Natalie is one of the available voices for Synthesizer V, a vocal synthesis piece of software. There are demos online about how it works. Essentially it's a mini - DAW, either standalone or 'nested' in another DAW (pro tools for instance) as a VST plug in. At the first instance, you have to input notes, either by playing them in via MIDI or by drawing them on the time line, then define what the words are (highlight them and enter in a dialogue box). That's it at a basic level. It just sings them. You can edit all sorts of qualities about the voice & notes (tonality, pitch bend and vibrato etc etc) either globally or dynamically via little graphical 'lanes'.
In my case, I just messed around until I got what I liked, I'd like to say there was some rhyme or reason, but it was just reacting and correcting in an iterative way.
The entire album came about from discovering these artificial voices. I came across the demo (thanks to Damon who found it) and thought wow, this is really close. What if I made the 'bugs' into features, and made it overtly an AI singing about being an AI. This was pretty much as Chat GPT 3.5 happened, and everybody was exploring its weirdness. That wrote half the lyrics for me.
RE DG (I haven't said this!) I understand it's nearly done and an announcement will be made soon, but this is all second hand from Damon. I've not been involved at all, DG's been working with Charlie Andrew as a producer who comes with his own engineer, and I'd said I mostly want to step away anyway (enough already!). I haven't heard any of it, but Damon says it's somewhat different, just by virtue of working with a different producer.
