I for One Welcome Our New Computer Overlords.

I was at NAMM (the big music industry tradeshow) at the Anaheim Convention Center last year, and I picked up a demo CD of Melodyne, a pitch- and time-shifting software package for editing audio tracks. It was pretty impressive, fixing wrong notes in a singer’s performance, building harmonies (very natural-sounding), and constructing arrangements from a single instrument. It was amazing stuff, much more expressive than the industry standard of a few years ago, Antares AutoTune (notorious for giving even crappy singers good pitch, and for creating the “Cher” effect — that robotic, stair-step kind of vocal).

Well, they’ve gone past that. As blogged about in Panopticist, Celemony, the Melodyne folks, have devised a way to take a polyphonic audio track (that is, one built from multiple simultaneous notes, like piano or guitar chords) and allow the user to edit individual notes. They call it Direct Note Access (DNA… Skynet, are you listening?). Make sure you check out the video on Panopticist…

This is a sea change in audio manipulation, people. This is huge—something that was, until recently, pretty much written off as theoretically impossible. Sure, you can pitch-correct a singer or a sax, but that’s simple monophony. But take a guitar chord, extract each voice in the chord, and alter those by time and pitch!?

So why am I upset? I’m both awed and a little nervous. I’m a musician (a bassist), and getting a part right is part of being a good musician. You practice and work on something, and then you record it.

With this package, you will theoretically be able to drop in a half-assed performance, and then fix it while the musician gets a sandwich. People complain about over-produced pop now… give it some time, kids. This will make tape-splicing drum parts look like fingerpainting.

But maybe that’s just me being shrill and alarmist. In reality, studio tricks have been around since the advent of the recording studio. Speeding up, slowing down, effects, tape-splicing (to build metronomic drum parts), and more (remember the outcry about synthesizers in the ’70s and samplers in the ’80s?). Look at the plotline within “Singing in the Rain”; film star Lina Lamont gets her terrible voice dubbed (”looped” in the parlance) by a better singer.

So this is just one more tool. It could mean better-sounding tracks, or it could mean that people with even less talent get recording deals.

Here’s an example of the TC Electronics Helicon unit that’s been around for a few years (generates real-time user-defined harmonies for live use) —pretty amazing.