Paul Licameli

In practice you don’t need a large noise sample. Even a second is more than you need.

> In practice you don’t need a large noise sample. Even a second is more than you need. Reiterating this. If you actually select so much that the progress dialog...

Remember that overlapping computation and i/o with more threads is a possibility that might make it faster than in old versions.

The PR has not yet passed code review. I am recommending some simplification of it.

Performance should also be tested in the wave-clip-refactoring branch after #6094 is completed.

> Judging by the code style you are pulling in some of your older work. > > I have only looked on `lib-concurrency`. I think there few changes needed, but...

> > At least ... on one computer, with the averages of a few before-and-after trials, with 6 cores and hyperthreading. > > You have 12 threads and just a...

I should have put the big red Experimental label on this bit of holiday hackathon. I didn't ask review yet, but thank you.

@saintmatthieu @vsverchinsky Rebased onto latest master. There is much cleanup of AudioIO.cpp that I prepared. It's still not all of the cleanup. Correct inter-thread communication of the stream of most...

I have completed a review of changes in AudioIO.cpp. If my contribution of per track meters is also wanted, there is a lot of needed work to make these compatible....