How do you melodyne vocals that are pretty in tune but are still not perfect?

[Karloff70]

Originally Posted by psycho_monkey

I would go back to the original untuned vocal and either use autotune in graphic mode (and just fix the notes that need fixing, although if you're doing tight pop/rnb that could be all of them!) or use melodyne and start again either fix the notes that need it or tune the lot by a percentage (and then fix the dodgy bits).

There ya go! Go back and draw lil red lines across the evil bits in graphic mode. Best results tend to come from mainly worrying about the entry and exit of words/lines. The middle best left alone most of the time. Unless its totally fecked of course...... Only way to use that dastardly tool anyway unless you like T-Pain.

[ddeez]

Originally Posted by Elk

The vocals here (both main and backing) are deliberately processed and auto-tuned.

Do you want "in-tune" or do you want the sound of the vocals?

Perhaps post a bit of the vocal you have and let people listen to it and provide suggestions.

are the vocals heavily autotuned? in what way are they processed? they just sound really good to me lol. i don't have the best engineers ear.

[psycho_monkey]

Originally Posted by quantumpsych

as far as i know, auto tuning a live performance with recalled parameters developed durring the tracking therof, is not practical enough to be widely implemented. the cost of doing this would be insurmountable for most bands, but certainly not for top-tier major acts.

this is obviously better than dubed lipsync but is it THAT much better that it should be acceptable among the performers? if they say no to their managers and producers, are they really that susceptable to being dropped? am i really that naive?

Actually it's very common that with a large scale production, a performer singing "live" to studio or concert audience is later re-tuned for broadcast. To take it to the extreme, a friend of mine worked on an S-Club live DVD a good few years back - his job was to replace all the main vocals with the parts from the record, and to tune the adlibs. now that's cheating!

That said, I saw Cheryl Cole perform on X-Factor a few weeks ago. She was either miming or had a very loud double track in the choruses, but the verse vocals were definitely live - weren't bad, but obviously nothing like the supertuned recording (it's just not possible to sing modern pop live IMO, even for a good singer). I think this was live to broadcast though, so no chances for technical magic there.

[Mr Funk]

Actually, that performance on X-Factor was delayed. It was performed in front of the audience, but it did not go out as it happened, so they could have easily fixed some dodgy bits.

With regards to those vocals you want to process/tune further. It is obvious that they have bee tuned already, but I think that is the sound you are after right? Non real, processed vocals? I think that style fits the track btw.

I however, being a singer hate the idea of being tuned. if I can't sing it, then i can't sing it! That being said, I think pulling in the odd dodgy note to save an otherwise good take is fine.

[theblue1]

Originally Posted by RockManDan

if your tuned notes are not sounding quite right with the track could it be that your instruments/backing are out of tune slightly. I had this problem once dealing with a vocal that i couldn't seem to get in tune even with autotune and turns out it was the gutiar that was in tune with itself just a little flat.

Although surprisingly few in the industry seem to grasp it, a good vocal that is intervallically in tune with itself will -- by the nature of the disjunct between the 12 Tone Equal Temperament system that most keyboards and guitars are tuned to and the mathematically correct, truly harmonious Just Intonation intervals -- be out of tune with such equally tempered backing instruments (although not necessarily out of tune with properly played brass, winds, and strings).

Check out the 'intonation gap' between the necessary compromises of 12TET and the harmonious intervals of Just Intonation:



Equal temperament - Wikipedia, the free encyclopedia


Of course, in an industry where big budget major product is often rife with obvious, annoying tuning wrenchmarks in otherwise 'natural' sounding vocals [not talking T-paining, here], I don't think we can expect much in the way of musical sensitivity or sophistication any time soon.

[Elk]

Given that you want a processed vocal sound, you should be able to get what you want by simply turning humanize to 0 or close to 0, and bring the retune speed up as high as you can without eliminating any slides or movement between notes that you would like to keep. Similarly, tighten the vibrato control so that whatever movement within the notes that occur that you do not like disappear. Play with each parameter so that you understand what it does and its impact on the sound.

You shouldn't have any problem getting the result you want with some patience and experimentation.

To speed up the process, try tuning a short (5-10 seconds) representative section of the vocals. Once you find settings that nail what you want, try it out on a full verse or chorus. Once you have this tweaked you probably will have what you want for the entire song, assuming that you want a more or less continuous sound. If this doesn't work - or if you want difference sections to be processed differently - do your work in chunks.

I think you are nicely on your way. It is now just a matter of experimenting and learning. It will take some time and patience. The next time however you will be able to do it a bit more quickly as you will have developed some skills.

[theblue1]

Originally Posted by ddeez

hey thanks very much for all the responses!! These aren't my vox these are a singers...i just wrote the song.

Here's the quality of vox i want...i hear these as being pretty much perfect.


