Zipper noise on channel and master faders

By “zipper” I do mean non-smooth change in gain: increases and decreases in gain in jumps. When these jumps happen quickly and this discontiguous gain change is applied to a pure sine tone , the spectrogram shows them as added frequencies.

Ok, the spectrograms for both the files you posted look very similar, though not identical. Not what I expected. Thank you for helping, though.

I can now reproduce this on a mac, an entirely different computer!

To easily witness the behavior on mac, first set the audio hardware for built-in, then set the buffer size to maximum (2048 samples).

Then open the attached project. The first set of automation targets the channel fader (volume). The second set targets the gain parameter of the TestGenerator. They are indeed the same automation data for both sets (cut-n-pasted).

The first set occurs as stepped. The second is smooth.

stepped_volume.cpr (124.4 KB)

Looks like 12 fixes this (hopefully?):

“Sample accurate volume automation: Volume automation is now near sample accurate and completely independent from the buffer size.”

1 Like

This is indeed improved in Cubase 12. The fader automation seems to be free of artifacts (although “live” fader changes are still audibly quantized.)