imitone community

Building a universal, human-friendly music interface. ...But first we have to fix the bugs!

You are not logged in.


Having trouble with imitone? Hit send on any Error Reports it creates, or use the in-app "send feedback" button.

#1 2020-01-23 15:49:06

Registered: 2020-01-23

idea to have an imitone finally in version 1.0


It's been a while since I decided to trust the promises made by imitone, but it's now 2020 and the program still seems to be at the same point.
Moreover it seems that there is now a serious competitor named Vochlea Dubler, who has new ideas like turning a beat box into a midi battery, for exemple.
But Dubler works only with a microphone provided and it puts it at an unattractive cost for my taste (especially since I already have a good microphone).

So here I am again, and Dubler inspired me some ideas for improvements for imitone.

The big problem of imitone for me is that, it's too sensitive and picks up too many unwanted notes in piano mode, certainly because a singer will tend to make legatos between notes.
Perhaps imitone would need an attack-recognition system to have only the desired notes.
Imagine a singer singing "LAAA LAAA LAAA" and imitone recognizes the "L" and the "A" as two different information.
Now let's say that imitone considers the "A" as notes and remains locked on regardless of the singer's tonal variations, and that the "L" is considered as interrupters to properly differentiate one note from another.
In addition it would be cool to have the tonal variations inside a note transformed into pitch bend to simulate the vibrato and mic volume to expression controller.

I think that would be an idea to try, what do you think?

Best regards.


#2 2020-01-25 16:02:38

From: Seattle, WA
Registered: 2014-05-23

Re: idea to have an imitone finally in version 1.0

Hello, Lychee —

We've been watching Vochlea since it was a university project.

Insightful suggestions!  I already have plans for imitone's fifth stage of research, which revolve around key, scale, tuning and diction.  By "diction", I mean the way consonant and vowel sounds in singing voice imply divisions between notes.  For example, when someone sings "haaasssss", we hear a tone in the "s" where there is none.  When they sing "Fallen", we perceive two syllables even though the tone is unbroken...  This type of work will make imitone even more specialized for human voice, of course.

Even before diving into Stage 5 there are some opportunities to improve imitone.  When the experimental 0.10 technology is switched on, you'll see some improvements in attack note tracking and fewer reverb-related artifacts.  I've already developed new approaches to vibrato and pitch guide which should help with "slippy notes", but there wasn't enough time to get them into 0.10.0.  As a result our pitch-to-note transcriber is largely unchanged from 0.9.2d.

Regarding pitch bend, it has been on my mind for a while to add some new options for vibrato control.  Our team has discussed a pitchbend-based vibrato which would be based on an LFO.  I remember thinking about an "exact" vibrato option but ended up focusing on other things for 0.10.  It creates some problems to think about (separating vibrato from slides) but could be really cool!


Board footer

Powered by FluxBB