YouTube - Chris Brown ft. Ester Dean - I Love You [No Tags]


RoyJeeBiv thanks for the great response much appreciated!

It probably would have helped up front if you said you were going for a T-Pain-Lite effect. People wasted a bit of time thinking you were trying to fix human-sounding, naturalistic vocals -- not super-effected robo stuff.

When I finally caught up with your clip (before I heard the 'model' one linked in the quote above), I thought I was listening to the wrong snippet, since it was clear you were going for a super artificial sound. (Maybe not as super-artificial as T-Pain, but, you know. Not human.)

[theblue1]

Originally Posted by psycho_monkey

Actually it's very common that with a large scale production, a performer singing "live" to studio or concert audience is later re-tuned for broadcast. To take it to the extreme, a friend of mine worked on an S-Club live DVD a good few years back - his job was to replace all the main vocals with the parts from the record, and to tune the adlibs. now that's cheating!

That said, I saw Cheryl Cole perform on X-Factor a few weeks ago. She was either miming or had a very loud double track in the choruses, but the verse vocals were definitely live - weren't bad, but obviously nothing like the supertuned recording (it's just not possible to sing modern pop live IMO, even for a good singer). I think this was live to broadcast though, so no chances for technical magic there.

Of course, the relatively new musical/episodic show Glee often adopts a robo-lite approach to many of the teen-oriented vocals, and I guess we have to accept that as understandable given the prevalence of over-the-top robo-tuning-as-effect in pop music.

But a few weeks back they had operatically trained and extremely proficient Kristin Chenoweth on (in a great part as an over-ripe white trash bimbo) and her songs were presented au natural -- for the most part. But in a couple of places in her 'big' show-stopping set piece number there were painfully obvious robo-tuning wrenchmarks. Right in the middle of otherwise apparently completely natural singing. It was like putting a big Jimmy Durante schozzola smack in the middle of Marilyn Monroe's face.

[ddeez]

Originally Posted by theblue1

It probably would have helped up front if you said you were going for a T-Pain-Lite effect. People wasted a bit of time thinking you were trying to fix human-sounding, naturalistic vocals -- not super-effected robo stuff.

When I finally caught up with your clip (before I heard the 'model' one linked in the quote above), I thought I was listening to the wrong snippet, since it was clear you were going for a super artificial sound. (Maybe not as super-artificial as T-Pain, but, you know. Not human.)

yeah im not really going for a Tpain effect... to my ears the example i posted isnt tpain sounding at all...it just sounds perfectly in key.


Elk Thanks for the great response. I really appreciate it.

[theblue1]

Originally Posted by ddeez

yeah im not really going for a Tpain effect... to my ears the example i posted isnt tpain sounding at all...it just sounds perfectly in key.


Elk Thanks for the great response. I really appreciate it.

By T-Pain-lite, I'm basically talking about a use where the use of auto-tuning is obvious and noticeable throughout the track (as opposed to something that apparently is supposed to sound like someone singing naturally but where one or more obvious retuning glitches make it unmistakable that the track's been doctored).

Obviously, a lot of stuff on contemporary pop and rock radio is aggressively processed for effect, with no bones made about it and no attempt to sound like natural, undoctored singing -- although still far short of the full on, rip-my-ears-off-my-head-and-throw-them-in-the-garbage-disposal T-Pain effect (although that Chris Brown/Esther Dean track is getting there in places.)

[Mark Kaufman]

In classic Gearslutz fashion, the answers are 10% addressing a way to do what the OP would like, and 90% criticizing the question.

[FeatheredSerpent]

Yes but he can't even hear that the vocals have been processed in the first example he posted up!

How is that even possible?

Unless you've only ever listened to that kind of nasty processed sound exclusively, as in 100% exclusively.


I think it's irresponsible to even provide suggestions for how to do it under these circumstances tbh.

[ddeez]

Originally Posted by Mark Kaufman

In classic Gearslutz fashion, the answers are 10% addressing a way to do what the OP would like, and 90% criticizing the question.

LOL! but you know what the one's who responded were very helpful and i'm very thankful for that.

[RoyJeeBiv]

Originally Posted by FeatheredSerpent

Yes but he can't even hear that the vocals have been processed in the first example he posted up!

How is that even possible?

Unless you've only ever listened to that kind of nasty processed sound exclusively, as in 100% exclusively.


I think it's irresponsible to even provide suggestions for how to do it under these circumstances tbh.

the more familiar you are with something, the more noticeable it can become. for example, when i first started mixing, i slathered reverb on everything, EQ'd tracks improperly, etc...only after i refined my mixing chops did i notice all of these 'mistakes'.

we work with audio every day, so these things can stick out to us -- so i don't fault someone for not being as aware about the procedures, because i know at one point i was (and i'm sure you were) there also. but hey, what do i know?? i'm 'irresponsible'...

