Thanks Mike! I’ll look more later today.
I skimmed the RFC and didn’t see this, but what kind of changed behavior may result? Anything breaking anticipated? I like the RFC but think we should document anything that would “break” from the fixing
Depends what we call “breaking” There are no API changes, only changes to what is output by the UGen, and this is almost entirely confined to the first block of samples (filter settling times), and more often just the first few samples (correcting a delay, or what would appear to the user as a slight change in initial phase).
My approach is that when there are any more significant changes in output, or judgement calls as to how things should be initialized, I’ll submit those as separate PRs for discussion. One example is the Impulse change, which I’ll also be taking back up as part of this effort.