I'm wondering whether or not anyone out there is currently working to utilize Perl for audio manipulation, in a (mostly) pure-perl implementation? By pure-perl, I mean the code written by the developer, not necessarily any modules being used.

What I'm really wondering is what is the best approach to working with the actual waveform data? Just using perl arrays seems easy enough, where each sample is expressed as an array element, but really bogs down in memory and CPU time when the number of samples exceeds a second's worth at 44,100Hz. I've been using a substr()-based solution, but this has a tendancy to be a big hassle when working with bitrates higher than 8 (and if you don't use pack/unpack, you can end up with garbled data). I've heard of using PDL for this, but can't find a reference to clear examples.

Some examples of using susbtr() to manipulate wave data can be found at http://www.digitalkoma.com/church/projects/perl-dsp.html.

Is anyone out there also wrestling with wave data, and have some suggestions?