The hardest vocal part in the world (the saga continues)
KeithHandy posted in Composing, Old skool, Producing on January 10th, 2007
So here’s my user-friendly (he says with a resigned sigh) road-map to the screenshot above. We’re somewhere in “week 2″ of recording four-part harmony vocals for This Is Your Chance. As I said earlier, it’s rather ornamental, non-repeating, dense, frilly, embellished. I still insist that at the core it’s not “complex”, but that’s a brawl for another pub. The bottom line is that it has to be done in tiny sections, tweaked, and edited as smoothly as possible. And I’ve been working on it for nearly two weeks now.
The top (red) track, “Guide (whole thing)”, is the whole song repeated four times, one for each voice, with a computer-sequenced speech synthesizer singing all the parts. It’s muted for the time being.
The next (purple) track, “Guide (what’s left)”, was an exact copy of the first track. Only, as you can see, most of it has been deleted by now. Each time I sing a bit and get it lined up and sounding decent and in-tune, I erase that part of this track so that I can easily see and hear what I still have left to sing.
The next track, “takes as they go”, is exactly what it says it is. This is my wide-open work area. Instead of singing along with the guide track, I find it a little easier to play a section of it and then sing it back; I can better hear what I’m doing that way. So I play a little bit of the guide, and then sing it. If I need to, I sing the bit several times until it “feels right”. I don’t have to line it up while I’m singing, because I can always drag it into place afterwards.
Even after I get a “good” take, though, I usually find I still need to edit it a little to make all the syllables line up right, and also want to slightly adjust the pitch of some of them. This is easy in Tracktion, because you can split a clip into smaller clips and type in a new “speed” value for any one of them. (Normally a clip’s speed is at 1.000; if you type in 1.01 the pitch goes up a percent, or if you type in .99 it goes down a percent — a semitone is about 6% lower or higher — I rarely have to adjust it by much more than 1 or 2%, which doesn’t noticeably impact the character of the voice.)
When I’m happy with how it sounds, I drag the edited clips (blue) down to the “accumulated good bits” track and cut that much more out of my guide track, again leaving the “takes as they go” track open for more raw recording.
The yellow track below that, “rough first take (test)”, is muted; it was just a rough run-through of all four voices without any attempt to perfect anything. As you can see it is 1/4 of the length of the other tracks, this is because I’ve dragged all four voices on top of each other to hear how they might sound together. It is quite out of tune, sloppy, and has a lot of random background noise. And it was important to do, because it gave me a vague idea of what I might have to look forward to if I put in the work.
The green track below that, “incomplete (test)” is something you can listen to right now if you didn’t listen to it last week. It’s a partial rendering of what I had so far of the “good” takes, again merging all four lines together, but with more than half the bits missing. It was good to be able to hear how things were progressing.
Now you can see how much more blue there is than purple … that shows how much of it I’ve done and how little I have left to do.
