Pitch by Ratio

Cubase
21

Anyway, back to the subject of tuning by perfect ratios.

22

I’m just using this VST instrument example to demonstrate. Yes, as you observed, Cubase can only tune VSTs that specifically implement that function: that includes some Steinberg multitimbral synths, Some VSL libraries and a few others.

You really can’t hear it?

Here’s a demo:

Just Intonation (Hermode Active)

Equal temperament (Hermode not active)

A Cubase project
Hermode.zip (664.5 KB)

Anywho, As far as “tuning by ratios” goes, There has to be a note used as reference. Then the note to be tuned is calculated from that. The Hermode feature does this on the fly, it dynamically calculates the tuning, and which note to use as the reference during playback.

So, are you hoping to select a note and then apply the ratio, getting the result of a second pitch?

23

I think maybe you were typing this before reading my reply, but I can hear it. But it only works with some of Steinberg’s plugins, as I’ve mentioned.

I’m not sure what you mean by you need a reference note. How do you think the pitch shift works right now? You can pitch shift any audio event. You don’t have to specify a specific note. There is a visual where your ‘root’ note is C and you select another note on the keyboard, but that’s really only a visualization. The note “C” has nothing to do with anything in the audio you’re actually pitching, and your audio doesn’t have to contain C in order for it to work. It just a GUI for selecting a ratio, but the ratios it applies are obviously skewed to fit equal temperament. The only real difference between what exists now and the feature I want is that there should be an option to do exact ratios instead of skewing them to fit equal temperament.

But anyway, the reference note is built into the ratio. When we talk about pitch ratios, we’re referring to speed. In a 2:1 ratio, it means our reference note speed is 1 and we want the new speed to be 2. But to actually apply this to audio, we need to translate this to audio length. Basically, you’d multiply the length of the samples in your reference audio by inverse of the ratio.

Let’s say we have an audio clip 500,000 samples long. If we want to apply a 5:4 ratio, then we simple do:

500,000 * (4/5)

So, our result should be 400,000 samples long. Each sample would be 4/5 the length, and if we’re already at maximum sample rate, then basically you’d be deleting 1 out of every 5 samples to achieve this. I feel like this should be obvious.