Tips for pitch correcting gruff/screamed vocals?

[blunderfonics]

Hey Everyone,

I tend to record a lot of bands on the loud/stonerly side of things and the singers usually have quite a bit of grind to their voices if they aren't outright screaming. Try though I might, its not always totally possible to keep redoing takes until they nail both the attitude and pitch as they tend to get blown out pretty easily and budgets aren't always huge. So when it comes down to it I'll usually err on the side of going with a take that has the right attitude knowing that the pitch is somewhat easier to deal with after the fact.

For the most part I use AutoTune Evo for tuning vocals when necessary and it can work fine on cleaner voices to the point. But with the grindier voices it seems that AutoTune has a hard time tracking the vocal correctly. There are times where it outright tracks the vocal in the wrong octave (with hilarious results on playback) or generally freaks out (less hilarious). In the cases where AutoTune is misbehaving I will usually try manually automating a pitch shifter plugin, but that is time consuming to get happening in a natural sounding way.

So has anybody else had to deal with this? Are there some different settings in AutoTune I can try?

I always work in graphical mode and have it 'make notes' but then go back and selectively edit the individual notes for duration and re-tune speed. Occasionally I'll switch to using curves if the particular part doesn't sound right using notes but that only seems to work occasionally so I just end up not using AutoTune on that section.

[honkyjonk]

Really? So the budget is or isn't paying for you to doodle with the plugin? I'm not much of an editor, but I'm imagining the argument would be that talented and efficient pro tools/autotune dudes could easily and quickly yield a usable vocal track with the right tips and tricks, and practice. I haven't had the patience for that though. I just keep finding out that doing another take or another take after that or another still is actually a quicker endeavor.

[kslight]

Hey Everyone,

I tend to record a lot of bands on the loud/stonerly side of things and the singers usually have quite a bit of grind to their voices if they aren't outright screaming. Try though I might, its not always totally possible to keep redoing takes until they nail both the attitude and pitch as they tend to get blown out pretty easily and budgets aren't always huge. So when it comes down to it I'll usually err on the side of going with a take that has the right attitude knowing that the pitch is somewhat easier to deal with after the fact.

For the most part I use AutoTune Evo for tuning vocals when necessary and it can work fine on cleaner voices to the point. But with the grindier voices it seems that AutoTune has a hard time tracking the vocal correctly. There are times where it outright tracks the vocal in the wrong octave (with hilarious results on playback) or generally freaks out (less hilarious). In the cases where AutoTune is misbehaving I will usually try manually automating a pitch shifter plugin, but that is time consuming to get happening in a natural sounding way.

So has anybody else had to deal with this? Are there some different settings in AutoTune I can try?

I always work in graphical mode and have it 'make notes' but then go back and selectively edit the individual notes for duration and re-tune speed. Occasionally I'll switch to using curves if the particular part doesn't sound right using notes but that only seems to work occasionally so I just end up not using AutoTune on that section.

If Autotune is tracking in the wrong octave?then manually bring it to where you think it is. There are presets in Autotune for specific ranges for it to track within also, perhaps you are not using those?

Honestly I'm not an Autotune kind of guy anyway, but in certain genres I think attitude > pitch more often than not, especially as much as those production styles rely on effects and/or burying the vocal in the mix. If the band isn't complaining about the pitch, it might be one of those things to let go. Playlist all your takes and if you have to go back and swap a few lines out then that's pretty normal for music of any genre to make a good vocal comp from several takes if they can't keep attitude and pitch going for the whole song.

[Nick Sevilla]

Stitching together a take first is best, followed by mangling the pitch with autofart.

Sometimes you can find a single syllable in another word, and stitch that in.

Good luck in your endeavors.

[blunderfonics]

Yeah, maybe I should clarify things a bit. I am comping the vocals first for the best performances and am not afraid to steal words, syllables or phonemes from another take or section if that is what works. In an ideal world that would be enough, but occasionally it isn't. And with these types of voices I find that there is a somewhat smaller window between where they are warmed up and mentally in the zone with the song and when they are blown out or ready to move on. So in those cases I turn to the tools at hand. I can tune a vocal fairly quickly once its comped and I can do it with headphones on the train or on the couch at home, so no it's not really eating into their budget for me to doodle with the plugin.

I do try messing with the 'Input Type' to see if AT is happier tracking in a certain range, but without much luck. And when it does track in the wrong octave dragging the notes or curves to the right octave either causes the whole vocal to shift to the wrong octave or introduces artifacts. It's very situation dependent.

