The established techniques for capturing stereo have been serving engineers well for decades, but they are not set in stone. We ask when and how they can be changed to work for you.
Using Two Mics For True Stereo
True stereo recordings bring depth and realism to recordings, which are entirely different to ‘panned stereo’ mixes derived from mono sources. It’s true to say that some sources such as a single voice are too narrow in real life to warrant anything other than panning for any stereo placement, however things are different when it comes to wider sources from acoustic instruments up to entire orchestras. These sources benefit from the true stereo treatment by employing long-established mic techniques to translate their size into two channel stereo. Incredibly, the techniques engineers used today to capture true stereo sound have their roots in research beginning in the 1930s, with their refinement carrying on ever since. These revolve around using two microphones placed in front of the source to create interchannel differences that the ear can use to hear stereo ‘images’.
The Established Techniques
For stereo to happen, two mics can capture these differences in one of two ways. Level differences can be captured with angled directional mics, or time differences (usually) with omnidirectional mics. Hybrid techniques using both approaches are common.
Of the level difference techniques X-Y is one of the best known, using coincidently placed cardioid mics angled at 90 degrees away from each other. The A-B (or spaced pair) technique captures time differences with mic spacing varying greatly depending on factors such as the source width and the effect desired. Hybrid techniques see directional mics angled and spaced away from each other. An example of this is the DIN (Deutsches Institut für Normung) technique which employs cardioid mics spaced by 20cm and angled 90 degrees apart.
Following research and experimentation from engineers going back decades, conventions and standards started to emerge in the approaches used. These mainly concerned themselves with factors such as microphone directivity, angle, and spacing, and differed between the institutions devising them. Because of the differing advantages and disadvantages of each, many variations emerged, with those devised by large European broadcasters gaining traction in the wider recording world. While each system’s ‘standard’ for polar pattern, angle and/or spacing had its roots in scientific research and practical listening tests, the main advantage was their learnable, repeatable nature. Many engineers today would concede that these numbers are not ‘magic’, and can indeed be changed according to the task in hand.
Changing SRA’s
The Stereo Recording Angle (SRA) of an array describes its ability to accurately capture the width of the source in front of the mics. When the SRA matches the source, sources fully left will be heard as such over stereo loudspeakers set up at a given (60 degree) angle to the listener, with the converse for right channel sounds. The lower the number, the wider the array will sound on a given source. An array will still ‘work’ when the SRA and the source don’t match, and as such the engineer may wish to narrow or widen the image artificially for things like drum kits or acoustic guitars. If the SRA is incorrect for the width of the source then you might experience issues such as sounds crowding towards the speakers rather than being spaced progressively across the soundfield. Whether or not this lack of accuracy matters depends on the context and many ‘wrong’ setups are very successful, just not necessarily accurate. The SRA for any given setup can be calculated using the excellent Neumann Recording Tools app.
Below are three stereo acoustic guitar recordings made using three textbook stereo microphone setups with a pair of Line Audio CM3 sub-cardioid mics. The pattern of these mics further changes the SRA of sources by decreasing the level difference between left and right. All SRAs quoted accordingly. No audio processing was used.
This array has an SRA greater than 180 degrees. This means that the source is technically too narrow for full width reproduction.
This array has an SRA of 131 degrees. The time differences introduced increase the width on playback compared to X-Y.
This array has an SRA of 130 degrees. Greater level differences countered by smaller time differences compared to DIN give both very similar widths.
None of the examples above is technically correct for the SRA’s of the setup used, owing to the physical width of the source as seen by the mics and wide cardioid pattern of the mics used. Do they sound good? Only you can decide.
The Sound Is Sacred
Although the established techniques can achieve predictable results, they do not represent the best possible technique for every situation. Factors such as image width, ambience, and perspective can all be controlled by adjusting the array’s parameters, trading one for the others. For any given source, with some thought and careful listening, these qualities can be optimised for the aesthetic sought. Certainly, deviating from the text-book setups’ parameters does not ‘break’ the sound.
I chose the three setups featured above because they are a great starting point, as established by a lot of experimentation over time by a lot of clever people. In the end, to my ears, I preferred the DIN array, even though SRA is wrong. It provides a good tonal balance with an image that is wide enough to give it noticeable stereo interest, yet doesn’t make the guitar sound like it’s ten feet long! Would I change that to make it ‘right’? Absolutely not.
Textbook techniques are totally invaluable when faced with time pressures or a large source such as an orchestra or choir in an unfamiliar venue. That said, given any time to experiment, adjusting things like ambience or spread can give better results than just sticking with an SRA. The SRA’s for the setups used in this article are not technically correct, but you might like some, or none of them at all. For these reasons, you might find, like me, that the textbook stereo techniques are not sacred.