This topic came up recently on Nick Galea's producer forum on facebook and it generated a lot of different responses.

I personally think that for producing electronic music, that you can make simple mix moves as you write your track,
but you should save the surgical work and the 'getting it big and loud as possible' stuff until the track is written. Same with
spatial design, buss compression and of course anything on the master, these should all come when the track is nearly 100%  written.

The following is an example of how i'd process an individual track once i'd written the part and figured out it's role in serving the song. Note: I don't start going through this process until i've got a few key elements in, usually kick, percussion and hats. These are balanced with faders only as drums are the most likely to be pre-processed and least in need of work.

Firstly, some initial parametric EQ to remove anything un-needed (eg: making room for
the kick and bass by hi-passing non-key elements with bass info, high shelving or high cutting background elements).

This is followed by some compression, using a compressor with either a specific vibe for that task, or a simple control set (2 to 5 knobs). Compressing only when necessary (a reference mix is handy to judge this).

Finally, a 'vibey', broad EQ such as V-EQ or Stillwell Vibe to shape the tone in an instinctive way. I do the first round of parametric EQ with the element in question solo'ed and the second with the mix all playing at once. Sometimes a saturator comes into play here as a form of EQ also.

I try to only spend about 2-3 mins max per-track on this whole sequence. Anything else I save for later. Reverb is applied within instruments (such as Sylenth1s internal reverb) if I feel it's a key part of their sound (usually on synths). Reverb for the purpose of  the mix such as room reverb on a send, i do later when the track has taken shape.

What are your theories on this? Should the mix process be entirely separate from the track creation or are the two processes intrinsically linked for you?

I agree with metta for the most part; I would say that there is nothing wrong with what I would refer to as a "feeling" mix. If something sounds excessively loud in comparison to something else, I find it a bit distracting and I'll just sort of bring the fader down to about where it sounds better, without concerning myself with it balancing perfectly or anything like that. I agree not to break down to analytics, but if you get a gut feeling about something that involves elements of sound design or mixing, I wouldn't discourage it.

It's good to separate your analytical and creative thought processes so you don't have to keep switching focus, but as you develop your skills in mixing and engineering you will incorporate a more creative and reactive approach to your entire project. One of the biggest lessons I've taken away from talking to my engineering teachers or watching videos about mixing is that the best mix is one where you are able to listen to and look at a sound, and have acquired enough experience to know what it needs without having to deliberate on the possibilities.

As you get used to how things go together in your mixing sessions, you will start to pick it out instinctively in the other sessions. Don't try to ignore a problem just because you're supposed to be wearing your composer hat at the moment - the distraction of noticing it might be worse than just tweaking it to your liking on the spot!

Well before I say anything: to each their own. If it works for you and you like the results, stick with whatever method you've been using!

As for me there are two things at play... first: I think the sonic quality of the mix during my writing process IMMENSELY influences what choices I end up making in the compositional phase. If everything feels really bright and big I might change my harmonic structure to match that 'vibe'. It's all about vibe to me and that means that I do about 75% of my mixing as I'm producing/composing the track. Very interlinked for me.

The second (and perhaps more important point to make) is that if you don't mix at all as you go, you might get used to the sound of your track and be afraid to make necessary changes because you've come to like the 'demo' mix. I've heard mixing engineers talk about this a lot. Artists give the engineer their rough mix but find themselves gravitating back to their original sound, even if the mixing engineer is making quality decisions to improve the mix. Point being, make sure to stay as objective as you can about your track and be sure to compare to reference tracks when in doubt.

I do a large amount of shaping of my sound prior to actually touching my mixer channel eq. This is because i separate the ideas from creating and mixing mode. It helps to keep that in perspective so i remained focused on a certain mindset. Now, that shouldn't be concrete, because as soon as you start adding samples or start playing a synth noise, is when you start mixing. Your input signal is probably more important than any thing you can do with an eq.

So, these two philosophies are really just one thing. Except for when you start applying eq that's last in the chain (my mixer channel eq), then, that's when you are starting the "final mixing" stage or what have you.

Often times, i'll find myself thinking of new layers when i enter that final mixing stage and i'll add them in and see if it works out. But basically i've crossed over into final mix when ever i feel to have enough musical material to have it be a song. So then the final mix becomes just the "finalization" phase as you iron out structure, Musical ideas, and mixing. If you have a powerful enough pc, you should be able to keep things as midi up until you actually render out your track for mastering.

I (at least when mixing electronic music) mix as I go. Separating the mix stems from when everything was recorded on to tape and then mixed on a console (but even back then people would EQ and compress .etc while tracking). But when working within a computer, with VSTs and such, it makes sense to just mix as you go along simply because you can. But to each his own.

Mixing is definitely a separate process for me. Of course, "mixing" happens while doing sound design stuff and writing, but I like to give myself a few days to focus on mixing 100% after the track is done.

I do most of the mixing as I go, then fix as much as I can when I get to that stage of production.

I've been thinking about it, and I honestly don't see why you can't do both - version it out! If you feel comfortable mixing your track while you produce, have a "progress_mix_1" version and then a "separated_mix_1" version. Clean up your mixdown in the in-progress version to where you like it, then next time you come back to the project you open up the separated mix. Reset all your faders and pan pots, and give yourself a shot at mixing from the ground up. Bounce out your two mixes and A/B them side by side. Go with the one that you like more.