Forums > Dev

[ot] how to recover from blown up coreaudio driver?

October 9, 2009 | 7:16 am

hi all,
i recently switched from a good old ppc laptop to one of those shiny new intel macbooks (actually a mbp). well, i’m certainly enjoying the speed of the machine, but feel that some strange changes have been made to how audio is handled by the os.
as an example, i’m working on some audio externals, and whenever some values go astray, coreaudio bursts into a loud click and dies.
and this does not only affect the program running (max in this case) but the whole os – no sound at all, or completely distorted.
and this is not restricted to max, of course. it also happens to me, when trying out au-plugins under development, e.g.

while i find this behavior less than desirable, i’m wondering how to recover from this scenario. quitting the affected program does not help, logging out/in does not help.
right now i’m restarting the machine every time this happens – reminds me of the glorious os9 days…

what do others do? there must be some easier/faster way.

thanks,
volker.


October 13, 2009 | 9:06 am

hm, silence…
this could mean
(a) nobody knows what the hell i’m talking about, cause i’m the only one who experiences this – reasons: broken machine, broken brain, broken whatever.
(b) there are indeed other people who see/hear the same behavior, but nobody has a better solution how to recover from this.

which is it, a, b or maybe c?

osx 10.5.7
max 4.6.3


October 13, 2009 | 2:32 pm

i’m having problems (audio dying) when changing samplerate, could be something similar


October 14, 2009 | 6:50 am

is it just dead and gone, or can you recover, e.g. by closing max?

here, if i send very high numbers or nan to the dac (which isn’t nice, and mostly unintended – but happens from time to time) audio blows up, like an exploding filter, and is dead – which is ok. but there seems no way to get it back working, other than restarting the machine, which is not ok at all.

i just noticed that max4 and max5 behave differently in this respect (max5 not killing audio) – and they _sound_ pretty differently for signals exceeding the -1+1 range.
seems to me that max5 is properly clipping signals before handing them over to coreaudio as opposed to max4, which gives a little more "creative freedom" with the added risk of screwing things up.

none the less, this audio lock up is news to me, since using 10.5 on an intel mac. can’t recreate it on ppc with 10.4 and max4.
and it’s not only happening in max. i had similar behavior with AU Lab, when testing some unfinished au plugins.


October 15, 2009 | 6:57 pm

You are correct that Max 5 clips audio signals in the core audio driver and that Max 4 did not.

Cheers


Viewing 5 posts - 1 through 5 (of 5 total)