audio time-out

    Jun 01 2006 | 5:46 pm
    Hi- I'm using max/msp/jitter (4.5 running on mac OSX.4.6 on dual-G5 powermac) for an interactive gallery installation. I've noticed that somewhere between 24 and 36 hours running, the patches stop producing audio, while the system sounds of the computer come through the sound system just fine, and the patch is otherwise functioning perfectly.
    I'm planning to put the patch in startup items, and configure the machine to periodically reboot and this should solve the problem from a practical standpoint. Still, I'm wondering if anyone has noticed this before, if it's documented anywhere.
    Oh, and I'm running two (quite different) patches on two identicle hardware setups, and got the same behaviour from both.
    any thoughts?

    • Jun 01 2006 | 6:06 pm
      so what is this hardware? maybe your audio-interface needs some sleep?
      m a x i m i l i a n m a r c o l l
    • Jun 04 2006 | 10:38 am
      A long time ago, Max used to have an absolute clock that would run out, and stop after a day or so. These days, I think that clock is now a very big number, but Max does have a sort of "mortality".
      I have a permenent sound installation, and I've implemented exactly your solution; it restarts periodically. I also have a product to restart my audio interfaces (MOTU 828's) when the computer restarts:
      It seems like the audio interfaces do lose it regularly, so restarting them is critical.
    • Jun 04 2006 | 11:54 am
      On all the gallery installations I have ever done I get the assistants to switch off the system at the end of each day, and switch it on again each morning. This has always worked well with a minimum of instruction to the assistants. One installation has now been running for four and a half years - there have been problems, but not because of the switching on/off.
      Lawrence Casserley - Lawrence Electronic Operations - Colourscape Music Festivals -
    • Jun 04 2006 | 1:05 pm
      I worked on an installation on a G3 which ran, as far as I'm aware, for a couple of years without interruption. (Or so I'm told; it was logging sensor input every 10 seconds into an SQL database, so I should probably arrange to check.) It wasn't doing audio output, but was analysing audio input signals from the sensors via one of the eMagic USB interfaces. I had an audio installation running for about six weeks on a Mac mini, and it ran continuously, 24/7, with no problems.
      I think it's basically a case of soak-testing, and if a particular interface causes problems, schedule a nightly reboot, or have it done manually, as Lawrence suggests.
      -- N.
      nick rothwell -- composition, systems, performance -- http://
    • Jun 04 2006 | 2:11 pm
      I should add the comment that most of my installations include a lot of other stuff besides the computer (lights, speaker systems, video cameras and projectors, various sensors, etc). If the gallery is open for, say, eight to ten hours, then there will be a fair amount of power being used for the fourteen to sixteen hours when it is closed. I think it is more environmentally friendly to shut it all down when it is not wanted. OK, the installation does come to rest if there is no input from the sensors, but there is a fair bit of standby power being used - and that adds up over a long show.
      Also, if something is going to be on continually for months, there is never a chance to test it for that long. I find it more comfortable to know that it will start afresh each day.
      Lawrence Casserley - Lawrence Electronic Operations - Colourscape Music Festivals -
    • Jun 04 2006 | 2:47 pm
      One of the installations was a rudimentary artificial intelligence, so it was rather important that the work was running all the time in order to model the mental behaviour (a mind cannot sleep or dream if it's turned off). Also, since the mind was encapsulated within a lamp post, it worked best at night.
      nick rothwell -- composition, systems, performance -- http://
    • Jun 05 2006 | 11:43 pm
      Quote: nick rothwell / cassiel wrote on Sun, 04 June 2006 07:47 ---------------------------------------------------- > the mind was encapsulated within a lamp post
      Isn't it morally irresponsible to use artificial intelligence in a lamp post? What if it becomes self-aware? Could you imagine life as a lamp post? A friend of mine was a spotlight operator for a nude cabaret in Vegas for a while, and even that was unbearable more than a few hours a night.
    • Jun 06 2006 | 9:01 am
      We'd obviously be morally obliged to provide some mechanism to allow it to deal with dogs.
      -- N.
      nick rothwell -- composition, systems, performance -- http://