rob aylestone
Moderator
I downloaded a demo of synthesizer V studio 2 Pro and thought I'd see how simple it really was, rather than take the word of the youtubers and the manufacturer.
It's really intuitive. I found a box marked tempo, another gave me a click. Without even setting anything, it picked up my MIDI keyboard and played a tone for each note. I tried pressing record and played in a simple melody. On the screen is a piano roll type display of each note. each one says 'la' - and sure enough the voice selected in the box on the screen sings la - loads of adjustments for the type of singing brash and loud or soft or breathy - loads of choices on knobs. Then you double click on the 'la' and type in words for each note, and the damn thing just works. The basic, not editing, not clever state made an amazingly good job. Next thing is importing a music track. I grabbed one Cubase had produced, and set the tempo. I pressed play, and the track started and I played in the matching melody. Swapped the 'la' sound for lyrics and then there's a box to export the vocal track. Bringing that into cubase, and job done.
The demo didn't allow the export, but the demo lets you try out the male and female voices and I found one I liked, and bought a bundle with that voice. For what I do, I now have a singer on tap with very little work. You can of course edit to a crazy degree to change how complex phrases can be done - but worth checking youtube videos for some of the examples. I never thought I'd be into AI vocals, but I can see how useful it can be. I'm not certain people will spot it isn't real.
It's really intuitive. I found a box marked tempo, another gave me a click. Without even setting anything, it picked up my MIDI keyboard and played a tone for each note. I tried pressing record and played in a simple melody. On the screen is a piano roll type display of each note. each one says 'la' - and sure enough the voice selected in the box on the screen sings la - loads of adjustments for the type of singing brash and loud or soft or breathy - loads of choices on knobs. Then you double click on the 'la' and type in words for each note, and the damn thing just works. The basic, not editing, not clever state made an amazingly good job. Next thing is importing a music track. I grabbed one Cubase had produced, and set the tempo. I pressed play, and the track started and I played in the matching melody. Swapped the 'la' sound for lyrics and then there's a box to export the vocal track. Bringing that into cubase, and job done.
The demo didn't allow the export, but the demo lets you try out the male and female voices and I found one I liked, and bought a bundle with that voice. For what I do, I now have a singer on tap with very little work. You can of course edit to a crazy degree to change how complex phrases can be done - but worth checking youtube videos for some of the examples. I never thought I'd be into AI vocals, but I can see how useful it can be. I'm not certain people will spot it isn't real.
Sheesh. That is not the point. The point is it only takes a little more work and you won't be able to detect it.