Yes, thanks for that link - I'll have to study it as it looks quite useful. I'm awareof the Goertzel Alogithm and its use in detecting specific frequencies but am trying to figure out omething with a more general application. Not, specifically dtmf tones, although this could be an application. What was getting me confused was what bin number corresponds to what frequency and where the "extra bin"comes from - if in fact it exists. Juging from the replies to te question I think there is fairly widespread confusion regarding this question. The quest for an "optimal" sample rate derives from trying to minimise the number of samples in order to use a small microcontroller such as the Atmel mega32. I have a 64 sample implementation that will perform an FFT in about 10mS - or about 1/2 that if I don't perform the square root stage - and detecting dtmf tones is an easy way of testing it. Another option is to sample at varying sample rates (for subsequent runs) and then correlating the results to a) get finer resolution of the result and b) enable alias recognition and therefore detect higher frequency (super-Nyqist?) components. Obviously this requires a more or less continuous signal. Just something I'm playing with - not aimed at a particular purpose at this stage. Richard P On 14/07/06, Scott Dattalo wrote: > On Thu, 2006-07-13 at 09:51 +1200, Richard Prosser wrote: > > > I'm trying to select an optimum sample rate to detect a > > non-harmonically related tone sequence (e.g. dtmf tones) but the > > simulation results I'm getting are rubbish. > > Why do you need an "optimum sample rate"? As long as you're sampling fast > enough then all you need is proper filtering. > > But, instead of FFT's you may wish to consider either DFT's or even better > the Goertzel algorithm centered at specific frequencies. > > If you're seeking an estimate for the frequency with the most energy and > you want the answer quickly, then you may wish to consider the algorithm > discussed here: > > http://www.dattalo.com/technical/theory/dtmf.html > > Scott > -- > http://www.piclist.com PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist