If high sample rates make no discernible difference when recording (which is effectively what Dan Lavry's much-quoted white paper seems to be saying), then what advantage is there in upsampling within plug-ins? For example, Cytomic's The Glue sounds better in the 4x upsampling mode, but why is this? And if working already at high sample rates in your DAW project, should upsampling make no difference at all?
Via SOS web site
SOS Technical Editor Hugh Robjohns replies: Lavry's basic premise is that there is no reliable evidence that the human population can perceive sounds higher than about 20kHz, and therefore that a sample rate in excess of about 44kHz should be entirely adequate for audio recording and reproduction, given a properly engineered system. Moreover, he highlights the fact that the temporal accuracy inherently diminishes with higher sample rates. For these reasons he argues that there is absolutely no benefit in recording with sample rates higher than about 96kHz, and that the optimal sample rate — not that we have the option to use it, sadly — is probably 60kHz to avoid some of the remaining practical engineering difficulties associated with 44.1kHz. However, we are talking here about recording and reproducing high-quality audio, which is a relatively straightforward thing. Complex non-linear signal processing — like dynamics and some forms of equalisation — is a whole different kettle of fish, and there are some significant mathematical advantages in converting to a higher sample rate for some forms of signal processing. This approach doesn't change the audio content in any way, it just makes the number crunching easier and considerably more accurate, with fewer unwanted artifacts. If you are already working at high sample rates, then further upsampling is unlikely to deliver any noticeable benefits, but only because it is very unlikely that there would be anything audibly significant in the upper parts of the captured spectrum anyway.