Would full 24-bit delays (with reduced max time) be possible? If I am correct in thinking the normal delays are 16-bit, then the same memory allocation should give 2/3 the delay time.
I think it would be good for applications where the delay doesn't have to be particularly long but quality is more important (like ensemble effects where I've occasionally noticed quantisation noise).
24-bit delay possible?
Sputnki
#2
Not sure about 24, but looking at delay objects' code i think it's perfectly possible to make them into 32-bit.
Since variables are 32bit you'd have to discard a byte per sample, so no practical advantage.
Ah, that's a good point. I was thinking 24 because the converters are 24-bit, but of course the internal sample stream is a 32-bit fractional, right? So yeah, what I was asking for was full depth delay, I just forgot what the internal bit depth was.