David Chocholatý

> This is an interesting Bit: https://breakfastquay.com/rubberband/code-doc/classRubberBand_1_1RubberBandStretcher.html#a3cd2ea056ec0eb94087f2fca2a41b080 Yeah, it looks cool for the R3 version. I have seen it once before and IMO it looks like they tried to make...

> > > I thing that is why our original approach equerry the group delay fails. > > > Do you mean with the "original approach" the approach in this...

> Cool, does it depends on the sample rate? I think we can find out it easily if we watch the ring buffer load in this implementation. Yeah, you're right....

> I expect they will vanish if we move this to a pull model from a ring buffer with size 8192. Oh, I think that I take it now all...

> Since to much group delay makes this effect hard to use and after this is merged we have always the worst case group delay. > We need to find...

> > I am right with that, sounds really awesome. > > Yes, that was the idea. Perfect, thank you.

I totally agree, the comparison shows us which approach is better and which we should prefer. There is just a little bug I found out and should be fixed or...

> For a decision, a graph group delay vs pitch change would be nice. This can be done with the metronome effect chained with the pitch shift effect, recorded with...

> you forgot to replace the place where you actually used `SampleBuffer::subspan`. Just change those and `amend` the commit. Yeah, thank you. I figured it out now too. It should...

The failing test for macOS CI: ``` [==========] Running 1 test from 1 test suite. [----------] Global test environment set-up. [----------] 1 test from EngineEffectsDelayTest [ RUN ] EngineEffectsDelayTest.DelayGreaterThanDelayBufferSize ../src/test/engineeffectsdelay_test.cpp:34:...