Scott Wilson

> A risk here is that we don't have a comprehensive SynthDef test suite Sorry to be a pain, but I really think we shouldn't merge this without a test...

Hi James, Agree completely that a test suite should be a group effort! It's been a long time since I've waded into this code, and I think it's been modified...

To get the ball rolling: Here are a few things to test that came up in past discussions over the years. Could be cleaned up and simplified. Swap LocalBuf and...

James, I think we're getting immersed in implementation details here, when what we really need to do is go back to the start and: 1. Come up with a clear...

> 2\. Come up with a comprehensive set of tests to capture that. BTW, I think – again given the centrality of this – a sensible precaution would be to...

> Users may expect to get triggers from `Impulse.ar(0, Phasor.ar(...))` (in the same way that we expect a sine wave from `SinOsc.ar(0, Phasor.ar(...))`) but in my testing, that doesn't happen....

> I didn't test it (I should -- this is a _very_ important use case!) but I believe it will be fine, based on lines 1023-1025 in the revision: Cool....

> I always assumed the compensation is for summing the input channels, not for the distribution of the output channels. E.g. if you had a `SplayAz.ar(1, [a, b, c])` you...

@adcxyz do you know what the original intention was?

> Actually, since SplayAz can also be used to scale channels up (distribute few into many channels), > the equal power by levelComp idea is somehow wrong conceptually already. >...