Thom Stark's Recording Technology Page

Home Articles STARK REALITIES About This Site My PGP Public Key

After Hours Ordinary Hero A Season in Methven Our Host Send Me Mail

Home Articles STARK REALITIES About This Site My PGP Public Key

After Hours Ordinary Hero A Season in Methven Our Host Send Me Mail

Home Articles STARK REALITIES About This Site My PGP Public Key

After Hours Ordinary Hero A Season in Methven Our Host Send Me Mail

Home Articles STARK REALITIES About This Site My PGP Public Key

After Hours Ordinary Hero

This entire album was recorded in my guest bathroom.

That's right. In the bathroom.

It's a tiny space -- you can span the walls with your outstretched hands. I put a plank across the vanity to make room for the monitor, keyboard and mousepad of the genuine NoName™ Pentium Celeron 500MHz system I used to record all the tracks, and sat my vanilla Korg Triton sampling workstation on a cheapo keyboard stand in the bathtub, so I only had to turn around to reach the keys.

The computer has 256MB of 100MHz RAM and a 30GB IBM SuperIDE hard disk in it. It also has an Echo Digital Audio 16-bit Layla digital interface installed, along with an Intel Ethernet NIC. I turned off the built-in audio in the BIOS setup -- it's hard enough to get all this stuff to work without different pieces of hardware fighting for control of the audio -- and I turned off the system sounds, too, because they mess with Cakewalk's default recording resolution level and that's Big Trouble.

Speaking of Big Trouble, configuring MIDI was a nightmare -- it put the system into an unbootable state a dozen times or more, before I finally got it under control. (I'd still be fighting with it, were it not for WinRescue 98. Hooray for Ray Geide!) Based on that experience, I strongly recommend that, unless you're using an old, old, gameport-connected MIDI interface, you disable the Microsoft MIDI mapper, because it just gets in the way.

The system runs Cakewalk Pro Audio on Windows 98SE. (I started out with Cakewalk 7, but it doesn't understand 32-bit FATs, so I had to upgrade.) I also use Sonic Foundry's Sound Forge 4.5 for mastering.

Both programs can use the same DirectX plugins. I regularly use and recommend Sonic Foundry's Graphic Dynamics volume-control plugin to fine-tune track volumes (I don't have an outboard mixer and Cakewalk's virtual mixer doesn't respond in real-time very well, so I do most of my volume adjustments to the actual source tracks, rather than try to use console automation and ride the virtual faders with my mouse), as well as their Paragraphic EQ and Time Stretch/Compress plugins. I also liberally used DSP/FX's Chorus effect to fatten vocals and sparingly used their Aural Exciter plugin to add sparkle to some tracks. Likewise, I employed AIPL's Warmtone plugin to add warmth to some tracks -- mostly violin and guitar parts. Otherwise, I used Cakewalk's own plugins pretty much exclusively for flanging, reverb and echo on both instrumental and vocal tracks and their pitch shifter to enhance background vocals.

I use an Oktava MK-319 microphone -- it's got a very warm, natural sound -- running through a JoeMeek Systems VC6Q British Channel (which just provides phantom power) to record acoustic instruments and vocals. For my electric guitar parts, I run my Gibson Paul through a Line6 Pod -- the original one -- and record them direct. When recording, I usually don't use the Pod's effects, since I prefer not to add reverb, echo, flange and so on until I'm satisfied with the quality of the performance. By cotrast, Kurt used his DigiTech GeNetX to get that squonkin' tone on his Gibson Les Paul Special for his solo in "I Am a Desperate Man".

I recorded all the vocals and acoustic instruments in my office, across the hallway from the bathroom where I have the Triton and the computer workstation set up. The actual computer, the Layla's breakout box and the Roland DR-770 drum machine all sit in the hallway on the OTHER side of the bathroom (it has two doors -- one opening into each hallway), to provide some measure of isolation from the machine noise. (The Layla box buzzes something fierce and the fan in the PC is really loud, too.)

For the Triton-generated tracks, I usually turn off the particular sample's internal effects and, as with the acoustic instruments and electric guitars, record the keyboard takes dry and add reverb and other sweeteners later. Because I'm a terrible keyboard player, I have record the left- and right-hand parts separately -- and I have to do a zillion takes, even so, and usually do some note-nudging with Sound Forge to correct gross timing errors afer the fact, as well. But I don't use MIDI for anything but basic drum tracks, because it's too hard to make it sound "human" -- at least, it is for me. Your mileage may vary.

Over the course of the project, I evolved a system for recording drum tracks where I would record each voice on a separate stereo track -- a kick track, a snare track, a hi-hat track and so on -- then edit and balance them against each other, before I mixed them to a single, stereo drum track. I tend to do the fills and turnarounds "live" -- I play the parts on the DR-770's drum pads and record them in real time -- but I let MIDI handle keeping the kick, hi-hat, ride and snare on time. (I discovered that I had to turn off MIDI echo, or it'd occasionally "hiccough" and drop a beat -- so, nobody's perfect. Not even robots!)

At the mastering stage, I used Sound Forge to identify and reduce peak levels by hand. (Yes, I could have just let Sound Forge compress them on the fly while it was normalizing the stereo mix, but that tends to flatten the waveforms more than I think is acceptable. Doing it manually preserves the waveform -- all it costs is time.) I also used Sound Forge to eliminate pops and clicks caused by splicing partial takes together. (Luckily, Cakewalk lets you shell out to Sound Forge to edit individual tracks. That's something of an art form, though -- you just have to experiment with each track to find out where the glitch is, then marry waveforms to eliminate the edit point. That works best if you're dealing with a snippet no longer than a measure or two -- it's a lot easier to find the noise and, since you wind up shortening the track every time you perform that surgery, you don't throw the timing off as badly that way. So I try to make a clip as short as I can when I'm eliminating sonic "gravel".)

I'm really grateful for the way technology has freed me from the tyranny of the studio clock. If I'd've had to pay by the hour for the time I spent recording this project, I simply couldn't have done it. Likewise, if I'd've had to pay session musicians to play the string and horn parts, this CD would never have happened, because I just don't have that kind of money.

Digital technology has let me substitute sweat equity for cash.

The "good old days"? You can have 'em.

I prefer living right here -- in the future.

(Copyright© 2003 by Thom Stark--all rights reserved)