I'm having a bit a problem with pd. It runs fine most of the time, but every once in a while there are DIO errors. Changing the "delay" setting doesn't seem to do anything. The pd CPU load is at most 5% and 1.2% in jack Pressing the DIO errors button reveals "A/D/A" sync and "unknown"
The patch is only doing some logic and playing back wav files.
Any ideas where I can look for the culprit?
Hey, are there any test patches that run the gamut of processing complexity?
TIA
brad
Brad Fuller wrote:
I'm having a bit a problem with pd. It runs fine most of the time, but every once in a while there are DIO errors. Changing the "delay" setting doesn't seem to do anything. The pd CPU load is at most 5% and 1.2% in jack Pressing the DIO errors button reveals "A/D/A" sync and "unknown"
The patch is only doing some logic and playing back wav files.
Any ideas where I can look for the culprit?
Hey, are there any test patches that run the gamut of processing complexity?
Just an update to help clarify:
I restarted pd to use ALSA. After initial playing of the patch, the vast majority of the errors were:
data late (quite a bit of them) A/D/A sync
Later, I received mostly
DAC blocked
and a few:
ADC blocked
brad
I'm having a bit a problem with pd. It runs fine most of the time, but every once in a while there are DIO errors. Changing the "delay" setting doesn't seem to do anything. The pd CPU load is at most 5% and 1.2% in jack Pressing the DIO errors button reveals "A/D/A" sync and "unknown"
The patch is only doing some logic and playing back wav files.
first of all, changing the delay, doesn't have any effect if you are using jack ...
for the DIO errors ... pd's dsp scheduling is independent from the hardware, which introduces additional latency and sync problems aren't necessarily xrun problems ... for the developer branch, i changed the scheduler, so that the dsp is computed during the jack callback ... this reduces latency and might help in you case ... installing is easy ... get devel_0_39 from the cvs, scons and scons install ...
i hope this helps ...
cheers .... tim
Tim Blechmann wrote:
I'm having a bit a problem with pd. It runs fine most of the time, but every once in a while there are DIO errors. Changing the "delay" setting doesn't seem to do anything. The pd CPU load is at most 5% and 1.2% in jack Pressing the DIO errors button reveals "A/D/A" sync and "unknown"
The patch is only doing some logic and playing back wav files.
first of all, changing the delay, doesn't have any effect if you are using jack ...
for the DIO errors ... pd's dsp scheduling is independent from the hardware, which introduces additional latency and sync problems aren't necessarily xrun problems ... for the developer branch, i changed the scheduler, so that the dsp is computed during the jack callback ... this reduces latency and might help in you case ... installing is easy ... get devel_0_39 from the cvs, scons and scons install ...
i hope this helps ...
Thanks tim. ( I realize that the jack issue is separate, I just wanted replies to focus on pd and not jack -- (so they wouldn't say: "check jack!"))
It's kind of a pain to hear clicks. I've reduced them somewhat by starting pd with no audio capture. Don't know why that would be since I'm not using that facility.
Thanks tim. ( I realize that the jack issue is separate, I just wanted replies to focus on pd and not jack -- (so they wouldn't say: "check jack!"))
oh yes ... to change the audio buffer size ... check jack ;-)
It's kind of a pain to hear clicks. I've reduced them somewhat by starting pd with no audio capture. Don't know why that would be since I'm not using that facility.
well, i doubt this will have a big effect ... it might make sense to disable the audio input in jack, though ...
also, it might be interesting to see, what how the native alsa interface performs ... i've had some trouble with my hdsp ... having a feedback ... but either it was a problem of the specific machine or i was the only one ever trying alsa/hdsp/pd ... so you might give that a try, too ...
cheers ... tim
Tim Blechmann wrote:
also, it might be interesting to see, what how the native alsa interface performs ... i've had some trouble with my hdsp ... having a feedback ... but either it was a problem of the specific machine or i was the only one ever trying alsa/hdsp/pd ... so you might give that a try, too ...
see my original reply to myself.