> Can You evaluate practical degree of cross-contamination while demosaicing of > RAW shots from popular DSLRs? Is it huge, moderate or negligible? Look at the noise pattern in the red channel of a raw shot for blue patch. Determine noise spectrum. Run demosaicking. See noise spectrum correlation in green and blue channels after the demosaicking. Depending on algorithm and implementation, from nearly none to substantial. -- Best regards, Iliah Borg