Let´s talk about Reatune on vocals

[Erkan]

I've tried GTune but I'm no experienced user of autotune type of plugins so I thought it didn't work too great. I don't really know what to aim for with the parameters. When I see attack and release I instantly go into compression mode and start thinking I'm looking at a compressor haha =) So what are you supposed to "look" for in a persons voice? How do you determine if you need slow or fast attack/release for the pitch correction?

 

[LeSedna]

I guess attack should be the time it needs to go from doing no correction to doing full correction (or X% correction you want it to do) once it detects an incoming signal.
Release the contrary.

"attack" and "release" are not related to compressors. It's just concepts that are obviously used in compression.

 

[Erkan]

Yeah I know and understand that but what I was asking was "how" do you know how much attack you need for pitch correction? Is there any reason why you would want an attack higher than 0ms at all? If the singer was singing a bit flat or sharp on a note, the attack time would keep the flatness/sharpness for that amount of time before correcting it. Why would you want to have that flatness/sharpness there at all to begin with? Obviously there are some things with autotuning that I don't understand and I hope someone can clarify it for me.

To me it makes most sense to just go with as fast attack as possible since I would otherwise have the out of tune vocals at the beginning of each out of tune note.

 

[narcossintese]

On Melodyne I always use the default settings with 100% correction. Then I listen carefully to the results part by part catching those that do not sound right. Just moving up/down the pitch, changing some parameters or deleting the little blocks. On Reatune the "little blocks" aren´t automatically drawn on the waveforms, but you can still see on which note the wave is, so you just need to put the lines when you want another pitch. The workflow is pretty straight foward but I think it´s really hard to decide which algorithm/parameter is the right one for the task.

using elastique pro here, with the "keep formants for most pitches" or something

Do you chose the "elastique pro" for some reason or just "because it works"?

 

[Yhanthlei]

Do you chose the "elastique pro" for some reason or just "because it works"?

well, I use it because after a bit of fooling around with the different algorithms it seemed to be the most transparent...

To me it makes most sense to just go with as fast attack as possible since I would otherwise have the out of tune vocals at the beginning of each out of tune note.

Try that, instant T-Pain!
I usually start with the default setting (250ms), and if the vocals need it pull attack time down until I start to hear it actually working and artifacts and shit, then put it up a little again.
I really like it at high attack times though for vocals that are pretty good already, just to get that intonation on longer notes perfect. Just sounds better, but you can't really tell it's been tuned

 

[Fragle]

well, I use it because after a bit of fooling around with the different algorithms it seemed to be the most transparent...



Try that, instant T-Pain!
I usually start with the default setting (250ms), and if the vocals need it pull attack time down until I start to hear it actually working and artifacts and shit, then put it up a little again.
I really like it at high attack times though for vocals that are pretty good already, just to get that intonation on longer notes perfect. Just sounds better, but you can't really tell it's been tuned

this. faster attack = faster, but more obvious correction. really fast attack times also seem prone to false detection. it's starting to work before knowing where to go so to speak.
i think there's not set and forget setting with pitch correction. just go by ear, if there's a part that just NEEDS tuning, automate it by hand so it's back in shape. if you need constant pitch correction there's something wrong with the singer anyways ^^