Re: [Audacity-nyquist] RE: New "Silence Finder" Plug-In, Need Help on Memory Use

[Alex S. B. <ale...@al...>]

WOW! Thanks for the feedback on the memory use and the rms function. I am
working on a version 2 using these techniques.

I believe that the biggest memory-use growth is in the "do" loop, not the
initial SETF.  I calculated that for a 20-minute segment, the loop is going 1.2
million times, so just a 200 bytes of memory use each round would lead to over
200 MB over the course of the function. Maybe that is part of the problem.

I am going to try using rms and a lower sample rate (maybe 10 to 100 Hz), and I
am also going to try iterating through the sample using SREF instead of
SND-FETCH. One of these techniques should turn out to be a winner! Just
reducing the sample rate should reduce the memory problems dramatically, if
they are happening in the "do" loop.

RE your PS: Audacity will split the tracks into WAV, MP3 or any other supported
format according to the labels.  Under the File menu, there are "export
multiple" commands that will create multiple files, one for each label. Nyquist
plus Audacity does the trick.

Thanks again for all the leads,
--Alex

Quoting "Roger B. Dannenberg" <rb...@cs...>:

> Thanks for the new plug-in. 
> I think the memory usage is due to (SETF S1 ...) near the beginning. This
> will assign the sound to a global variable, so as the sound is read, the
> global will accumulate samples until the entire sound is memory-resident. 
> 
<snip>
> PS Now, what will it take to have Nyquist split the tracks according to the
> labels?

----------------------
Alex S. Brown, PMP
ale...@al...
http://www.alexsbrown.com/