Nikolay Pokhilchenko wrote:
Graeme, I see an analogy with audio analog-to-digital conversion (2D quantization - time and amplitude). May be the profiler must automatically limit the harmonic spectrum of device response bumps when the bumps spectrum is wider than can "digitize" the A2B.
Hi, I think that Gerhard has answered this well. The smoothing parameters used in creating the A2B are not about anti-aliasing, but instead sets a tradeoff between accurately interpolating the sample point values vs. reducing the effects of sampling inaccuracies (noise). There is a fairly complicated process whereby the smoothing parameters are set according to the notional measurement uncertainty (-r parameter), number of sample points and number of input dimensions. Figuring out what numbers to use for combinations of these three variables is not easy, and I can well believe that they are not always optimal. Graeme Gill.