I just released another single where I played the drums. This one’s a more complex drum part than the previous release, and took a longer time to figure out and practice. Unfortunately I didn’t get around to making a video of the recording this time, but part of the bare drum track can be heard on the Substack post I wrote about it: https://mothershout.substack.com/p/two-lane-highway.
A couple of things that QfG folks might find interesting about it:
I wrote the drum part so that it’s nearly all dominant-hand playing kick, snare and hi-hat. The non-dominant hand is adding accent hits on kick, hi-hat and other drums.
To fill out the sparse hi-hat part I put a 1/8 note delay on it, so it sounds faster than it is.
Once again, I credit @Robert_Mathijs with the tricks of counting out loud, and linking the rhythm to body movements to help timekeeping. The song’s quite fast (170bpm) and the drums needed to be solid, and I like how they turned out.
You can hear it via the link above, or It’s on Spotify and on YouTube.
Great song! It slaps and definitely grooves, so that’s wonderful! I said it before but I love the step-by-step breakdown on the blog. Before I read it I was already thinking “dire straits” and indeed, that was the intention (sort of) Looking forward to the next one!
But they’re not AI. They’re synthesized, and there’s a big difference. An AI voice would (probably) do all the work; picking the notes, how they’re sung, how they’re articulated, and so on. With a synthesized voice, I do all of that. The only small machine-learning part is that the synthesizer has learned how the human original singers transition between notes; elements of their style. But everything else about the performance is hand-crafted.
Okay, that’s a bit of a rant, but I do feel quite strongly about this!