordered steps of deriving the motion vector predictor mvpLX:

If both availableFlagLXA and availableFlagLXB are equal to 1 and mvLXA is not equal to mvLXB, availableFlagLXCol is set equal to 0, otherwise, the derivation process for temporal luma motion vector prediction in subclause 5 is invoked with luma location ( xP, yP ). It seems that we should append a checking before deriving the TMVP candidate in the function "fillMvpCand" of the file "TComDataCU.cpp" of HM6.0-rc1 as follows:

if (pInfo->iN < AMVP_MAX_NUM_CANDS) { if ( getSlice()->getPPS()->getEnableTMVPFlag() ) { ........ } }

In this way, we can eliminate the unnecessary burden of deriving TMVP.

After deriving the TMVP candidate, we can delete the following statement, because it is impossible to satisfy the judgment. if (pInfo->iN > AMVP_MAX_NUM_CANDS)

{

pInfo->iN = AMVP_MAX_NUM_CANDS;

}