Writing and recording a song in 48 hours.

In the last 48 hours, an entire song came together, from draft to full demo mix

Usually, I work much slower and take a lot of time to sit with a song and let myself find the things that are good about it and develop them further until I end up with something refined that I feel is as good as it could possibly be. But not this time-

Last week I gave a breakdown of my process for recording and mixing my demos, and while much of that is true (those steps do usually happen in some form or another), it's also very true that I'm constantly trying to re-invent my process and shake things up. I think I'm the most excited and interesting creatively when I feel like I'm breaking my rule, or doing things wrong. I don't think of it as taking shortcuts, I think of it more like... cheating, lol.

Trying to get away with a process that I know is strange or flawed puts me in a position where I have to come up with interesting solutions to interesting problems. I think some amount of problem solving and discovering the sound of the song as it emerges is good, and makes the mix feel more flexible and alive.

Late Monday night, I was sitting at my piano and stumbled into a song that I thought had some potential. It came together quickly- my first pass at lyrics and melody worked out okay (but I tweaked some things later) and the structure was alright, not too crazy, but not too boring. I feel like the verse sections still were kinda weak, but the chorus presented some interesting development, so I decided to run with it.

The next day sitting at my desk at work I jotted down a midi outline of the song with beepbox (an online midi editor designed for quick sketches that runs entirely in your browser). I do this at some stage for most of my songs, just to better understand/ visualize things link structure and flow. Sometimes, I export these sketches and upload the midi data into my daw and let that serve as part of the skeleton. This whole process can be good for getting a project organized and identifying any areas where unexpected complications might exist (an odd measure or beat here or there that sounds perfectly natural while playing, but really throws things off when trying to put the song on a grid).

I also spent some time editing a voice memo recording I made of the song last night, and formed that to the grid in my daw as a scratch track. Along the way, I had a few ideas for percussion, and dropped some samples in at some key moments that really helped give the song some dimension.

I've been experimenting with incorporating more of this midi tracking into my mixes lately. I wanted to experiment with keeping things simple this time, and playing with a lot of layers that mostly all play the same thing in unison. I knew I could use the midi outline to control my synth and do a few takes of that when I got home later, so I went ahead and set up the midi outline. Because this song has some tempo and time signature changes, my midi export was not able to mirror the song exactly, and I had to do quite a bit of chopping, stretching, and restructuring after import. Normally, a lot of these sorts of kinks would work themselves out over time as I sit with and work on the song in the writing stage, but this song didn't get that.

I really love using beepbox to develop ideas for percussion and rhythm. Because beepbox is built around repetition (and my normal sample-dropping percussion process is so... not...) the ideas I tinker with here can sometimes provide a useful structural foil when they work their way into the mix later. I also love beepbox because the pallet of sounds in one channel is very consistent and limited, which allows me to focus on the beat in very broad terms- "does it go up/down, and how loud?" Any questions about voice and tone get put aside and answered later.

Until now, a huge problem for me has been how to make the beepbox percussion export control something that sounds good in the context of a more developed mix. The raw beepbox percussion noise I use (rendered as a wav file) often doesn't sit right in a bigger mix and doesn't hold up well to being tweaked with extensively because of its limited bit depth and sonic range. This time, however, I found a way to route part of the midi information exported by beepbox (essentially the note volume envelope) and connect it to a very noisy, full-sounding reverb plugin, TAL ii. This, combined with my current go-to midi voice bank, the Ramzoid 808 Cooker, allowed me to emulate something functionally similar to beepbox's noise-sample based percussion, that sounded satisfying and could sit nicely in a more developed mix.

Another challenge was to find a way to make a track of steady synth notes that runs through the whole song. Usually this is easy for my synth because everything is already lined up and plays through at a consistent tempo, but since I have a few sections that change the tempo and time signature, I needed to find a way to either map my tempo more accurately, so the actual DAW grid was 100% true to the song as it should be counted, or make a midi track that I could stretch and chop the same way I did the other midi tracks so it would stay consistent. I opted for the latter, and went back to beepbox and made a new track that was straight 8th notes, and ported that in, and stretch and chopped it in all the same places so it would line up. (This ended up being a very interesting move, and would set me up for interesting creative choices later.)

That night I got home and recorded a few isolated piano takes. As I was recording, I realized this song sounded better and more more interesting when I kept everything a bit more staccato. Spending a day with the mono-phonic midi tracks made me accustomed to hearing the song with no reverb or tail on these notes, and I was curious how everything would blend together if I kept this more neutral. As I started to add synths, the sound developed more and more, and this really felt like the right call.

When it came time to do the track that was straight 8th notes, I realized that my midi-track based solution was forcing not only timing, but also notes from my synth, so if I wanted to change notes, I would have to do it in some way other than inputs from the keyboard... I ended up muting the primary oscillator and slowing controlling the pitch of the second oscillator manually with a knob. This created lots of interesting in-between spaces, and let me control where, when, and how the pitch shifted from one note to the next. I also decided it might be interesting if the pitch only shifted very gradually and infrequently, and only goes upward, building energy through the song, rather than following the same contour as every other track (which, recall, are all playing a layer of the same thing). This track was interesting, but not amazing. I did a few takes to try to get whatever goodness I could out of the idea, but ended up taking some of the basic principals and applying them to a new track, where I did essentially the same thing, but played it live and controlled the pitches in a more traditional way. This gave me something that was both easier to work with, and more organic and natural feeling in its timing. Ultimately, I kept both, and layered them together, and this felt very right.

Then I dove into vocals. I didn't want to do the vocals "right" so I dragged my mic into the bathroom and did a few takes in there with all the hard echo-y surfaces. I just let myself have fun. I don't think the lyrics are perfect, but the basics are there and some of the moments felt strong enough to run with. After I did a few takes, making sure to at least get the timing and energy right at every part once or twice. I made a new track and recorded some "crazy" vocals, full of energy and drama. I layered this into the choruses and felt pretty satisfied with how much they lifted things up.

And that's where the song is now- I did a bit more mixing and balancing today. I don't think the song is perfect, and there's a lot more that could be done to develop and refine things. I feel like the song isn't fully realized yet, and if I'm being honest, I still think the verse need some work, but there are a lot of moments in this demo that I'm quite proud of too!

This was a fun little project, and if you want to listen to it, you can check out the full demo now on my Patreon! 😁