I'm currently playing with making talking basses, as heard in various Dubstep music - a genre which I'm currently very curious about. I really like the bass-led style that characterises this genre of electronic music.
The fun part is that I don't have any of the Native Instruments synths, like MASSIVE or RAZOR, which everyone else uses for this type of thing. This means that none of the online tutorials are really that helpful.
However, I don't want to do this enough to spend a load of money on it. NI Komplete, their collection of synths and other plugins, seems to cost a fortune, or a small fortune for the cut down version, and the large scope of Komplete makes buying individual components for £100 each seem wasteful. Finally, only the really expensive version of Komplete includes RAZOR which has a whole load of exciting looking vowel related controls.
Thus conquering my gear lust, I started out by looking at the vocoder included with Logic. This is the EVOC 20 PolySynth, which is a basic soft synth with virtual analogue and FM modes, and a pair of "formant filterbanks" which are at the heart of the vocoder bit.
According to the EVOC 20 manual:
A vocoder analyzes and transfers the sonic character of the audio signal arriving at its analysis input to the synthesizer’s sound generators. The result of this process is heard at the output of the vocoder.
So, in effect it extracts some kind of essence from an audio recording, and superimposes it on the synthesiser notes. The "some kind of essence" here is something to do with formants. On which more later.
I created a software instrument track for the EVOC 20, and then created and recorded an audio track. In this case it was my voice saying "Listen to the bass" which sounded cool to me at the time, but now sounds like American political advice from The West Wing.
In order to get this piece of wisdom from its audio track into the EVOC 20, I selected the audio track as the side chain input in the drop-down at the top right.
Turns out, though, that the whole thing of using an audio track as the source for the EVOC 20 is a bad idea. This is because playing the EVOC 20 just outputs whatever is playing in the audio channel at the point when the note occurs. If the transport isn't moving, nothing. Repeated notes don't restart the phrase.
In order to get round this, I right clicked on the audio region and converted it into an ESX24 sample instrument. It took a bit of faffing about in ESX24's Edit screen to get the sample to play, and to stop ESX24 from pitching it up and down depending on the note. Finally, I routed the output of the ESX24 to a bus, and then used that as the side chain input to EVOC 20. Yay!
In the first part of this clip, you can hear "Listen to the Bass" vocoded via the EVOC 20. In the second part of the clip, the rather indistinct, but otherwise much cooler, "Okaaay" reply was done with Zebra 2. I'll write up how I did that one in the next post.