DEBRIS.COMgood for a laugh, or possibly an aneurysm

So, I spent last November recording drums for five songs, and then over Thanksgiving I engineered a two-day session in which my ex-bandmates overdubbed guitars — lots of guitars — bass, and vocals. I have not written about those sessions yet, but plan to, just as soon as I write about the 2nd half of my Utah parks trip (December 2002), the 2nd half of my Greek Islands trip (Sept. 2003), or the middle of my British Columbia trip (July 2004).

I won’t be mixing all these tunes, but I’ll definitely be mixing the drums for all of them. I started with the hardest song of all, my own tune Ode to Soup, because (as reported previously) it was recorded in sections. This means I had two tracks of snare drum, two of kick drum, two hats, four toms (two stereo pairs), and four overheads. Attempting to EQ, compress, and reverb all 14 tracks was killing my workstation’s CPU, because essentially I was trying to do twice as much work as I needed to. Two possible answers were evident:

• buy a dual-processor G5
• combine redundant drum tracks to halve the amount of processing required

Needless to say, I queued up store.apple.com right away… no, actually my G5 Upgrade fund has been appropriated and renamed, something about gradeschool tuition? Fortunately my workstation isn’t a complete dog. I recently upgraded the CPU to a 1 GHz model and added another 512 MB RAM. The CPU upgrade came with a tiny bumper sticker reading, “Don’t Laugh; It’s Paid For.”

Pictured is my ProTools session before combining the individual drum tracks. In most cases I simply solo’d each pair of tracks (e.g. the two snare tracks), set the faders to 0, and used Bounce to Disk to create and import the resulting mix.

In a few cases I applied gating and EQ as well, in an effort to reduce some of the processing I’d need to do later. The problem with applying filters in two passes like this is that fixing mistakes can be time consuming. I had to re-bounce two of the tracks multiple times when I later discovered that some of my filter settings were whacked. Given that I have not completed the final mixdown yet, I may again discover that some of my filter settings were whacked. The beauty of nondestructive editing is that I can always go back and redo this stuff, assuming of course I have the million years necessary to iteratively tweak every picosecond of audio in this song.

During playback I realized that I’d inadvertently achieved a different snare sound during the choruses, because I was playing more of a rimshot. As a result I re-mixed the snare channel with plug-in automation, enabling additional EQ during the choruses to reduce the nasty ringing sound the mic had picked up.

I did some editing too. I had captured pretty clean takes, but there’s always room for improvement. One downbeat in particular was killing me; it was a couple hundredths of a second late, and absolutely murdered the groove coming out of a fill. The repair was easier than I expected… and would have been even easier if I’d discovered the Nudge tool in Pro-Tools.

Then, left with just one set of drum tracks, I replicated the session setup advocated by David Franz in his DigiZine article on Phase-Coherent Drums. The final drum sound will consist of a dry submix, a heavily compressed submix, and a reverb track… 13 faders in all.

I need to play with EQ and compression levels still, but some of that has to wait until I’m ready to mix the bass and guitar. You can hear what I have thus far; the following clip contains an alternating 2-bar phrase to illustrate the change. I’m pleased with my progress, but I think I’m not done yet.

Ode To Soup Drum Excerpt (2 bars dry, 2 bars wet, repeat)

The ‘dry’ sections sound closed and boxy as compared to the ‘wet’ sections. The ghost notes in the dry sections are completely buried, in fact inaudible at low playback levels. The snare sound is much improved — check out how anemic it is, dry — but I’d like to get more of a crack out of it. I experimented with an additional snare track, created by running this track through a sound-replacer plug in (aptrigga, ~$46), but the results didn’t justify the cost. I’ll revisit that decision further in the mixdown process.

The next step is to extract a single guitar track from four takes plus an overdub (plus two harmonies and a powerchord track too!).