>Mike Berry wrote:
>>
>> Soundhack analysis files can be used in csound.  So you can do the
extraction
>> in soundhack, analyze the output files, and use the analysis with pvoc.
>> --

>Richard Dobson Wrote:
>I don't have a Mac, so Soundhack is denied me.

Exactly! One of the reasons for the original request was because although I
have access to Macs, I am primarily a 'PC' user. If the feature was added to
Csound anyone could use it.

>I would like to know what the
>criterion is for marking a frequency  as stable or otherwise - ie - over
how
>many windows?
>

Soundhack allows the user to set the number of frequency bands used in
analysis (large no. = better freq reponse, small=better time response etc.),
and the analysis frame size, which is defined in terms of the FFT size (e.g.
a value of 10 will use 10 FFT frames for the spectral analysis). Transients
are then taken as partials which have a deviation of more than 'x' Hz per
analysis frame, and stable frequencies are ones which deviate less than 'y'
Hz per analysis frame.

Another reason for adding this feature to CSound is that (ideally) the
deviation settings could be dynamic, thus greatly increasing the
power/usefulness of the function.

Regards,

JamieB