[dfuruta]

Do this singers care whether it's in tune or not? Being out of tune seems like part of the charm of the style, to me...

[kslight]

Yeah, maybe I should clarify things a bit. I am comping the vocals first for the best performances and am not afraid to steal words, syllables or phonemes from another take or section if that is what works. In an ideal world that would be enough, but occasionally it isn't. And with these types of voices I find that there is a somewhat smaller window between where they are warmed up and mentally in the zone with the song and when they are blown out or ready to move on. So in those cases I turn to the tools at hand. I can tune a vocal fairly quickly once its comped and I can do it with headphones on the train or on the couch at home, so no it's not really eating into their budget for me to doodle with the plugin.

I do try messing with the 'Input Type' to see if AT is happier tracking in a certain range, but without much luck. And when it does track in the wrong octave dragging the notes or curves to the right octave either causes the whole vocal to shift to the wrong octave or introduces artifacts. It's very situation dependent.

Well the singer being blown out is something totally different and shouldn't even be accepted in the recording process. The irony is that most of the metal bands I work with are flat broke and want to make an album in like 6 hours (3 days, tops) but we all know that most of these guys even on a good day cannot pull that and sound good take after take. Vocal tuning is never a concern (unless they are actually singing?) so eh I just kind of think maybe you are trying to fix something that isn't broken. I don't touch Autotune unless:

1. The out of pitch part sounds worse than the Autotuned result.
2. The singer is not available, not practical, or not capable of redo-ing part correctly.

[losthighway]

I usually stick to the classic stoner metal approach to the use of autotune: none.

"Sing it a gain mcgruff, try to relax more on 'Wizard'. It's sharp. That's a tough vowel sound but you'll get it."

[Bro Shark]

As someone guilty of employing said vocal styles, I want to stress that the imperative strategy - while this may seem counter-intuitive - is to stay relaxed and go easy. Even hold back a bit, especially in early takes. Ease into the zone.

My voice can last four to six hours, at which point the body will be fatigued and there's no point going further. When I'm done recording vocals I feel like I've been punched in the gut repeatedly, but there is no throat damage/failure. The power comes from the gut.

Also, fuck autotune.

[Injured Ear]

Melodyne will track gruffer vocals better. It's a bit fiddly to get used to, but in the end you will get usable results where Autotune won't grab onto a pitch.

[Orpheus]

As both a producer and consumer of such styles I think that the best way to record it was figured out 30 years ago....reverb and delay. The contemporary slick sounding "death" metal is crap.

[Bro Shark]

As both a producer and consumer of such styles I think that the best way to record it was figured out 30 years ago....reverb and delay. The contemporary slick sounding "death" metal is crap.

The contemporary slick sounding anything is crap. There are a ton of great bands today doing it the right way though. No need to pay attention to the goofy shit. Here are some modern goodies.

Necrovation (SE)
Obliteration (NO)
Dephosphorus (GR)
Diskord (NO)
Artificial Brain (NY)
Gorguts (CA)
Incantation (NY)
Disma (NJ)

etc.

[GussyLoveridge]

I usually stick to the classic stoner metal approach to the use of autotune: none.

"Sing it a gain mcgruff, try to relax more on 'Wizard'. It's sharp. That's a tough vowel sound but you'll get it."

yis b'y.

[hunterchristy]

i would employ the following before resorting to auto-tune..which i dont even have.

1. get him to do it again. I record a vocalist who usually does one or two songs, really well, a day. I've worked with him multiple times, and we've mutually found this to be the best way to do it. I just realized that, even though it's a hassle to wait on the singer, I spend less time comping / editing / massaging tracks. I bet, really, if you factored the time you spent fighting autotune into the budget, you'd be saving them money by having them redo some stuff.

2. double that shiz: if it works for ozzy, right? RIGHT? haha. if you can roughly comp together a "best" take and a "second best" take, sometimes just playing two together, and the way the two takes phase or modulate with one another, can help a less than stellar vocal sound better.

3. mr roboto this is a weird one, and usually only works for more metally stuff, where maybe you'd apply a little gain to the track anyway, but blend the live vocal with a vocoder track--like a noisey sawtooth.

[casey campbell]

this situation is where melodyne really shines. the ability to snap everything to a key, and then undo certain wavs within that take...or manually adjust them in such a precise way is really awesome for screamo dudes.

i know antares can do it too, but i find it easier in melodyne.