Re:DRM’s Dead Canary: How We Lost the Web, What We Learned, ...
by StavrosK on Tuesday November 28, 2017 @11:12AM
(#15798496)
I use floats because it's most natural to represent duration in time. Nanoseconds are going to be about as lossy as floats too, I guess (although for a different reason). If you're going for maximum accuracy, you'd need to represent the duration as the number of samples.
I wonder if floats are going to be a significant problem, hmm.
I looked at your playlist format specs, why do you use floats? I'm currently coding on a GO audioserver and save time intervals as Nanoseconds (time.Duration). This makes more sense to me, as floats may behave badly in some contexts.