[Dysanfel]

I listened to the clip. Look more carefully at the chord being played when she sings "...shine" and "...time". Seems to me there is something not quite right there.

[Elk]

Originally Posted by FeatheredSerpent

Yes but he can't even hear that the vocals have been processed in the first example he posted up!

How is that even possible?

Lots of exposure to processed vocals in many forms of currently popular music, lack of experience listening to artifacts, little experience tracking vocals with an excellent mic and clean chain, new to the craft, etc. All recording is artifice.

Most people don't hear MP3 artifacts, the artifacts in satellite radio, compression, limiting, etc. until it is pointed out to them. Once you know what to listen for it becomes obvious.

I think it's irresponsible to even provide suggestions for how to do it under these circumstances tbh.

Perhaps. But maybe the best approach is to encourage the OP to listen to what the plug-in is doing, to learn how what he does impacts the sound, to encourage making comparisons between his work and the good work of others, and to learn.

Most of our first recordings and first mixes were dreadful. Yet we learned from them and hopefully our recent output is better.

I work almost exclusively with classical music and musicians in real spaces. Yet I learn a great deal trying and experimenting with popular music techniques and equipment. It improves my ear. It also heightens my respect for those that record and mix good quality rock, pop and rap recordings. These engineers have amazing chops.

[Mark Kaufman]

Originally Posted by FeatheredSerpent

Yes but he can't even hear that the vocals have been processed in the first example he posted up!

How is that even possible?

Unless you've only ever listened to that kind of nasty processed sound exclusively, as in 100% exclusively.


I think it's irresponsible to even provide suggestions for how to do it under these circumstances tbh.

Keep in mind that a lot of people submit their music to publishers only to get returned for "pitchiness". The truth is, it's easy to hear "wildly out of pitch" but not so easy to hear "slightly out of pitch"...this makes his question very relevant. If you know it's a little off, but you're not entirely sure where, then what's the best way to use Melodyne to address the issue?

Whether or not to correct pitch always turns into some sort of ethical dispute, but as a business matter, it can be a make or break decision for a submission.

[theblue1]

Originally Posted by Mark Kaufman

In classic Gearslutz fashion, the answers are 10% addressing a way to do what the OP would like, and 90% criticizing the question.

Well... I didn't understand the question, since the OP seemed to be talking about naturalistic vocals with some transparent correction -- but was apparently actually talking about retuning-as-effect as evidenced in the example and his own clip. But he seems like a guy who's really open to learning, so, you know, it'll all work itself out.

[theblue1]

Originally Posted by Elk

Lots of exposure to processed vocals in many forms of currently popular music, lack of experience listening to artifacts, little experience tracking vocals with an excellent mic and clean chain, new to the craft, etc. All recording is artifice.

Most people don't hear MP3 artifacts, the artifacts in satellite radio, compression, limiting, etc. until it is pointed out to them. Once you know what to listen for it becomes obvious.



Perhaps. But maybe the best approach is to encourage the OP to listen to what the plug-in is doing, to learn how what he does impacts the sound, to encourage making comparisons between his work and the good work of others, and to learn.

Most of our first recordings and first mixes were dreadful. Yet we learned from them and hopefully our recent output is better.

I work almost exclusively with classical music and musicians in real spaces. Yet I learn a great deal trying and experimenting with popular music techniques and equipment. It improves my ear. It also heightens my respect for those that record and mix good quality rock, pop and rap recordings. These engineers have amazing chops.

[Bold added for clarity]

Great post. I think it's particularly worth noting the bolded part when considering that Elk works mostly in classical where the impression of hearing a very naturalistic, real performance is crucial.

Some classical recording is done in a straight documentary type fashion, but, even in the classical world, there is often editing and artifice.

That said, while I'm not a huge fan of classical and operatic vocalism, I'll personally come after the first SOB who robo-tunes my homegirl, Marilyn Horne.

[Tomer1]

I can see a situation forced upon you as a mixer, where you have no accual control over the recorded matirial handed over and you have to make it work,
But if your the producer and perhaps also the recorder I might assume why bring yourself into this situation in the first place?
If possible why not take the time and overdub passes which you dont like?

Just trying to understand the logic or couses to this...

[crypticglobe]

Originally Posted by ddeez

i have a song that i produced that i recorded the vox on and i autotuned the vox (in auto mode) and they sound good but they still dont sound perfect. So when i'm in Melodyne it's not like i need to change the note to an entirely different pitch but each note isn't perfect the whole way thru...so how do i get each note perfectly in tune? thanks a lot in advance! much appreciated.

With Cubase 5 varipitch.... to me it works better, is faster, and sounds better than anything out there. I just clip the segment I need to change and drag it to where I want it. If I want to flatten the pitch a little... drag the slider... done!