Simulate tone of human voice with Syntakt?

In my explorations with Machines and LFOs and such I have occasionally made sounds that feel similar in tone to a human voice (fragments of speech or song)

I’m not interested in generating words or anything, just want to approach a point where a sound is similar enough to a human that you question it. These sounds are more noticeable with modulation so you hear an evolving sound.

I suspect FM synthesis is the path… do any of you folks know any theory or technique to apply here?

Good example of what I want here:

5 Likes

Have you check this out?

https://dood.al/pinktrombone/

9 Likes

I’m crying the sounds im making are so funny :joy: :joy: :rofl:

Now how do I control this with MIDI :face_with_monocle: :nerd_face:

1 Like

I haven’t tried on the ST but as a general rule you can modulate dual bandpass filters with high resonant peaks. You could try on an analog track and then route it through the fx track to get a second bandpass.

4 Likes

Excellent suggestion, thank you.

The technique is known as “formant synthesis”. Some devices have engines specifically for this task. Much useful information here (and in the whole series of articles of which this is just one part):

https://www.soundonsound.com/techniques/formant-synthesis

6 Likes

Experimentation is fun :slight_smile: I had a quick go with with SY Tone, a couple of SY chord and then tried the dual filter peak that @Jacob suggested. Also tried a little audio rate FM to mimic the YT example using the LFO, but it didn’t add much. I don’t think the LFO is quick enough.

I guess I didn’t get very close to the example, but with more experimentation you might get there.

30 Likes

The falsetto voice and then the last one, yaaii yaaaiii, those were really good.

3 Likes

Very cool, impressive results :astonished:

2 Likes

Haha awesome! It sounds a bit like the mellotron choir. Are you using one lfo to modulate the other? So that the tremollo can gradually slow down?

2 Likes

Yeah they’re all kind of quirky, in a fun way :smile:

The first ones are all SY Tone, the two falsettos that join the choir late are SY Chord, and the last '“yay” is two identical Sy Bits machines set to sawtooth and resonant low pass.

No, just a simple LFO to vibrato in most cases, with a little fade-in. One of the voices has a very mild audio rate LFO to amp, where I tried to dial in a frequency that played well with the basic patch.

I can share the patches here later tonight, so you can experiment further :slight_smile:

5 Likes

Yeah they are :smiley: Got it, thanks for the info. Ahh ofc, just using the fade! Need to experiment with it more. That’d be awesome :slight_smile:

2 Likes

This has more soul than many choral libraries! Fantastic work.

3 Likes

Reminded me a bit of that TE wooden choir :slight_smile:

3 Likes

Sysex file, one pattern with the voices in the audio file.

Cool if someone would develop them :slight_smile:

Voices.syx (35.4 KB)

8 Likes

Thank you, i will explore these. :cowboy_hat_face:

1 Like

This takes me back to the days when that singing monk VST (was it something like delay lama?) was all the rage - well done!

2 Likes

i’m frequently asking for the Vo-6 machine from the monomachine to make a return in syntakt or digitone 2 format , maybe one day elektron will do it, for this type of sounds it’s incredible.

2 Likes

Nils,
I can’t seem to load this (project?) file…
I transferred it through Elektron transfer and it doesn’t seem to show in project menu of Syntakt.

My bad :frowning: In my haste I forgot to mention that I could’t find an option to export only a pattern in Transfer, so I used C6 instead, which only exports raw sysex files.

Do you have a software that can import syx files? Eg Midi-ox, C6 or similar?

If not, does anyone here know if it’s possible to export single patterns+sounds from Transfer?

1 Like