variaudio: can't stop it from pitch quantizing

I’ve got a sustained high recorder note. Just as an experiment, I went into the editor for this part, picked the variaudio tool, and pulled the right handle down so that the pitch line, instead of being horizontal, now slopes down at an angle.

But now when I use the the speaker tool to audition this in the editor, the pitch does indeed fall downward as the line shows, but it does so in a “quantized” way… sounds like it steps down chromatically.

In the editor inspector, the “Quantize” field shows as 0.

What do I need to do to get an actual continuous smooth downward slide?


By “right handle” do you mean the tilt handle? The transition sounds smooth here when I use the tilt handle. How many semi-tones are you dropping and how quickly?

Yeah, “tilt handle”. I did a little more testing. Used retrologue and generated a couple of bars of a sine wave at C4, C5, and C6. Then rendered that out to audio.

Edited the resulting audio part, used the Variaudio tool to tilt the right half of each of those three notes downward varying amounts.

The stepwise glitchiness is easier to hear on the C5 and C6, but on C4, with a smaller downward bend (a couple of semitones), you can hear a/some steps there too. With a steeper bend, C4 sounds pretty smooth. C5 and C6 always sound steppy though. It is not a smooth bend at all.

Just wanted to give this a bump… anyone hearing what I’m hearing in the above test? Is there anything that can be done about that, or is that just the way it is? Is it particularly aggravated by high notes (as seems to be in my case)? I like to make my tuning subtle; steppy correction kind of defeats that. Although I think the “steps” are finer-grained than simply chromatic. I think there are more than two quanta per whole step.

Could you answer the question of brycem? I mean the one about how many semi tones you are “tilting” in total?
I think this is important, since the function and algorithm (of this “tilt”) are not built for other than subtle corrections of one single note - the intention behind being to correct “drifts” that are typical intonation errors of singers. It is a kind of micro-tuning and the algorithm most likely will create artefacts.

BTW: Your thread title is misleading. Because there is a difference between the “Pitch correction” and the “micro adjustments” (like “tilt”, etc.).

Yikes, I just tried this with a sine wave at 440 hz [edit- was actually at 880 hz, got my octaves muddled :wink:], and now I can hear the glitchy pitch stepping, even on a subtle half semitone downwards tilt :open_mouth: The pitch doesn’t drop smoothly, it clunkily steps down. I’m sure it sounded smooth when I did it with a vocal… Gonna do some more testing to try to work out what’s happening.

I bounced the variaudio processed sine wave to a new track, then looked at it in a spectral editor. Lined up with the variaudio tilt ramp, one can clearly see the stepping in the frequency response. Variaudio curve is on the bottom half, spectral response on top:

Tested with 220 hz and the tilt sounds smooth! So the variaudio algorithm must be struggling with higher frequencies.

Variaudio tilt with 220 hz sine - no stepping, sounds smooth:

Thank you, it is now obvious that the algorithm for the “tilt” function (pitch changes in a single “note event”) is sensitive to both, waveform and frequency.

Brycem, great visual illustration of the problem.