Bojcha told me yesterday that he did some tests and - as far as I have heard - he did not see any crashes if he didn't restart it very fast; if that's indeed the case I could just add some delay between successive restarts when you're changing the buffer size - and actually even if that doesn't fix the issue it should make the chance of getting driver crashes much much smaller. Hm - maybe you can confirm this? It would be good to know if it acts the same on multiple pc's. Test is: Set normal input/output to something non-Juli@, then turn ASIO on and off a few times per second.).

Yes, I can confirm this, it pisses my Juli@ off..