5.1 recording with a Williams star array

I’m becoming interested in surround-sound (5.1) recording. For me, this doesn’t mean upmixing a multi-track recording and encoding as 5.1, but capturing the real ambiance of the environment in true 5.1.

There are several methods of using multiple microphones in array to record in surround, and they are discussed in reasonable detail at DPA’s Mic University. The Fukada Tree looks like a good design, but the array is three metres from front to back, so for occasional use it is going to be a pain to assemble.

Then I found this paper on Multichannel Microphone Array Design by Michael Williams and Guillaume Le Dû which describes several different configurations using cardioid microphones. All of the variations are more compact than the ones described by DPA. I picked the layout shown in Figure 9 to try, since the front-left and front-right pair almost resemble the ORTF stereo layout and it should be possible to extract a reasonable stereo recording from the same array.

Williams microphone array
Williams microphone array

A while back I laid out this design of Williams array using multiple mic stands and checking the alignment with a ruler and a protractor. It was quite awkward because the feet of the mic stands kept bumping into each other. The mics in use here from front to back are a Behringer B-1, a pair of Behringer B-5s and a pair of Behringer C-2s.

Williams microphone array
Williams microphone array

The result was quite good though. I wrote about it at the time and there are sound samples too. Since doing the above test, I got hold of some more small-diaphragm condenser microphones and in future I’ll use a pair of Sontronics STC-1 for the front pair, the Behringer B-5 for the rear pair and a Behringer C-2 for the front centre – all with cardioid capsules. The STC-1 mics are particularly good.

I also set about building a microphone array holder to enable quick and accurate setup with one microphone stand. Woodwork is not a particular strength of mine but I find it oddly satisfying, even when I make a mess of it. I made a full-size diagram of the array on the back of a piece of wrapping paper (it was all I had that was big enough). The extent of each wooden arm is shorter than the theoretical diagram above suggests, as the body of each microphone forms part of the length.

Sorry for the bad photo – it’s surprisingly hard to photograph a large piece of paper evenly. There’s a 30cm ruler in the picture for scale.

Williams array layout
Williams array layout

I found some scrap wood (pine, with cross section of 15×25mm) and made cross joints to create an aeroplane-shaped form. At the “nose”, the tips of the “wings” and the tips of the “tail fins” I drilled 11mm holes to enable 3/8″ Whitworth bolts to pass through, which screw into five standard mic holders.

Cross joint
Cross joint

I painted the entire construction with extra-matt black paint (left over from blackening parts of my darkroom) to help keep it unobtrusive, and match every other mic stand ever made.

Williams array
Williams array

Finally, I thought the array would weigh too much and be unstable on a single mic stand, so I looked for alternatives. Fortunately this problem has already been solved in the world of photography, where cameras regularly weigh several kilograms. I took a Manfrotto quick-release plate intended for use with Manfrotto photo tripods, removed the standard photographic 1/4″ captive bolt and screwed it into the wooden frame of my array using wood screws through two holes that were already in the quick-release plate. Setup is now much easier: set up tripod, snap the wooden array onto the tripod, attach the 5 mics, plug them in and start recording. I’m considering adding a spirit level to the centre of the cruciform to make sure it’s all level.

Williams array
Williams array

I plan to use this array to make a 5.1 recording of one of the Christmas services at church, and this should be a good real-world test. As I mentioned above, it should also be possible to extract a reasonably good stereo recording by taking the front pair feed as an approximation of an ORTF recording. There will be samples when it gets its first use!

Be First to Comment

Leave a Reply