The Vocoder, was developed as a method to encode and encrypt a human voice for radio transmission, and decoded at the receiving point. Leave it to musicians to find a peacetime use for the technology. You are using the microphone to input all but pitch content... you provide the pitch from the notes of the synth. Yes, it can be described as a robot voice. The synth sound is the Carrier, the Microphone Input is the Modulator. And simliar to FM, you do not hear the Modulator directly, you hear its influence on the result of the Carrier.

The character of your voice will be lost, it is almost impossible to tell who is singing the Vocoder. Those things that make your voice recognizable from others are not involved in the Vocoder (encoded) signal. You for have the option of mixing in your own "Dry" signal with the Vocoded signal.

Perhaps what you are looking for is beyond the scope of the Vocoder. It's sound is fairly well known from Cylon robots on Battleship Galatica to the vocals on Daft Punk's Get Lucky, and scores of hits inbetween. The charm of the Vocoder is that it does sound robotic, and not human. It's the ultimate synthesized voice.

That said, quietly Yamaha has been working on "humanizing" scoring of vocals. Just FYI, I provide you a link to where things are after some 20 years of development... mostly sold in the domestic market, their have been third party Libraries made in English.

Yamaha Vocaloid

https://soundcloud.com/vocaloid_yamaha/sets/vocaloid4-library-cyber

A vocoder is not really meant to transform a voice into a different voice. It's meant to make synth sounds adopt some of the characteristics of the modulator (microphone=voice activated input) which end up sounding voice-like in quality. Usually, this is "robotic" and is exactly the desired effect. This is similar to talk-boxes on guitar. Only these actually pass the sound through your vocal tract (mouth) so the shape of what's in your mouth ends up effecting the sound. More of an analog method of achieving a similar result. Again - not meant to change your voice into another voice - but meant to make the guitar sound inherit voice-like qualities.

You may be able to combine your voice (dry signal) with the vododer output to try to alter your voice. Use of EQ can alter your voice slightly. I don't have my Montage setup right now - but perhaps you can alter the pitch of the vocoder (octave up/down).

Other misc ideas from the net:

http://www.musicradar.com/tuition/tech/15-ways-to-make-something-good-out-of-a-bad-vocal-635291
http://www.artsjournal.com/lies/2013/01/making-bad-singers-sound-good-at-the-touch-of-a-button/
https://www.gearslutz.com/board/newbie-audio-engineering-production-question-zone/826880-trouble-making-voice-sound-good-most-common-helpful-vocal-effects.html

... of course there are thousands more

Vocaloid! It even works in Cubase! Thanks!

Thanks for the links, J.

Michael Trigoboff wrote:

Vocaloid! It even works in Cubase! Thanks!

Yamaha has not actively promoted Vocaloid in all markets. But you can't watch any anime in Japan and across some parts of Asia where it is not used as the singing voices. Initially, they concentrated on Japanese, quite naturally, not just because it's their native tongue but there are a very finite number of consonant sounds to recreate (like 50 but don't quote me). The technology works with a sample library - actual vocal recordings of a male or female voice hitting each vowel sound through the voice range. Read through how it's done. You can literally play the melody, type in the words, and the Library Voice sings the part. You can add articulations like vibrato and pitch bends in a totally intuitive way.

At first you heard it and you would think, well, it was good enough for background vocals, but each version I hear it getting better and better, amazing technology. The question is who wants to have an automated singer? Well, as you have expressed, perhaps there is an audience. Certainly, it's a tool worthy of exploring.

When I reviewed the content of Vocaloid my main concern was with timbre customization. The inflections and other nuance perhaps you can "dial in" (program) to be unique - but the core voice will be something that others can copy unless the samples are easy enough to generate with your own hired vocalist to give you a custom Vocaloid "instance". Or perhaps other simulation (virtual vocal tract, etc) technology can be applied to the sample to generate an "infinite" number of variants from the original recording so this doesn't just create the same "set" of vocalists (or something very close to the same vocalists) on all tracks that use the technology.

With speech recognition getting better - perhaps vocal scratch track recognition can get to the point where you don't have to enter in either the lyrics for Vocaloid to sing or some of the inflections. Narrowing down the process to microscopic surgery rather than hammer and chisel work.

Still, an interesting path to arrive at a destination even with the state of Vocaloid today.