Inside the recording process

The vast majority of music that I hear is recorded, and if you’re reading this the same is probably true of you. Most people don’t have a clear idea what the recording process is like, especially using computers. Here are my adventures in recording.

I grew up in the eighties. Cassette recorders were just starting to be ordinary household gear. My sister and I made a bunch of random tapes as kids, not knowing what we were doing or why, just that it was fun. We also taped songs we liked off the radio. We waited until the song we wanted came on, and then held up the tape recorder to the radio speaker. Go ahead and laugh, millenials, but this was such a widespread practice among my generation that there’s a whole Facebook group devoted to it.

Recording to a single-track tape from a single mic was the only way to record music until 1955. In the single-track era, music was recorded more or less the same way it was performed for an audience. There was a single mic in the middle of the room, and everybody played into it simultaneously. The only “mixing” was done by placing quieter instruments closer to the mic and louder ones further away. Recording as an art form unto itself came into being with the invention of multitrack tape, which made it possible to record different sounds non-simultaneously.

Multitrack is an enormously big deal for recorded music. It enables you to capture ideal performances more easily, since you record each voice or instrument in isolation from the others. An error on one track can be fixed while leaving the others intact. Multitrack also opened the door for mixing, since you can manipulate the volume and tone of each sound independently of the others. This might not seem like such a big deal, but that’s because we’re all so used to spectacularly high-tech sculpting of sound. When I listen to old jazz records, the bass is a vague muffled presence buried in the murk of the low end. It took until the sixties for recording engineers to really figure out how to make the bass jump out of the speakers; now we take for granted that it’ll be as crisp and defined as any other sound.

Even with all the flexibility it offers, tape recording is still relatively unforgiving. I recorded a few songs on tape with my first band in college. Correcting mistakes was tedious and took considerable skill and timing on the engineer’s part.

Since 1997 or so, everything I’ve recorded has been on the computer. There are some pros and cons. The major con is sound quality. Tape is analog. The waveforms it captures are infinitely smooth and continuous. By converting the continuous electrical signal from the microphones or instruments into digital files, you necessarily sacrifice some signal quality.

So that’s the bad news. For me, and for most recording musicians at this point, the good news enormously outweighs the bad news. Digital recording is cheap and constantly getting cheaper. Good quality audio tape is expensive; hard drive space costs next to nothing. A computer costs a heck of a lot less than a decent tape recording console and you can use it for other purposes. But cost is only the tip of the iceberg. The really big deal with the computer is that it visualizes music, turning it into screen objects that you can drag, drop and otherwise manipulate the same way you’d manipulate words in a word processing document. For a visual thinker like me, this is a transformative and revelatory change. It’s radically easier to do complex edits on the computer screen than keeping track of a bunch of pieces of identical-looking tape.

The other big deal about digital audio is perfect copying fidelity and endless editing. Every time you copy a tape, the sound quality degrades a little. Also, as tape ages, it chemically degrades. Digital audio files are highly robust. They’re just long lists of numbers, so you can copy them flawlessly and endlessly across any data storage medium. You can edit digital audio non-destructively, so you can try out ideas to your heart’s content without ever harming or losing your original tracks. Digital audio is also nice and portable. You can lay down basic tracks in your basement, overdub more sounds in someone else’s bedroom and then mix and master in a million dollar studio. And while there’s no undo with tape overdubs, you can effectively undo anything you do on the computer.

Music is intellectually a lot easier than it looks. The big challenge for me, and for most would-be musicians I encounter, is anxiety. We have a crippling fear of being judged, and when we’re doing a recording, the panel of potential judges is enormous. Digital recording has done a lot to reduce my anxiety in front of the microphone. Knowing that nothing is carved in marble takes a lot of the pressure off. I’m much likelier to lay down a perfect take or a cool new idea if I’m feeling relaxed, and recording in my apartment on a computer is as relaxing as it gets.

I’ve been recording an acoustic singer-songwriter’s album for the past year. Aside from the vocals and guitar, everything on the album is fake: the bass, the drums, the percussion and keyboards. The vocals and guitar are processed using Auto-tune, digital EQ and reverb and compression, and various other tricks. The “performances” are stitched together from many different takes, with sections repeated and individual notes corrected for timing and volume and decay. None of these techniques are unusual in the age of computer recording. Some people feel that the computer is harming musicianship by making it so easy to sculpt a flawless performance. My feeling is that the computer just shifts the locus of creative work from the original performance to the editing process.

After doing enough of my own projects using the full digital toolkit, I started questioning the wisdom of recording instrumental performances at all, when it’s so much easier to use sampling and synthesis. The turning point came while working with a soul/R&B band called Love Child. The singer and I were writing and arranging songs using samples, drum machines and all the other hip-hop tools. We gave these tracks to the band to teach them the parts. I made charts too, but the tracks were better for conveying the vibe and nuance we were after. We had a bunch of ace musicians in the band, but they never sounded as good as our sample-based tracks. We’d meticulously sequence a bassline, and then the bassist wouldn’t play it exactly. He’d do variations and little improvs, the usual embellishments that musicians add almost unconsciously. The problem wasn’t his ideas, they were all good. The problem was that by straying away from the extremely sparse parts we were writing, he was deflating the tension, turning our hip-hop feel into a generic-sounding funk.

So it went with all the musicians. Also, it was a logistical nightmare getting everyone together, and it cost a fortune. Eventually we asked ourselves, why are we doing this? The songs sound better on the laptop, why don’t we just commit ourselves to life in electronic world? So we started doing gigs with just the laptop and singers, and it sounded terrific. I feel bad for contributing to the rapid drying up of gigs all musicians are facing in the computer era. But meanwhile, we were going for a sound, and the human beings weren’t giving it to us.

Samples and loops give you a lot of freedom. They also carry their own constraints. When you use, say, two bars of a Miles Davis tune in a particular scale with particular chords to a particular beat played on particular instruments, that forces you to fit the rest of your musical elements to fit. This constraint is a stupendously valuable songwriting tool. Repeating the loop identically is easy and varying it is hard. So by default, sample-based music uses a lot of repetition, and you have to justify each variation because it takes so much more effort than another copy and paste. You’d think this would be true with live musicians too, but it’s not. Getting a band to play a loop without variation is just about impossible. I’ve tried many times, everyone gets bored or feels the need to express themselves. We in the western musical tradition undervalue repetition, and having the computer encourage it has improved my writing and arranging enormously.

Sampling is such a useful framework for structuring musical ideas, now I take a sampling approach to live recordings of instruments whenever I can. If I’m doing a rock track with Barbara Singer, we’ll record a take of her flailing freely away at the guitar over a beat, and then find the best bar or two and loop them. If we need a variation or another section, we’ll use the second-best bar or two, and maybe the third. The less material we use, the better it sounds.

In the future I would wish for a more porous barrier between the recording artist and the listener. It’s been a bottomless source of pleasure for me to remix and mash up other people’s recordings. With all due respect to my fellow musicians, I know what I like better than they do. For the vast majority of recordings I have, I’d rather hear the key musical ideas repeated identically in groups of four or eight over hip-hop beats. If recording artists don’t want to oblige me by structuring stuff that way, I can just edit their music to suit myself. It would be a lot easier to do this if I had access to the individual tracks. A few, very few, artists release tracks with the stems separated out. I wish for the day when it’s standard practice.

Update: for hilarious insight into the process of making a top ten hit in 1988, don’t miss the KLF’s Manual.