corpus experts- midi input and tuning

[graphixsounds]

im trying to use corpus as an effect to play melodies- is it normal for it to be pretty out of tune? any way to reel that in?

[JoshG567]

Corpus is an effect, what are you feeding into it, a synth? Are you using the default or a preset? I ask the latter b/c maybe the LFO or Transpose is on.

[graphixsounds]

im feeding percussion sounds into it and then using the midi input played by a keyboard to control the pitch, however im noticing the pitch isnt exactly what it reads out- its slightly off. my point being why have a midi input if it doesnt actually pitch correctly to the note you're giving it?

[JoshG567]

Corpus won't alter the pitch of the incoming sound, we can only tune how its resonator responds to incoming sounds.

If one tunes the resonator to C while the note coming in is B, it's going to sound dissonant.

[graphixsounds]

got it but if its at full 100 percent wet shouldnt it pitch correctly no matter what sound sorce it is fed?

[yur2die4]

There are a lot of things that influence the pitch other than the pitch of the resonator itself.

In a lot of cases, the resonator is basically just a really tiny echo. You can get a similar effect by using Simple Delay and setting the delay time to 1 ms or 2 ms etc with very high feedback. You go from having an echo to having a resonant tone being generated.

The thing is that even if the echo is at the frequency of C1, a bass drum or snare could still have strong underlying tones that cause the overall combined frequency to sort of blur. Still, you'll likely 'resonate' the frequency you are looking for, it'll just also have that underlying tone too.

It is a lot like a guitar body. A guitar body might mostly resonate certain frequencies, and that isn't going to change because it is a solid object. But the strings will change in pitch. If you were to try to measure the tone of Just the body, you would still hear it blended with whatever tones were playing.

I guess a best testing method would be with staccato white noise and really harsh corpus settings. Even then, you have to make sure you don't have the pitch offset via different parameters.

[TomViolenz]

got it but if its at full 100 percent wet shouldnt it pitch correctly no matter what sound sorce it is fed?

I think this may be the misunderstanding here: The Midi side chain tab in corpus doesn't receive any sound source it just receives the Midi note information from another track so that you can have some melody on that and corpus on the affected track will move the pitch in tandem.

But it's just Midi, it's not analyzing any audio from the sidechain source track.

So if for an extreme example the source track is a DrumRack where the Midi note triggers a sample of which the pitch is totally unrelated to the Midi note, the Midi side chain info would be totally useless.

[graphixsounds]

ok got it- so what is the point of the midi side chain? if not to be in tune- whats the point?

[Angstrom]

Some of the models stay in tune (using midi side chain) but one of them is weirdly out of tune. I forget which one, the Tube perhaps. But I've used the midi input on the membrane to play tunes quite a lot. It's fun!

Make sure the inharmonic dial isn't adjusted from the centre obviously.

[TomViolenz]

ok got it- so what is the point of the midi side chain? if not to be in tune- whats the point?

Well the point is what you assumed, but since it can't really work as you expected it, it lends you a hand to work that way if you use it just right as far as it can.

So let's say you want to affect a kick sample this way and put corpus on the kick track.

You would tune your kick sample to a known pitch. Let's say C1.
Now you want to affect it with corpus such that a playing melody on another track applies the effect with the changing note pitches from there.
For that too, you would use a synth on that track that is tuned correctly so that the triggering notes correspond to the played pitch.
Now when you choose that track as the sidechain input
for corpus your kick starts to play that little melody too.

Got it?!

[yur2die4]

You've still never mentioned what kind of material you are affecting.

If it is a snare or sidestick, some hats, white noise, you might get some pretty clean results.

If it is a bass tone or synthesizer, you might need to reconsider the effect you're trying to accomplish??

1. What precisely are you using it on, and 2. What KIND of sound do you hope to accomplish as a result of resonant tuning?

