Vocal Overdubbing Techniques By George Massenburg

73*

Listening is Paramount
For those of you who don't know who George Massenburg is, google him or check the sticky post about Legendary Producers/Engineers at the top of the forum. The man is nothing short of a genius.

The following is taken from the GML website (http://www.massenburg.com/cgi-bin/ml/vocal_overdubbing.html)

Extraordinarily kind of Linda to mention our work together in the December 2000 issue of Mix Magazine, but perhaps too generous. We've learned a lot together. I'm sober enough to recall the many things that I've screwed up; it serves to eliminate head-swelling.

I've overdubbed vocals on band tracks for as long as I remember. In the 60's we were overdubbing by adding vocals machine-to-machine (playing back on one, recording on the other) and assembling by hand cutting 1/4". As soon as we had more than one track available (the advent of 3 and 4 track machines) we started punching in, and the fun began. After several years of 'punching' lead vocals, it dawned on us that there was more to making 'great' vocals than we had thought.

If I had to point to one person who had the greatest influence on me - to try to do it better - that person would be Lowell George. He was never satisfied with the status quo and avoided the quick and easy answer where possible. Thus, Lowell and I wanted to try some new ideas for the lead overdubs on "Thanks, I'll Eat It Here".

The first time I had ever heard of 'combining' many vocals to a single track (as contrasted with live switching, say, a couple of tracks), it was with reference to the record that Jackson Browne was working on at about that same time (was it 1976?) Jackson (an obsessive genius and a great lyrical songwriter) was doing more 'whole' passes, and punching in corrections less and less. His theory, and what was becoming very clear to us, was that a vocal track at it's best could be a complete entity with an integral musical story with it's own ebb and flow and dynamics.

Listening carefully to what happens when an isolated word or phrase is casually dropped in, we often notice that for whatever improvement is made in pitch (or phrasing, or interpretation) we take a 'hit' on the reading...very often we are distracted or pulled away from the story in some subtle way. We also find that when we take longer pieces that there is more of a performance in general from the artist and less concentration on, for instance, that one word, and more of a song.

Thus, we overdub vocals in complete takes as a rule.

In general Linda will start with a tune as follows. First, before she sings it, she will immerse herself in the tune. If it's an original, she'll live with the rough, or demo (she's been known to play it day and night for several days); if it's a cover, she'll listen to as many versions of the tune as we can find, and choose several from among them to listen to repeatedly.

Then we start singing. Generally, we'll put down groups of up to six tracks in a row to eventually combine to a single track, going over tracks if she's still obviously warming up, getting the most flattering headphone balance (too loud and the singer tends to sing flat and ahead of the beat, too soft and the singer will tend to be sharp; do a truly ****ty headphone mix and I'll get an Indian wrist burn).

Listening to these tracks is collaborative. We'll usually listen to several whole passes and pick the best, most complete take. Then we'll listen to each section (verses, b-sections, choruses, bridges, outro's) a track at a time to get some idea of what's happening on each track; I'll take very detailed notes on each track; often I'll have notes on many individual words. I'll try to listen with both brains - that is, I'll keep track of both the 'feel' (how the track 'works') and of the technical performance.

I have a system of notation that's loosely based on chess notation. I'll keep track of words and phrases of questionable pitch with up arrows and down arrows as near as possible to the spot that is out of tune (again, if you listen *really* hard, you'll often hear just one spot with bad intonation). Same with phrasing...I'll notate a vocal leaning ahead of the pocket or sitting back on it with left and right arrows.

A word about vocal pitch here, as we (hopefully) edge into an era of AutoTune backlash. I think we've noticed that the ability to perfectly tune vocals has unfortunately led to it's extraordinary overuse. Clearly, it can easily result in a 'vanilla', soulless performance - that in fact there's alot of "story" in a singer's inflection and intonation. Linda and I'll tune vocals only with reluctance, and never, never "AutoTune". We find that often a phrase or line with just one word or one syllable out of tune is mistakenly perceived as an entire out-of-tune line, and if we go in and just get that one spot the whole line improves.

Same idea with overdoing phrasing correction, although we don't have AutoPhrase yet (thank heavens). If just a word or phrase is out of the pocket, we'll nudge it...again, reluctantly.

So, back to the notation system. Acceptable lines are at least noted with a dot. The good lines are noted with single exclamation points. Fantastic turns of phrase get double exclamation points. Suspect spots get a question mark or two. Really screwed up spots get x's. Should something go by that I wouldn't want anyone else to hear (perhaps embarrassing to the artist), I'll put a tiny swastika under a spot; should that spot be so egregiously bad that it should never be heard again by humans, perhaps it'll get the rare double-swastika. Other marks indicate mismatches in levels (a circle with an arrow) and wacky textures (wavy lines). Should the artist exclaim that she loves something or hates something, I'll note that impression with her initials alongside. In the end, I'll have more or less of a map to follow, or at least to talk about.

So after listening to a section we'll stitch a comp together (most of the time I'm doing it D-D to either 3348 or ProTools); we'll listen to the whole comp, then perhaps listen to the comp against our favorite un-comped track. Maybe we'll refer to the demo or rough or original and compare where we're going with that. If we have the right direction, we'll carry on to the end. If not we'll go home and try it again another day.

If we choose to carry on we'll sing another three to six passes. By working together on the comp we've both learned a little more about the tune - these new passes may expand on the performance, or maybe not. We'll listen to the new tracks against the comp, dialing in spots from the new tracks when it makes sense. We don't really know until we finish out the new singing passes and the re-comp that we're getting somewhere. On occasion, we'll have one transcendent take, but more often we won't even know it until we've done a couple of comps - we'll find that we keep prefering that track...that track that's proving to be unimprovable.

Sometimes we'll do this several days in a row. We'll know we're done when we're clearly not improving the track anymore. On a particularly difficult tune like "Wish I Had A River I Could Skate Away On" it's taken us four years and three big tries at arranging and cutting a track to arrive at something that was even remotely singable.

Sometimes we'll do this in several different keys.

Sometimes we'll do this and throw the whole tune away.

[SIZE=+1]George [/SIZE]
 
Back
Top