So, feeling compelled to write something following a post on forum regarding whether or not analogue summing was important, I thought I would publish my response here as well:
Summing is a tricky one. Back in the days of DAWs with bad arithmetic, a lot of people thought that the DAW never got its summing right. These days, the maths that drive the DAW is really pretty good, and the sum output should be the arithmetically perfect mix of what went into it. So, as with the whole analogue thing, what people look for are the imperfections.
A lot of the interviews with producers in Sound on Sound mention analogue summing, saying things along the lines of “the Dangerous 2-Bus gave us a sound that summing in Pro Tools just couldn’t”, but as with all the discussions of DAW vs analogue summing I’ve read in the past, there is nothing scientific about these comments, everything is always subjective. It would be very nice for someone to conduct a proper double-blinded comparison with sufficient replication to try and get some statistical significance (or not).
So, with analogue summing, using a regular desk or summing mixer, you’re possibly looking to impart the sound of tubes, transformers etc, rather than getting pristine quality. Additionally, you also have to consider the fact that you can end up needing a lot of D/A conversion to sum outboard, and that gets expensive and, for many producers, may not even be possible. Let’s say that I have an Apogee DA16X, giving me 16 channels of superb DA conversion. Do my productions only have 16 tracks? No. So, therefore, I’m going to have to sum internally down to 16 tracks. And if I’m going to do that and be happy with the quality for that, why not just sum to a single stereo buss and then just send two channels out and back through something analogue?
That said, from a mixing perspective, there can be advantages to doing things that way, even if you go down to 8 channels. I generally all of my mixing in the box, all automation, effects etc. What I’ll then do, is run out my kick and bass as 2 mono channels and three stereo pairs of stems to my mixer or other outboard. I’ll zero everything, then mix those stems bringing up the faders and possibly adding a bit of eq and/or compression, then print that mix back as a final stereo mixdown. Then I’ll zero the controls, and do it again, and again, maybe over the course of a week. The differences in the final mix can be subtle, but sometimes one just sounds ‘better’. Yes, you can do that in your DAW, but sometimes breaking free of complete recall of every parameter can be a blessing, if that’s how you like to work of course…
If funds were unlimited, I’d probably mix or sum in the analogue domain, but would that be for sound quality? Probably not. However, as they are very tight, I think the money is best spent on good monitors and acoustic treatment: a mix is only as good as the weakest link in the chain.
A couple of articles you might want to check out:
Fynn's House and Audio Tech Notes 
Thanks for this great info.
Professional Audio Mixing and Mastering: http://CrystalMixing.com