Can you clarify whether or not you already know the file names of the mp3 files that you want? Your post would suggest you do not know the filenames.

If you do not know the filenames, is there a web page that lists them? Your original post would suggest there is not.

If you do not know the filenames in adance, and you have no way to obtain the filenames, this isn't a Perl issue. You have to solve one of those two problems before you can consider using Perl tools to automatically crawl the podcasts with something like WWW::Mechanize