Notes about mixing: when do you start?
When it comes to mixing for electronic music there's kind of two big schools of thought: mix as you go vs mixing as a separate process after writing. I used to be big on mixing-as-you-go because I saw that as part of the sound design process, and sure it can definitely help locking down parts of the track to build from. It also helps with the vibe, as for example a well produced and polished drum track will be a lot more inspiring to jam and record to.
But, and here’s the big but, for the sake of not losing my mind I've stopped doing that almost completely. I was spending too much time deep inside plugins’ guis and that kept getting in the way of actually writing good music. I believe the amount of time you can work on a track before burning out is finite and I think I'd rather try to come up with a catchier melody than a more polished sound. Deep down I also believe anybody would choose to listen to a great track mixed badly than a shitty song with a well balanced mix.
Thinking in those terms (good song >>> good mix) might make for a nice and clean argument, but it’s too simplistic right? Why can’t one aspire to have both good writing and good mixing? Sure, why not, and also, If I may, those aren’t two complete separate entities, and they influence one another in many ways. When you’re working on your own music it’s often way easier to fix a problem by reworking the arrangement, or choosing a different synth patch rather than trying to square the circle with eqs and compressors without ever reaching complete satisfaction.
Let’s be clear about something, I’m not advocating slapdash mixing as a creative choice, I’m just saying that in my own experience when I pay more attention to the writing and the arrangement I can then get my mixes to sound better and projects take less time to sort themselves out and get to the finish line.
While mixing-as-you-go it's really easy to get into shaping every single layer to make it sound punchy and awesome in itself and lose track of the big picture. If you get to mixing only after the arrangement is finished you have a much clearer picture of what needs to be fixed, or improved, because all the elements of the song are already there and you're very familiar with them. Knowing what’s the loudest part of the track before setting the levels is super helpful. Sure you can eq, compress and distort your kick drum and bass synth into oblivion from the get go, but that just starts an arms race that ends up eating into the mix headroom, and there's little point in my opinion in spending an hour with the bass track soloed trying to make it sound HUGE only to have to dial it a down a couple of notches later because it's swallowing everything else.
The only concession I make now is setting sensible levels with the gain and the faders sooner rather than later. I've had the privilege of sitting in a lecture with Gareth Jones and one thing he said on mixing really resonated with me: you can get so much done with just the faders. Relative levels between the different instruments will shape the song in a major way. Is this a kick driven track? Is the vocal front and center and everything else supports it? Those are broad strokes that will help you define the idea of the track and guide you moving forward. Also in an absolute way, setting the levels will ensure a correct gainstaging, which in turn will prevent certain plugins from misbehaving and will safeguard enough headroom to take some creative decisions and make some boosts later on.
After playing with the faders a bit I'll just write and arrange the rest of the parts, listen to the whole thing and try to fix whatever problems might present themselves to me and then I'm done. Sure, easier said than done, but I think that this way of tackling the decisions helps me be more productive and backtracking less. Because yes, no matter how solid or reasoned your blueprint is, you’ll end up backtracking a lot, and reassessing your decisions and changing your mind as you reach the end of the mixing process. That’s absolutely fine, and part of the nature of the beast.
It's kind of a philosophical shift of perspective: i'm not necessarily thinking about trying to make stuff sound 'better', I'm just fixing problems. If something doesn't feel right I try to address it at the sound design stage (picking better samples, programming better synth sounds, whatever) or writing and performing better parts.
TLDR: I think you can kinda boil it all down to 'start with a good thing and try not to fuck it up too much'