As I understand it, asynchronous sample rate conversion should, in principle, be able to produce results with no audible degradation. Essentially, what asynchronous sample rate conversion involves is simulating an extremely high resolution D->A conversion in the
digital domain, followed by a simulated A->D conversion at the required sample rate. At least, that's how I think it works... ...