miller puckette's book explains how to use pd in a mathematical way:
http://crca.ucsd.edu/~msp/techniques/latest/book-html/
julius smith's book explains all about filters:
http://ccrma.stanford.edu/~jos/filters/
this thread discusses interpolation algorithms (and shows that (imho) miller made a bad mistake in his 4-point interpolation algorithm):
http://lists.puredata.info/search/PD-list?query=better+tabread4~
other than that, pd audio algorithms are fairly standard/generic/simple, and not really documented other than in the help patches and examples (which concentrate on how to use them rather than how they work) and the pd source code isn't particularly descriptively commented (but is the best place to get details).