Bill Tracey wrote:
OK -- I'll take a look at adding support for handling +/- 1 sample offset in the DttSP audio callbacks...
FB. Not sure how you want to handle this, and I'm absolutely certain you don't need to be reminded of it, but there should surely be some mechanism for reinitialization of the delay memory on arbitrary resets.
Perhaps you could handle this ziplessly by pre-performing dummy write(s) into the appropriate ringbuffers where required. That's a way of handling arbitrarily-long delays without having to explicitly allocate delay memories for them.
Also +1 isn't necessary, ever. The delay of 1 sample (delay of -1) is a causal linear filter. +1 corresponds to having ESP about what data will happen in the future and is accordingly nondeterministic and nonlinear. No need to make more trouble yourself than you have to :-)
73 Frank AB2KT
