Pitch-to-MiDI without Autotune

Continuing the discussion from Pitch accuracy for vocals to midi conversion when not in real time:

Is it not possible to translate actual incoming pitch to a target pitch plus pitch bend information? I don’t care about autotune; I just want a way to translate vocal pitch into a midi timbre. That is really the only reason that I bought the Dubler for — not autotune or beatboxing.

Hi Rick. Apologies - I’m not 100% sure I understand. If you could elaborate that would help! In the mean time here’s some information on Dubler features that might be helpful.

Pitch Bend - Dubler takes in audio and extracts the raw pitch information. Dubler then uses this pitch information to trigger the nearest MIDI note*. MIDI notes are limited to tones on the chromatic scale. If you want greater resolution of pitch you can use pitch bend in Dubler. [Control Settings]->[Pitch]->[Pitch Bend Range]. If you increase the pitch bend range from ‘0’ Dubler will still trigger the nearest MIDI notes but also continuously output corresponding pitch bend information relative to that note. E.g. A pitch bend range of 2 will allow you to bend up/down 2 semi-tones from the note initially triggered. For this to sound right it is important pitch bend range matches the pitch bend range of whatever synth you are controlling.

MIDI Timbre - Dubler also allow you to use vocal timbre to control MIDI CC values. These are the 3 dials “AAA”, “EEE”, “OOO”. If you sing “laaaaa” it will turn up the “AAA” dial and if you sing “Doooo” it will turn up the “OOO” dial. You can map these dials through MIDI mapping to whatever effects, controls, parameters etc you want in your DAW. The dial farthest right is envelope “ENV” this outputs a continuous measure of volume so you can do things like add swell to your synths.

Hope that helps, but if not please let me know and i’ll have another stab!
All the best, George

*Ok it’s not necessarily the nearest MIDI note - this does depend a little on your settings - i.e. if you have a key fix applied this will change things.

I second this. I really wish the software could make use of pitchbend info to match the actual notes being sung into the mic. It’s crucial for those that are working in microtonal music or to hit the in between notes so to say. Maybe making use of the MPE protocol in future updates would fix this.

Hi Bandenoire. It would be good to better understand what you’ve got in mind here, so we could spec it out. Are you after a microtonal system (discrete notes but at a greater resolution than western chromatic) or the ability to output any note on a continuous range? The second of these is already possible with Dubler using pitch bend. But I’m guessing you’re after the first?
As you rightly say this wouldn’t be possible with traditional MIDI which is tied to the western chromatic scale. MPE is a bit of a dead-end for us as MIDI 2.0 is going to be coming up in 2020 (which will almost certainly include microtonal notes).
I think this is different to what Rick was requesting though? Anyway, as a quick overview - this is how the system works at the moment:
[1] Voice in -> pitch analysis to identify frequency: This gives us a raw pitch. And it is pretty raw, even accomplished singers will constantly shift between flat and sharp of a specific note.
[2 A] Frequency -> Output MIDI Note Using a variety of factors we decide which MIDI note should be played.
[2 B] Frequency -> Pitch bend (If a user has pitch bend activated). Dubler will send out pitch bend data to the DAW based on the difference between the input pitch and the note being played by Dubler. The traditional MIDI protocol for this is a value between 0-16383 based on whatever pitch bend range is set in Dubler. I.e. If you have a pitch bend range of +/- 2 a pitch bend value of 8192 will result in zero pitch bend, a pitch bend value of 0 will bend the note down 2 semi tones and a pitch bend value of 16383 will bend it up 2 semi tones. E.g. If I have a bend range set to 2, start singing a C then vocally bend up to a C# the pitch bend number sent out will be 12287.
Two reasons this is important [1] the pitch bend range on your synth needs to match the pitch bend range on dubler. When MIDI 2,0 comes out we might be able to do something cleaver about this but for now you have to make sure it matches manually. In our above example if you had the synth set to a pitch bend range of +/- 4 then that 12287 you expect to come out as a C# would actually come out as a D. [2] Pitch bend is high resolution. So small variations in input pitch will make their way through to the synth. What we could maybe do (as an interim solution) is reduce the resolution of pitch bend so it is more stepped, we could even tie the steps to a microtonal scale. Anyway just some ideas!

Well, 2 years later, MIDI 2 still seems to be mostly vaporware, so maybe MPE is a good idea.

But now I am wondering — after also experimenting with Entonal Studio — if microtonal scales in the standard scala format might be able to be imported into Dubler 2. (I tried using the Dubler together with Entonal [also beta], but the result was chaotic, with unintended modulations, etc.) At any rate, being able to add to the scale library with a standard microtonal scale format — scala — seems like it could be very useful.

New to dubler 2 Studio. My main difficulty is to translate my melody into a relevant midi track; I mean that the pitch, with or without bend or sensitivity settings, does not match exactly what I sang. A suggestion for next upgrade of the software: have a live quantization feature ! Thanks.

1 Like