Whew – as I write this, it’s about half way through what I hope to be my work day on Feb. 11th, and I’ve just put the final scratch track to bed. I’ve got bounces of all 10 tracks queued up in REAPER for some critical listening, to make some notes before I get into tracking “for real.” That leaves me with 17 days and change, to arrange the songs, record all the parts, do the mixes and put together a final master. I like those odds.
Today, I’m going to write about my approach for recording scratch takes, what I hope to do with them, and why I thought it was a good idea to spend the first third of the RPM Challenge recording parts that I more or less intend to abandon.
In the past, when I’ve recorded in full-band context, we’ve started with recording bed tracks as a full band – getting down the full arrangement of the songs, as we’ve rehearsed them, played live in a studio. All of that is intended to get polished up with overdubs and edits later, to whatever extent needed, but everything is building off a rehearsed and well-performed outline. Working solo, and intending to record something above and beyond a single voice and guitar, that’s not really an option for this project.
At the same time, starting with that solid performance is a huge boon, so there’s a lot of benefits to having a voice-and-guitar performance to work off of – so I started each song, or at least most of them, by laying down a voice-and-guitar performance.
Because I’m doing to be laying down all sorts of sequenced instruments on these songs, tracking the scratch takes needs to be done to a click track. However, I’m pretty god-awful at just playing to a click. While I attempted it for most of these songs, I find that there’s never quite enough information in a click to actually keep in time.
Instead, I found myself beginning by sequencing a simple drum beat, usually just kick, snare, and hi-hat, for each song. For some reason, even if it’s just grinding away quarter-notes, I find that a lot easier to actually lock in with than a click track doing the same thing, and it gets me thinking of the arrangement right off the hop. (As I wrote that, I realized that I can set up the metronome in REAPER, like in any DAW, to play kick and snare samples instead of just clicks. I’ll keep that in mind for next time.)
Once I had the drum beat more or less with the right feel, I just set it to loop for eternity and tried tracking a voice-and-guitar performance on top of it, without really sweating the finer technical points of recording. After several false starts to sort out the tempo and beat feeling, I was generally able to make it through a solid take – for most songs. A few of the songs on this record have a few twists and turns in the timing – either an odd number of bars here and there, or a few bars in 3/4 in the middle of a 2/4 or 4/4 piece – and by running through the song I was able to get all those changes mapped out in the REAPER project, then circle back and run a few takes for real.
After that, I would label the regions in the song, delineating all the verses and choruses, so I can navigate through the project more easily when I get into the real tracking.
So why go through all that effort, only to abandon what I’ve come up with? For better or worse, these songs aren’t all rehearsed or arranged well enough to have a reasonable shot at getting solid keeper takes right off the hop, so these performances are just scratch – sketches of where what the songs will look like and be structured, so I can then start building the music for real.
I actually found that to be really liberating, once I got down to it. Because these pieces are just building blocks, I can slice and dice them with impunity: while I want to avoid using things like auto-tuning and time-stretching on the final takes of “live” instruments, I used them heavily on the scratch takes to get rhythms and melodies just so without spending an eternity trying different takes. I was also free to do more drastic things, like slicing and dicing up performances to add space for instrumental verses, or changing tempos and keys without worrying about losing good takes.
For example, late last night I gave everything I had done to date, and realized that “Turpin Hero” was about 20% too slow, and probably in the wrong key – so this morning I just pitch-shifted everything down a fourth, and sped it up. That mangled the existing takes beyond all recognition, but proved it was a worthwhile change to make. Here’s a verse of the original, the pitch-shifted version, and my second take after I decided to keep the change:
That’s still not a ground-breaking performance, but it’s an improvement to my ear – and one I probably wouldn’t have felt free to make, or would have felt very discouraged by, if I was committed to keeping any of the scratch takes.
I used these techniques for some other things, as well, like pitch-shift and auto-tune to rough in harmonies that I intend to dub in later, but that’s a story for another post. For now, I really need to get back to these songs, and start tracking the “real” takes.