Tricks for good voice separation?

I’m using the tracks for four individual (mostly mono) voicings, and am struggling a bit with getting good note separation. Generally speaking i tend to use the tracks from low notes to increasingly high register for each voice (kind of like SATB voicings).

My general approach is to have each voice interfere as little as possible with the others, without doing away with too much of the overtones and/or character of that particular voice. Since there’s only two outputs, i feel they leave a bit too little leeway for post processing. Instead, i’ve been trying to cut frequencies as much as i can within the box, using the filters.

With the more overtone rich patches, this becomes really challenging. This is FM after all. And adding any/several of the effects quickly turns into a mess to my ears.

Soooo… Any advice is appreciated. I feel like i’ve exhausted the possibilities of the dual filters, and i can only wish for individual outs for a future mkII model.

1 Like

The Digitone Keys has individual outs.

1 Like

Overbridge allows processing individual track sounds.

2 Likes

There’s a chance that the parametric EQ which Digitakt got in 1.3 would come to the Digitone.

Oh yes, let’s hope so

A little brain dump:

  • SATB don’t separate that much; choirs often aim to gel, to sing “as one”. So the metaphor breaks down quickly
  • it sounds like all your sounds are melodic. Do you need four melody/harmony lines?
  • play with volume: our hearing has separate detectors for quiet and loud at the same frequency, so you can hear a quiet sound amongst a loud sound better than you might expect
  • play with reverb: put some sounds out in space, some right against your face
  • change how the tonal structure of the sounds evolve through time: some sounds to have their overtones early and envelopes fast, others to start dark and bloom late
  • if you have a sound with really enticing or gnarly modulations or trails, have it play on its own, everyone else go quiet to highlight the flourish
  • it’s as much about musical effect as it is about frequencies
11 Likes

If your music allows use strong transients.

1 Like

Could you elaborate on this?

These are interesting advice, i’ll try them out!

If I remember correctly…

The little detector hairs in our cochlear are sensitive to both frequency and loudness… there’s multiple hairs for each frequency (band), some responding more to loud, some more to quiet. So, whilst it’s true that one sound can mask another if the frequencies match, there’s also some circumstances when you can hear a quiet sound amongst a loud one. For example, you’ll probably hear a quiet rhythmic figure under a louder but less transient one.

You’ll get more milage faster by dividing your sounds across the frequency range, but don’t be afraid to experiment with other strategies to arrangements and mixes.

1 Like

Also, don’t be afraid to mix yourself into a corner sometimes. Say you have a powerful melody sound that just sounds best eating up loads of the frequency range… let it. Accept the other instruments have less to work with. Done carefully, you can end up with more characterful mixes.

e.g. the drums on Jimi Hendrix “Are You Experienced” are really muted and dark, probably intentionally to leave space for his voice and guitar. People love it. Modern recording technique would probably give a different mix, but no-one minds the 60s sound.

2 Likes

Pan sounds left / right a tiny bit
Don’t overlap any sort of accent / high volume levels between the notes on each track

Use delay on one , reverb on another sound

Have any ‘instant’ sounds slight difference with different attacks , if they trigger on the same step at lest there’s a subtle difference in the transient.

Find a track you like and see what they did.

1 Like

This is good synth mix advice in general. We are all such modulation junkies…with good reason! But it’s easy to lose the effect or obscure the sound. Also, modulation on everything can be a mess sometimes.

1 Like