The hardest vocal part in the world, post #3
KeithHandy posted in Composing, Old skool, Producing on September 4th, 2007
As I mentioned in a comment on another mortal’s site, I sometimes tend to “micro-blog”, meaning I write about the details of whatever I’m working on without always giving a clear overview of what I’m talking about.
And today’s post will be no exception!
So, back to tweaking and cleaning up what I’ve here referred to half-jokingly as “the hardest vocal part in the world”. So ya’s doesn’t have to go searching through my backposts for the details, the song is called This Is Your Chance, one of the songs left off the 1998 CD version of my rock opera for time considerations, and it was written in traditional four-part SATB format. That’s where “traditional” ends, though; had I handed it in to my college music theory professor, he would have drawn a truck driving through the gaps between the notes. (He always used to do that on the blackboard.)
I made up my own rules for how to arrange it. I figured since it would be sung over rock instrumentation anyway, it wouldn’t need to stand on its own, so if it had some questionable voicings in it, that would be fine. I wouldn’t have gotten as complicated as I did with it, except that I was working double shifts at a gas station in 1994, and needed a challenging project (besides counting packs of cigarettes) to keep me awake in the wee hours.
This song, or at least the vocal part, is a rare case where my prime motivation is “climb the mountain because it’s there”. Truthfully, though, it’s coming together nicely, and not as likely to be an acquired taste as I’d thought.
Here are my most recent issues with the vocals and what I’ve done about them:
1. External genetalia. SATB stands for “soprano, alto, tenor, bass”. Two of which I’m not qualified for. One more of which I’m only semi-qualified for. So some thirteenish years after writing out this arrangement, I decide that it can’t hurt anything to just try knocking the two higher lines down an octave, since like I said, the voicings were weird anyway, so it can’t make them any weirder. And a lot of it does sound just fine this way, but some of it is very “clustered” sounding, like if you played a keyboard with your fist.
Solution: I created a varying delay effect that was pitch-aware (requires Praat and does not work in real time, sorry peeps), so that I could make versions of the soprano and alto lines that were consistently delayed by exactly one half of a wave cycle. When mixed with the original part, it cancels out the fundamental pitch and all the odd harmonics, effectively making the voice sound an octave higher. By changing the volume level of this delayed sound, I could gradually shift the emphasis back and forth between the higher and lower (original) octave. As the soprano/alto voices went lower and got too close to the tenor/bass voices, I increased the effect — and as they went higher, I used it less, because there was already enough space between the voices (and the effect would have sounded ridiculous on those higher notes). I think this even fixes my amateurish voicings, but how anyone would ever perform it live is… not my problem.
(I say “not my problem” now, but just watch, in another ten years I’ll be working out a five or six part version for live performance, to emulate the recording I’m making now.)
2. Intonation. Hearing the voices sound the way they’re really going to sound makes it easier to pick out where the tuning issues still are. When I recorded them, I was extremely fussy and did a lot of tweaking by ear, but that was only with a horizontal (time) reference. Meaning, I was working on each line by its lonesome, tuning it to itself and not to the other voices. And for the most part, I did a decent job of this; but with aural fatigue, and perception naturally distorted by the repetition of listening to a phrase over and over, one is bound to be off here and there.
Solution: well, this is simple enough, because I now have all four lines synced together on four tracks. If a chord doesn’t sound like it’s quite hitting it, I mute different combinations of tracks to see which tracks do sound in tune with each other, and this makes it easy to zero in on the culprit. Isolating the offending syllable (snip snip) and shifting its pitch by 1% or 2% (less than a quarter tone, since a semitone is about 6%) in either direction usually is enough, and only takes a few seconds.
There aren’t that many occurences of noticeably wacky pitch anyway, and besides, some syllables — and this goes for any vocal performance — are more important than others. It’s best not to waste too much time on un-accented “in-between” notes, especially if they fly by so fast that you barely perceive them as pitched. Instead, it’s better to waste time writing a blog post about it. That is to say, if I hadn’t stopped to write this, I’d be done now. Curse you, internet!
