I have a PD-extended sketch running on a RaspberryPi (RPI) that plays wav's from 32+ tabread4 array's. They can be selected via OSC messages sent over network. I noticed that every so often the audio would cutout. The following is a recording from PD on the RPI. The cutout is heard clearly at 22seconds and 50 to 54 seconds. This happens with both the internal sound card of the RPI and when using a USB soundcard.

For comparison here is a similar graph from running the same sketch on OSX. The waveforms differ because the tabread4 arrays played are a bit random. But the graph shows that the sway from the 0 middle line is much more on the RPi than OSX.

To attempt to resolve I have to be certain the sketch DAC is given a volume of around 0.5. But perhaps someone knows better what is going on and could advise me?

Sketch elements:

@cyphunk Hello..... first idea is that osx is coping (through some mechanism such as its auto leveller) better with the dc offset that you have (the wave drifting up and down),
Many of your samples seem to have a dc offset, and when they are added together the drift is (maybe) large enough to trouble the Pi??... as it will be trying to dump a higher current into your amplifier and maybe the power supply is dropping from 5V.........
I would start by adding a [hip~] before your [dac~] and see if that helps...... maybe [hip~ 40] or even higher, and maybe normalising your samples in an external editor (many allow you to correct for dc offset at the same time.
The hip might help because it will diminish frequencies below its cut-off frequency..... and dc has a frequency of 0Hz......
David.