Multi-channel USB microphone experiments
Karla Youngs on Thursday 04 November 2010 Tweet this!
Setting up multiple USB microphones for multi-channel recording.
In the course of researching an upcoming guide to USB microphones I've come across quite a few people wanting to use more than one USB mic at the same time to record in stereo, multi-channel, or to pick up multiple speakers or interviewees, but who are unsure how, and indeed whether it can be done.
We have several identical USB microphones - the Samson Go Mic - which were bought to use in our screencasting workshop, so I thought I'd try to set up a stereo pair to see if I could get it working reliably and what sort of results it might give.
The two Mics clipped together quite neatly, and after a bit of experimentation I managed to find an orientation where I could get the mic heads quite close together and, using the ball-and-socket joints, at angles of between 90° and 180° to each other, with their axes in the same horizontal plane. This gave the option of adjusting the stereo image by changing the angle between the two mics, and provided a near coincident pair, which is the technique I was trying to achieve.
Of course with the Go Mic's switchable pattern you could set up a spaced omni pair, as an alternative to this XY configuration or the similar ORTF and NOS stereo techniques, which also use two cardioid directional mics, but I'll save that for another day. It didn't look too rough and ready either I don't think:

They actually work even better (and are rather more secure) on a boom stand, as the mounting thread ends up horizontal, but that was less photogenic ;)
Now to get them both feeding signals to the computer at the same time. To my knowledge it is unfortunately nigh-on impossible to configure multiple audio devices for simultaneous use under Windows without Steinberg's proprietary ASIO software, and the Go Mic (in common with all USB mics I know of) does not have an ASIO driver, so that was a non-starter. As far as I'm aware (and please correct me if I'm wrong) it can currently only be done a Mac, and this is how:
The OSX Core Audio drivers are configured in Audio MIDI setup, and since OSX 10.2 there has been the facility for configuring Aggregate Devices, which act as a single driver for multiple devices. Plug in the Go Mic and the Mac recognises it immediately, and it becomes available as an input/output device. I then ran Audio Midi Setup (Applications>Utilities) and added an Aggregate Device consisting of two GoMics (plugged into the two USB ports of a Macbook Pro) as shown in the screengrab below.

I chose to use the Macbook's headphone output for monitoring, and the latency turned out to be respectably low - certainly good enough for this purpose, but I assigned this output within GarageBand, so it didn't need to be part of the Aggregate Device. The only minor niggle was that there is no way of labelling the mics to differentiate which is which, and it can be easy to lose track as you add and remove them from the Aggregate Device!
The next challenge was getting them to clock together. Initial recordings started out sounding OK, but degenerated quite quickly into distortion and finally complete meltdown and granulation as the digital clocks gradually drifted out of sync. It started as what I mistakenly took to be abnormally high background noise, but the longer the two mics cohabited in the Aggregate Device, the worse it got, to the point where it sounded like the signal was going through some kind of sonic bacon slicer.
Only disabling and re-enabling one of the mics, or switching master clock source would 'reset' the clock (and then only temporarily - the cycle begins again). The Aggregate Device master clocking is a bit hit-and-miss, and it rather bizarrely seemed to make a difference which order I enabled them as aggregate inputs, but after quite a bit of headscratching and tinkering I arrived at a set-up that worked and was stable, and the background noise all but disappeared [edit - after trying this set-up on a few different Macs, newer machines seemed not to have this problem, despite running the same version of OSX. I'm still unsure why this would be...]
It might seem a lot of trouble to go to, but now it's working I have a good quality, very flexible little stereo condenser mic pair which needs no soundcard, and I think represents very good value indeed for the quality of results. Where I think they'll really come into their own is recording interviews or groups, as each participant can now have their own microphone.
My next experiment is to get more than two going at once, for multichannel or even surround sound recording! But for now I have run out of USB ports on my laptop, so I'll have to try that on a desktop machine...
If you're interested in stereo audio recording, and for some more background to the techniques mentioned in this blog, take a look at our guide to Stereo Recording Techniques.



Older | Newer