In the latest HEVC spec, there is no specified behavior when transform_skip_flag is 0. We suggest the following semantics change.

Before:

transform_skip_enabled_flag equal to 1 specifies that transform_skip_flag may be present in the residual coding syntax. transform_skip_enabled_flag equal to 0 specifies that transform_skip_flag is not present in the residual coding syntax.

transform_skip_flag[ x0 ][ y0 ][ cIdx ] specifies whether a transform is applied to the associated transform block or not: The array indices x0, y0 specify the location ( x0, y0 ) of the top-left luma sample of the considered transform block relative to the top-left luma sample of the picture. The array index cIdx specifies an indicator for the colour component; it is equal to 0 for luma, equal to 1 for Cb, and equal to 2 for Cr. transform_skip_flag[ x0 ][ y0 ][ cIdx ] equal to 1 specifies that no transform will be applied to the current transform block. When transform_skip_flag[ x0 ][ y0 ][ cIdx ] is not present, it is inferred to be equal to 0.

After: transform_skip_enabled_flag equal to 1 specifies that transform_skip_flag may be present in the residual coding syntax. transform_skip_enabled_flag equal to 0 specifies that transform_skip_flag is not present in the residual coding syntax and transform will be applied to the current transform block.

transform_skip_flag[ x0 ][ y0 ][ cIdx ] specifies whether a transform is applied to the associated transform block or not: The array indices x0, y0 specify the location ( x0, y0 ) of the top-left luma sample of the considered transform block relative to the top-left luma sample of the picture. The array index cIdx specifies an indicator for the colour component; it is equal to 0 for luma, equal to 1 for Cb, and equal to 2 for Cr. transform_skip_flag[ x0 ][ y0 ][ cIdx ] equal to 1 specifies that no transform will be applied to the current transform block. transform_skip_flag[ x0 ][ y0 ][ cIdx ] equal to 0 specifies that transform will be applied to the current transform block. When transform_skip_flag[ x0 ][ y0 ][ cIdx ] is not present, it is inferred to be equal to 0.

Section 8.6.2.1.1 Table 8-58 -- Specification of hPos[2] and vPos[2] according to the type of sample adaptive offset process

When sao_type_idx=4, the sampling positions are (1,-1) and (-1,1). hPos and vPos should be:

hPos[0] = 1 hPos[1] = -1 vPos[0] = -1 vPos[1] = 1

The TAppDecTop::xFlushOutput() caused memory leak at decoder when DYN_REF_FREE is 0. The reconstructed pictures in DPB are not released.