Trouble Shooting Streams

I have been having some problems with streams while monitoring gate signals. They appear as audio signals, green yellow and red and fade slowly rather than dropping off immediately. This seems to happen more on the left channel than the right channel however it sometimes happens on the right channel. I tried calibrating streams and had no trouble on the right channel, however while trying to calibrate the left channel the oscillator wasn’t really coming through except for a few moments after which it stopped. After trying several times I finally was able to determine that the calibration procedure only seems to work properly if the vca curve is set to linear not exponential, once I took care of that everything seemed to be working fine. I am still having some problems with monitoring gates however it seems to be reduced and is only cropping up when I have very short attack times and exponential curve. I think it may have to do with noise on my system.

Summary:
1: Streams seem to only work properly when trying to calibrate if curve is set to linear not exponential. If this is correct adding this to the manual may save someone else quite a bit of time.
2: Streams seems to be displaying some inconsistency when determining if a gate is audio or cv this seems to pop up more on channel 1(left channel) with exponential curve and short attack times. My guess would be this is on my end with noise… I don’t know where the noise is coming from perhaps flying buses and my 4 ms psu, although I have streams connected into the first port on the flying bus.
3: Calibrating helped with the gate issues although did not totally resolve them, and streams seems to be functioning perfectly regardless.
4: sorry I keep having problems… (don’t fret, I’ve already found 2 bugs in Moogs Mother 32 so I guess i’ve got sort of a knack for it, mutable firmware seems more solid than Moog!)

1. I confirm that the calibration needs to be done with linear curve.

2. Streams use two pieces of monitoring code:

  • “Audio style” monitoring (rectification + low pass filtering, dB scale, green green orange red)
  • “CV style” monitoring (direct visualization of the voltage, green = positive, red = negative)

The criterion for determining which to use is the following:

  • If the zero crossing rate goes above 10 Hz, we switch to audio style monitoring
  • If the zero crossing rate goes below 5 Hz, we switch to CV style monitoring

Enough noise on the Gate input could indeed causes the detector to incorrectly switch to audio monitoring. It would be doable to add a noise gate here to make the detection more robust.

Thanks! I really just wanted to know about the linear curve, that took me a minute to figure out. I think that the monitoring on streams is what is creating at least some of the noise and luckily it is easy to not monitor with streams. It sounds great!