At the church where I mix I'm usually responsible just to get the sermon recorded and so I'm set up to do that pretty well. Beyond that I usually take a quick board mix so the band can hear themselves but it's not a great mix and it's really more for hearing musical issues than enjoying the performance. When Easter rolled around this year we decided to video the whole thing to post on the net. I'm kind of limited on my console (it's kind of strange to be frustrated with a Midas L3000) because I've got so many monitor mixes, so it's pretty much a matrix out or nothing for a recording.
That said, my system is also in mono so there's no panning and thirty some inputs in a mono mix starts to sound flat and squashy pretty fast. Now despite the Mythbusters proving the contrary the old saying, "you can't polish a turd" pretty much holds true. Please understand that this is in no way referring to the music, but to the method I had to use to record it. You put garbage in and the best you can do is try to clean it up and finish with clean garbage.
So here's the tricks I used. I started with my flat, uninteresting mix (which was for the room and not the recording so yes, the horns are down and the vocals are too hot) and normalized it so that my highest peak is at max level. Then I added some compression. Not a lot, because the console was squeezing things a bit already, but just to get the big hits under control a little.
Going a step further, I stopped right there and set up my bus compression on the two mix. It doesn't matter if you're mixing two channels or two hundred, if you start with your end of the line compression in place and mix into it, you'll know when you're overdoing it as you work, instead of finding out when you try to stick it on last thing. In most cases I like a multi-band compressor because it almost allows you to remix something once it's down to a stereo mix. You can take four bands and decide if you want the kick comping with the bass or the bass to compress with the guitars, isolate the vocals a little and manage the cymbals separately, all without having any one element "pump" the mix.
With my dynamics in place I went back to my single track and split it, that is I copied and pasted it. With two identical tracks I then panned them out hard left and right. This didn't do anything right away, although it does sound slightly different than hearing a mono track in two speakers. The kicker is adding just a smidgen, an atom of micro delay to one side. Humans understand the space around them by the delay in time of sounds reaching each ear. So you can head fake people with a little delay. I used to try this with delays in the ten to thirty millisecond range, but it would always wind up sounding slappy or phasey. With delays of less than one millisecond (I landed on .9ms in this example) you're more closely simulating the time it takes a sound to hit one ear and then travel by your head and hit the other one.
You have to experiment a little because what you'll get is some comb filtering where some frequencies interact and start to disappear and others add together and get stronger. Depending on the type of music and the environment you're trying to create you have to put up with some thinning of the sound. But if you listen to the examples below, I feel like the flat one is a little fatter sounding but boring and the spaced out one is a little thinner but more engaging. With that in place I spent a little time running the slider back and forth from wet to dry and once I was happy with the space I had created I went back to the faders and balanced things out again. The delay had skewed the sound way to one side so it took a little compensating with levels to pull the image back to center.
For a final touch I added just a touch of reverb, a room that's similar in dimensions to ours, but that doesn't sound as dead. You can hear it on the drums at the start of a track and it's a little weird, but this is a situation where quite a few compromises needed to be made to make an hour long viewing experience feel alive and not canned. Somethings sound a little thinner and the cymbals are sometimes a bit flange-y, but I chose a delay that left the vocals pretty solid. With a quick check in earbuds and headphones I was ready to print.
You can A/B them for yourself with the links below. I'm not saying I made a silk purse out of a sow's ear, but I feel like it was an improvement. It's one of those things where you show it to somebody and they start to open their mouth with a few criticisms, but then they ask what you did it on and their eyebrows go up.
Hit the comments or drop me a line on Facebook or Twitter and let me know what you think or what you would have done differently. Someday I'll have a digital console in my venue and be able to dial up custom recording mixes, or even take multitracks of everything. Until then it's all about post production
trickery er... sweetening.