Recording vocals to be autotuned for effect

[thebigcheese]

Ok, so I'm not real big on hip-hop and rap myself, so I'm not really looking for a bunch of people telling me I just shouldn't do it. Having said that, as I understand it, T-Pain's vocal effect thing is accomplished with Auto-Tune. I use Reaper and I don't have Auto-Tune, but I do have a basic understanding of how to use it and Reaper comes with a similar plug-in. When I have used it, it has been to make minor corrections and thus has been relatively transparent in sound, but artists (if they are to be called that) like T-Pain make it very obvious that their voice has been "corrected." How do you go about doing that? Do they just speak for the recording and worry about the pitch later? As a musician, I feel like it would be hard to not already have pitch ideas in mind and to get the right kind of inflection for the mood I'm trying to set if I'm just speaking into the mic.

I know that a lot of people just think the effect should be done away with, but I like to consider all the musical options available to me and try out new ideas. In this case, however, I'm not really sure how to proceed, so I could use some input.

 

[d.bop]

Well, for one, T-Pain couldn't carry a tune if his life depended on it, which is what makes this effect actually work.

With REAPER's "ReaTune" you can.....sort of get this effect. It wasn't really built for extreme correction though. Lower the attack time to get closer to what you're looking for.


The other option is to download GVST's GSnap plugin.

This will give you MUCH MORE control and is pretty comparable to Auto-Tune (it even accepts midi input!)

1.) Set the key you want the part to "sing" in and check "Set threshold to fill gaps."
2.) Set the "Amount" to 100%.
3.) Turn the attack time all the way down and adjust the release for whatever sound you're going for.

That's basically it. Read the manual and mess around with the settings until you can get a desirable effect.

Hope that helps you out a bit!

edit
I've also seen people achieve this type of effect with a talkbox. Might be fun to try sending some keyboard or guitar into a talkbox and mouthing the words.
...haha I think I'm going to try that right now actually!

 

[thebigcheese]

I've got a talkbox, which I'm not particularly good at using, but I was just curious how it was actually done for T-Pain. So are you saying that he doesn't sing it, he just speaks it? That's the part that I'm having trouble figuring out. Granted, I haven't actually tried it...

 

[d.bop]

No, he sings. He just isn't very good at singing on key which makes autotune work harder than it normally would, giving him that effect.

 

[d.bop]

As long as you set the correct musical key and adjust attack and release times it probably won't matter how well you sing. The tool will, by nature, tune everything to the correct pitch automatically, regardless of whether it is actually on pitch or not.

I say just sit down and mess with it. Hit record and slide your voice up and down, do scales, sing a real song, anything really. Then start messing with settings and see where it gets you.

I really do recommend the GSnap plugin over ReaTune, though. It gives you a lot more control, and it's free!


I found this video, I think it might help: https://www.youtube.com/watch?v=7BWEpEvixME&feature=related


(Here's another clever use of auto-tune by the way: Auto-Tune the News)

 

[thebigcheese]

So, since I am fairly good at singing on pitch, I would have to purposely sing terribly to get it to do what I want? That sounds difficult...

 

[d.bop]

Not really. The best way to learn this, or any, effect is to just sit down and try it out. Play with settings, record demos and concept tracks and just get to know the tool.

 

[thebigcheese]

Right, I plan on doing that if I want to use it, but if I sing on pitch, wouldn't that make the auto-tune changes fairly transparent, thus not really having the same effect?

 

[d.bop]

Not necessarily. There are other things, like vibrato for example, in the human voice that can be manipulated via the software. You can program the pitch-corrector to completely remove vibrato and get rid of those very subtle pitch variations as you sing and get a perfectly "tuned" sound that no human voice could ever achieve. On the graph, you'll even see the "shape" of the original vocal is a little bouncy, and the shape of the new pitch-corrected vocal is perfectly straight.

Also, with the attack set to 0ms, the "snapping" effect actually won't be transparent. Attack time tells the corrector when to start working. So if you set it to zero, that means it will start correcting the pitch in that first millisecond that it "hears" a transient. The longer the attack time, the longer the pitch-corrector waits before it actually starts correcting. This will give you a more "transparent" effect because the listener will hear the real voice initially and usually not even notice that the vocals were pitch-corrected.