Is it ever a good idea to record the vocals first?

[Michael Hopcroft]

Since I normally perform a cappella and rarely play instruments (I can't play them at all well), it's natural that I will want to record my audio tracks first. But I get the impression that recording vocals first and building everything else around them does not usually work. Years ago, I was put in a position of trying to do some recording, and my producer had recorded backings and had me record one verse or part of a verse of my vocals at a time. Since that's not a very natural way to sing a song, it went nowhere, and the album I'd mapped out never came close to completion.

The way I sing more naturally, especially since I sing by myself most of the time, would be to take the whole song at once and sing a few takes of it, and then go from there. But I have never once heard of that actually working as a production technique. So what do people with producing experience normally tell artists in my position when they want to record?

[msmcleod]

The main reason for not recording the vocal first is that:

1. Without a rhythmic reference, most people's timing won't be consistent
2. Without a harmonic reference, most people's tuning won't be consistent

This is why most singer/songwriters record their vocals with an acoustic guitar or piano as backing, as it provides both a rhythmic and harmonic reference.

While you can use melodyne to fix any tuning inconsistencies, vocals tend to lack any meaningful transient information to extract a tempo map.

I'd at least recommend singing along to a drum beat or metronome.

[mettelus]

There is nothing wrong with recording a vocal/melody first, since that with give overall key and feel for the piece, and can often be the core/backbone of the final piece. That said (HUGE caution here), for the very same reason that Mark mentioned above, never (this is pretty much always a "never") try to "shoe horn" a vocal like this into a final work. Better to think of them as a "demo track." As a piece fleshes out you will tinker/revise/adapt that initial thought into its final product (as with all instruments involved). There is zero harm in using that initial vocal as a creative backbone, but realize there will come a time (often several times) you will need to remove the original from the project and record a better version.

That same issue exists with folks on other instruments... a "first pass"/improve/whatever isn't the final version. Repeition leads to refinement, so brace yourself for the moment you highlight that original clip and hit the "delete" key and replace it with something better. A lot of "one hit wonders" fall into this category... they literally played that hit hundreds of times before the final was released; but when asked to repeat that, the "refinement" stage goes away do to time constraints, and the results reflect that.

[SuperFreq]

Hey if you have a good voice and you're confident with it, then HELL YEA! I say record vocals first.

The song will be more organic.

It will be more organic, natural and expressive for exactly the reasons msmcloud said above, but I think that's a good thing. (Yes difficult to work with, but not impossible)

I refer you to the best songs in the history of songs.

When the Levee Breaks... basic tempo 72 bpm, but Bonzo varies it plus or minus almost 10 bpm. The effect of this on expression and urgency is incredible.

Varying your tuning, even drifting as far as changing the key. There's nothing wrong with that, as long as all the parts and instruments follow. I refer you to Ladysmith Black Mambazo (famous for singing backup on Paul Simon's "Graceland" album), the song "Beautiful Rain" starts in sorta E, a little sharp, gradually morphs up almost a half step to F. It just flows naturally, and if some producer tried to autotune it to a consistent key it would lose that human touch.... and isn't that what music desperately needs to get back to?

So I say GO FOR IT MAN