Actually that is not the case. See: http://www.underbit.com/products/mad/
libmad is used in a replacement MPEGA.library for amiga, resulting in significantly faster decoding than when using FPU: http://aminet.net/package/util/libs/mpega_libmad
Yeah, I always used integer back in the day. The FPU accelerates floating point, but fixed point is faster. Plus fixed point is more accurate.
All floating point maths does is sacrifice accuracy for range for lazy coders. You never use floating point for financial calculations, because it's innaccuracy means you lose money.
