Re: Most accurate method of fundamental frequency determination ?

Sorry, of course I know that there can be multiple
zero-crossings within a given period of a multi-frequency
tone. I did say PERIOD zero crossings. I hoped I didn't
have to explain about low-pass filtering and that I was
referring to only to the period epochs that repeat
themselves. I rarely used the zero-crossings myself, but
measured from max peak to max peak or min dip to min dip. I
usually used AMDF for most of my work, but the writer wanted
super fine detail of the FO and you can't get that with any
correlation or frequency domain method. Measuring the
samples within each period epoch gives the most accuracy.

--
Chip Wood

"robert bristow-johnson" <rbj@xxxxxxxxxxxxxxxxxxxx> wrote in
message news:BF87FE37.BA52%rbj@xxxxxxxxxxxxxxxxxxxxxxx
> in article djtohm$ni6$1@xxxxxxxxxxxxxxxxxxx, Chip Wood at
> chip.wood@xxxxxxxxxxxx wrote on 10/28/2005 13:52:
>
> > While I appreciate all the involved frequency domain
> > approaches, what ever happened to the KISS time domain
> > approach of just counting samples between period
> > zero-crossings?
>
> Chip, when the OP asks about "fundamental frequency
determination", i
> presume that the input waveform is not restricted to pure
sinusoids but
> probably some general periodic or quasi-periodic function.
zero crossing,
> of course, will not work for that.
>
> but your KISS principle is still a valid point. why not
simply use some
> form of AMDF or ASDF?
>
> --
>
> r b-j rbj@xxxxxxxxxxxxxxxxxxxx
>
> "Imagination is more important than knowledge."
>
>


.