> I'm trying to develop a SAOL version of the CSound's foscil
> opcode, which is an FM oscillator.
Some ideas which may help
-- iphs isn't used in your math
-- sfront is currently doing linear interpolation on the tables (the
"normative" default for SA), if CSound does something better by
default this could cause spectral differences for sure. Try making
the table really big and seeing if the spectra come closer.
-- try running sfront with the -O0 option (the letter oh and the numeral
zero), and see if the spectra changes -- this turns off all optimization,
maybe there's a bug in the optimizer (although these days, its more
probable the non-optimized version crashes out, given that its under-tested)
--jl
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 11:46:43 EST