From: Chris B. <ch...@cn...> - 2006-12-27 16:00:23
|
Reuben Thomas wrote: > On Wed, 27 Dec 2006, a quegg wrote: > > >> On a similar vein, whilst removing null effects from >> the chain was admittedly only a nicety in most cases, >> for resampling, it fixed the bug that these effects >> could not handle having nothing to do, so we'll need >> something here to handle that situation. >> > > This is in my todo list. Let's fix the effects that can't handle doing > nothing (my removal of the special case was with flushing this sort of bug > into the open in mind). > > (I haven't looked at the code we are talking still but...) Most of these you can't fix really. The math to resample from 1 rate to the same rate just breaks down for example. At best you can do is catch the bad case inside the effect and instead do a memcpy from inbuf to outbuf; but then its kinda a waste of time. Optimizing out/stripping out these type of effects from the effects chain is going to have the best performance. Chris |