An element name of poc_lsb_not_present_flag[] in slice_segment_header(),i.e., if( nuh_layer_id > 0 && poc_lsb_not_present_flag[ nuh_layer_id ] | | ( nal_unit_type != IDR_W_RADL && nal_unit_type != IDR_N_LP ) )is missed in WD.

nuh_layer_id should be changed to LayerIdxInVPS[nuh_layer_id].

In I.7.3.8.5.2 Coding unit extension syntax, depth_dc_sign_flag is signaled when depth_dc_abs[x0+k][y0+j][i] is greater than 0 as follows:

dcNumSeg = DmmFlag[ x0 + k ][ y0 + j ] ? 2 : 1 if( depth_dc_flag[ x0 + k ][ y0 + j ] )

for( i = 0; i < dcNumSeg; i ++ ) {

depth_dc_abs[x0+k][y0+j][i] if(depth_dc_abs[x0+k][y0+j][i] >0)

depth_dc_sign_flag[x0+k][y0+j][i]

}

On the other hand,in HTM-12.0, it is signaled when the absolute value of DcOffset is greater than 0 as follows:

Void TDecSbac::xParseDimDeltaDC( Pel& rValDeltaDC, UInt uiNumSeg ) {

UInt absValDeltaDC = 0; xReadExGolombLevel( absValDeltaDC, m_cDdcDataSCModel.get(0, 0, 0) ); rValDeltaDC = (Pel)absValDeltaDC + ( uiNumSeg > 1 ? 0 : 1 ); ... if( rValDeltaDC != 0 ) {

UInt uiSign; m_pcTDecBinIf->decodeBinEP( uiSign );

} ...

}

Signaling condition for depth_dc_sign_flag in WD should be aligned with HTM-12.0 as follows:

for example

dcNumSeg = DmmFlag[ x0 + k ][ y0 + j ] ? 2 : 1 if( depth_dc_flag[ x0 + k ][ y0 + j ] )

for( i = 0; i < dcNumSeg; i ++ ) {

depth_dc_abs[x0+k][y0+j][i] if((depth_dc_abs[x0+k][y0+j][i] - dcNumseg + 2) >0) or if((depth_dc_abs[x0+k][y0+j][i] + (dcNumseg > 1 ? 0 : 1 ) >0)

depth_dc_sign_flag[x0+k][y0+j][i]

}

Introduced by m22570: Mismatch with software related to floating point /integer aritmetics in wedgelet pattern generation process. It is preferred to have integer arithmetics in the software.