Re: supernova sample-based scheduling bug?

On Sat, May 22, 2021 at 11:33 PM <[email protected]> wrote:
> On the flip side, compared to my original testing (https://scsynth.org/t/imperfection-of-language-based-timing/350), the bespoken issue seems to have dramatically improved. Now I have tested on OS 10.15.7 with 3.11.2 – no idea what caused the change.
> These awful jumps (in the size of dozens of samples) that happened every 10-20 seconds or so in earlier tests (which I only performed in scsynth though) seem to have completely disappeared. ...

I wonder if NTP has anything to do with this in Mac?

I was experiencing some quite nasty timing instability in Linux. I
finally switched off NTP and I haven't seen any of that kind of
instability in a couple of months now (even had a show last weekend,
which went off without a hitch, timing or otherwise).

I don't expect turning off NTP to result in sample accuracy, but -- if
something in your system is changing the system clock time
periodically, and SC is resyncing to the system clock every 20 seconds
or so, that's asking for trouble.

hjh

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/