Re: [Audacity-nyquist] Using audacity to automatically generate amplitude data
A free multi-track audio editor and recorder
Brought to you by:
aosiniao
From: Roger D. <rb...@cs...> - 2007-10-07 18:05:18
|
Mitch asks about getting amplitude data after time points specified in a file: Since your input is a text file and audio, and your output is a text file, I would certainly use Nyquist directly for this problem. With Nyquist, you can read segments of a sound, e.g. from a time point for 0.2 seconds, making it especially easy to iterate through the labels, compute the RMS of a segment, and save the answers to a file. Also, Nyquist can read/write text, so it can be programmed to save the data where you want in the form you want. Edgar writes that analysis in Nyquist is not practical because of memory issues. This is definitely a problem, especially in Audacity, but only because the simplest or most natural-looking programs have a tendency to save samples in memory. There are ways to access sounds such that samples are not retained in memory. I'm not 100% certain of this in every case because there are subtle interactions with Audacity. If there's a generic problem, e.g. find the maximum amplitude for each spectral bin, that characterizes analysis problems of interest, I'd be willing to work out a constant-memory implementation that might serve as a template for other analysis tasks. -Roger |