[graphixsounds]

i think i got it- im feeding it all kinds of random sounds some pitched some not :/- would be cool if it worked in a way so that its in tune no matter what is sent- ableton's resonator works this way no?- just no midi input :/

[TomViolenz]

i think i got it- im feeding it all kinds of random sounds some pitched some not :/- would be cool if it worked in a way so that its in tune no matter what is sent- ableton's resonator works this way no?- just no midi input :/

I don't think it does.

It changes the pitches just like the different parameters determine, but it does not compute the source materials pitch. So if you choose a combination of input to resonator setting that is not in tune, the output won't be either.

Or am I mistaken on this?!

[Angstrom]

You can feed audio into it as an effect, then play the pitch of that effect by using a second midi channel, via the "midi side chain". The midi source needs to play notes, which corpus will then follow monophonically. The midi note value controls the pitch.

[tylenol]

i think i got it- im feeding it all kinds of random sounds some pitched some not :/- would be cool if it worked in a way so that its in tune no matter what is sent- ableton's resonator works this way no?- just no midi input :/

I think it might be helpful to understand better what a resonator is. The best metaphors are a plucked guitar string, or air blown over a tube of a certain length, where length of the string/tube correspond to the tuning -- you can't just blow air at any speed to get something to happen -- the resonator prefers a certain oscillations and your action (plucking or blowing air) supplies a starting point that can lead to some parts of those oscillations. Except, corpus has somewhat more involved resonators than a simple tube. The resonator is always in tune with the midi input, but the interaction of a pitched input sound with some of the resonator types is extremely complicated, because resonators are complicated.

For pitched sounds there are going to be interactions with the harmonics of the pitched sound and the frequencies at which the resonator oscillates, and if the input sound doesn't supply the resonator's fundamental (the note the resonator is tuned to), you may basically lose that harmonic (think of this as a bit like playing harmonics on a guitar string, i.e. when you physically damp the largest oscillation). This will result in a harmonic other than the resonator fundamental being the prominent one, and this can sound extremely weird. For example, if the resonator is tuned close to the input pitch but not identical (say a whole step off), the input pitch usually "wins", i.e. the resonator won't actually achieve stable resonance at the pitch it is tuned to. In some of my quick testing while writing this post I was playing a saw wave at 440Hz (A) into a corpus with a string resonator tuned to B (493.88Hz). Not only does the 440Hz harmonic make it through as the fundamental (494Hz is damped by resonance), but the input frequency does supply a harmonic of the B at a very high D (this should be 4698.63Hz I think?), which resonates becoming extremely prominent audibly, and sounding odd:



Another example, saw wave at 440Hz (A) into a corpus with a string resonator tuned to D (587.33Hz). It turns out that a saw wave (at least operator's 64 saw) with a fundamental at that frequency has harmonics at 1760 and 3520, which of course are also harmonics of A 440, and the resonator had its modes at these frequencies; 587.33Hz was damped, going to 0 after the attack, and the A did resonate but not as loudly as the, about -10dB relative to the loudest harmonic:



Finally, depending on the attack of the input sound and your settings corpus's resonators won't settle immediately, and so on the attack you'll hear more of both the input's fundamental and the resonance fundamental.

Summing up: when you provide a pitched sample with rich but specific and tuned harmonics to a complex resonator that prefers to resonate at other harmonics, the results are unpredictable at best, possibly (likely) losing your intended fundamental frequency as the fundamental. Your best results will come from a relatively full-spectrum, flat, noisy input that allows the resonator to resonate at the right harmonics relative to its own tuning (i.e. unpitched percussion samples). Or, tune the input sound as well as the resonator -- this should be doable pretty easily with sampler/simpler. I'd also encourage you to play around with spectrum a bit more to get a sense for what is happening. Btw Resonator is not really any different, but just has a much simpler resonator model.

Edit: one more comment. The denser the harmonics of the input, the less predictable it will be, as you'll probably get even more of a mix of harmonics from the input and the resonator's tuning.