I've been VJing professionally for ~12 years with some friends and we have always exclusively used Windows machines because our internal testing showed that Macs were rarely enough to handle high-quality multi-screen projections properly. For the same price as a top of the line Mac, we were able to build a PC with roughly 2x the specs (and always a better GPU). We did however rely on Macs for QT encoding (usually PhotoJPG for best quality/size ratio), because even our fastest Windows boxes couldn't keep up with our slowest Mac (our conspiracy theory is that Apple intentionally crippled QT encoding on Windows). Over the last couple years we have switched purely to content production so we let another external company do the projections + mapping and we don't care what they use because it's doing mostly just video playback.

Do you have a writeup on your content production pipeline? I'm terribly interested. I used to use VVVV for sound-reactive video that I would project at parties and small DJ shows around town, but then I spent 8 years living out of a suitcase and couldn't keep up with the hardware demands.

Edit: low res demo from 2008: https://vimeo.com/2577056

We started out in a similar way, doing parties and shows, barely surviving. Then HD came along when all our hardware was PAL (720x576) i.e. useless. We couldn't afford to buy replacements for our MX50 (never mind that such a beast didn't and probably still doesn't exist), EDIROL V-4s, reliable DVD-players, projectors, multi-channel MPEG decoder cards, etc. We have several rooms full of PAL hardware we will probably never touch again, and which nobody would probably buy except for nostalgic purposes.

I continually experimented with automating sound-reactive video, but it was unpredictable and difficult to get a pleasing, "fluid" result which didn't cause seizures, so we always did it by hand using hardware such as EDIROL V-4 or the MX50. It was also a lot more fun to punch buttons and slam sliders to the beat of the music ;)

Do you mean current or previous pipeline? Unfortunately most projects are somewhat unique and require at least some variation, if not a complete overhaul in workflow, depending on what the customer wants. I never got a chance to use VVVV in a commercial capacity, although there is an upcoming project where we are seriously considering it. Generally we used custom in-house software (e.g. for 360 projections using 2x 4-channel MPEG-2 decoder cards) or VJ software such as ArKaos and Resolume. Like I said before, these days it's purely content production, so the pipeline is relatively similar to a game studio. We still encode everything into various QT formats and we rent hardware for shows because as you said, it's hard to keep up with ever-changing hardware demands.

If you are interested in our projects, check out http://www.motionlab.at (go the old site for now, the new site is WIP).

Some of that gear may be worth more than you might expect.

The V4s still sell quite well for some reason, some of the other analog gear is still in demand for circuit bending.

If the projectors are decent they can still hire pretty well too.

The seizure-inducing aspects are my favorite part! One of the DJs once complained that every time the video came on people would stop dancing and stare at the screen with dazed expressions. :)

I'd probably take a look at Touch Designer or Unity for sound reactive stuff now.

Keijiro Takahashi at Unity Japan has a great github repository full of relevant libraries.

https://github.com/keijiro

Given that you're mostly doing synchronized video on the projectors, I'm curious: if you did this setup today, would you still attempt to use a multi-GPU setup? Or would you prefer something like a bunch of cheap video-decoding-capable chip PCs attached to each projector/display and then networked to a computer that 1. sent (MIDI?) control cues and 2. acted as a NAS for the devices to stream from?

We have worked with both, but I would vote for the former - there are less points of failure and you don't need an IT person to make sure every node is running smoothly. We built the latter using multiple 4-channel MPEG-2 streaming cards + MIDI racks and used it for all sorts of events (the biggest was 360-degree back in 2006: http://www.motionlab.at/projects/content-production/urban-ar...) but there were always problems and I often had to rip hardware/software apart to figure out what was wrong. That's why I ended up writing a minimalistic tool for controlling the cards, the vendor's included software wasn't meant for our type of work - thankfully there was a decent C++ API available. Workflow-wise it's also much easier/faster to render one big video than a bunch of smaller videos, not to mention you can check final output on one screen. We wired up 8 mini-screens in the rack to preview the streams, but they were quite small, making it difficult to determine if the footage was ok.

This is exactly where I'm stuck right now.

Using a hackintosh for mapping/vj'ing as the off the shelf hardware isn't at all practical and I want to keep using OSX only software.

Honestly it's a nightmare.

I want as many video outputs as possible, enough power to run multiple layers of video, be running game engines at the same time, transcode a video in a pinch and more.

Switching to windows is probably the only option going forward.