Hi Bray, Thasp, all,
Some years ago, I was part of a listening panel doing double-blind tests of compression algorithms to be used for professional digital broadcasting and music distribution (including some from Dolby Labs). These tests were also duplicated by panels in the U.K and in Australia.
If the data rate was high enough--320 kbps--there was little or no difference in the various compression schemes. As the data rates went down, percussion instruments often showed up flaws. Castinets were really critical and cymbals as well. Also the sibilants on unaccompanied vocals would cause problems.
Pop/rock music generally wasn't as critical as classical works, but the sibilance flaws showed up on some vocals.
With the higher data rates, the compression schemes were essentially transparent. I've done comparisons of Dolby Digital 5.1 (a lossy compression scheme) with master mixes (no compression) and DD5.1 is essentially transparent.
Regards,
