Understanding mid-side audio

This page is all about working with mid-side in audio recording. I often hear sound artists talking about mid-side in ways that to my mind are not quite accurate. To explain this is a fairly involved process, so I’ll start with a short summary for those who want to cut to the chase.

Mid-side made simple

  1. The mid-side microphone setup involves one directional mic facing forwards (the mid), and one figure of 8 mic which picks up sound at both sides simultaneously. This can be done with two mics positioned one above the other, or the two capsules built into the same microphone body, referred to as a single shank stereo mid-side mic.
  2. The main reasons to use a mid-side microphone setup are: (i) if you want a stereo recording but one which deals well with sounds in the centre of the stereo field, e.g. for recording a point source plus ambience; (ii) for portability when using full sized mics; and (ii) if you want the best possible quality for centre sounds when summed to mono.
  3. A mid-side mic setup doesn’t give you better control over stereo width than other stereo mic setups. This is a common misunderstanding. You can get mid-side width control with any stereo signal by converting it from a left-right to a mid-side signal, adjusting the width, and then converting back to left-right again. There are plugins which make this easy, or you can set it up manually in a mixer.
  4. Sound quality is a key consideration with any microphone setup, and mid-side is no different. Many single shank mid-side mics have shotgun mid elements, which to my ears are not great for soundscape work. I would tend to choose a good-sounding XY mic, such as the Rode NT4, over a not-so-good sounding mid-side mic, such as the AT4029. Of course, what I mean by ‘good sounding’ might not be the same as what you mean. A lot depends on your purpose. If you want to pull voices out of a cluttered background, then a shotgun mid can be very useful for that. Two industry standard mid-side setups that I’ve used and found to have good sound quality are the Sennheiser 30/40 pair, and the even better-sounding, lighter, smaller Schoeps CMC6 bodies with MK41 and MK8 capsules. Both these set ups are extremely expensive, but can be rented for more affordable amounts of money.

Now the longer version

I’ve pitched the remainder of this page at serious recordists and sound artists who want a better understanding of what MS is all about. Send me comments if it’s not clear and I’ll try to amend things.

Mid-side (abbreviated to MS) is often referred to as a stereo microphone technique, which uses a directional mic such as a cardioid pointing forwards, and a figure of 8 mic picking up on both sides. These signals are then added and subtracted to produce left and right stereo channels, usually called LR stereo (L = M + S, R = M – S).

MS, as a microphone setup, is a type of coincident stereo setup, i.e. the mic capsules are located as close together as possible, directly above each other, rather than spaced apart. This means that, in theory at least, the differences between the signals registered by the two channels are only differences of level, not of timing (and hence not of phase). As a result, the stereo signal can be summed to mono (two channels mixed together to produce one channel) without causing phasing issues. (With spaced pairs, summing to mono tends to cause phasing, which cancels out certain frequencies and gives the sound a hollow quality).

It’s often said that MS mic setups have the advantage of allowing greater flexibility to alter the stereo width of a recording in post-production. For a narrower stereo image, increase the level of the mid signal relative to the side; for a wider stereo field, increase the level of the side channel relative to the mid.

That description is not exactly wrong, but it’s misleading because MS is not only a mic setup, it is a way of encoding any pair of signals. If you record an LR signal using any stereo mic setup – an XY pair or spaced A-B pair, for example – you can convert that into MS, adjust the relative levels to change the stereo width, and convert back into LR. MS conversion is often used in mastering for example, particularly for vinyl. The physics of etching two channels of sound in a single groove make it important for engineers to be able to control phase relations between the channels, and MS is great for that.

In other words, you don’t need to use an MS mic setup to get MS control over stereo width. It might be a bit easier to do this if the original recordings are made in MS, but in modern digital systems it is fairly simple to convert between LR and MS. Logic, for example, has a plugin called the direction mixer. You can feed in either LR or MS signals. If you input LR, it internally converts to MS, has a ‘spread’ control which changes the balance between mid and side to control stereo width, and then outputs an LR signal.

I also use Sample Manager by Audiofile Engineering, now called Myriad, which can convert single files or batches of files from LR to MS and MS to LR. It’s also not too difficult to set up an MS routing in a mixer, either in software or hardware. There are online tutorials for this.

If any stereo signal can be turned into MS, what is the advantage of an MS mic setup? I think there are three situations in which an MS mic setup is an advantage.

Reason for using an MS mic setup #1. If you want a stereo recording with good quality in the centre. An MS mic setup lends itself to what I would call ‘point source plus ambience’ recordings. For me, the typical example is recording an interview in situ for what I would call ‘environmental ethnography’ – environment not in the sense of natural environment, but any environment. In this situation, I want to pick up the ambience of the location in stereo, but I also want to person’s voice, in the centre of the stereo field, to sound good too. That same logic could apply to other point sources where you want to record them clearly, but with a strong sense of surrounding context, rather than isolating them from the wider soundscape.

Why is an MS mic setup good for this? Well, imagine you use try to do this using a spaced pair of mics. Spaced pairs are great for getting a wide sense of space, but they can suffer from a ‘hole in the middle’ effect, because centrally located sound sources are not close to either mic.

To avoid this, we could use a coincident pair, like XY, with two directional mics above one another but set at an angle. Now both the mics are closer to the central sound source, but neither is aimed directly at that sound source. In technical terms, a central sound source will be off axis for both mics in an XY setup. Chris Watson put expressed this with perfect clarity during a workshop some years ago, when I was recording him using a Rode NT4, an XY stereo mic with two capsules angled at 90 degrees: “neither of those mics is pointing at me”. Why does this matter? One of the features of directional mics is that they tend to sound best when the sound source is on axis – i.e. aligned with the middle of the microphone’s polar pattern. These mics sound less good on sources that are off axis, which is understandable because directional mics, by definition, are designed to attenuate sounds that are not on axis. With an XY pair at 90 degrees, a central source is off axis by 45 degrees in each mic. Try talking into the side of a classic cardioid mic like an SM58 and you’ll notice that not only does the sound get quieter, as you’d expect for a directional mic, but it also starts sounding a bit thinner in terms of the frequencies picked up. Directional mics are specifically designed to deal with what is in front of them, not sounds off to the side.

[diagram]

With an MS mic setup, central sound sources will be solidly on-axis for the mid mic. That will tend to improve the sound quality for those central sources. Now, in some cases we’re talking about quite subtle differences in quality. I’ve often used XY setups for recording voices and other point sources, and the quality is fine. Also, expensive mics tend to sound less bad off axis – Schoeps are very good for this. Nevertheless, if quality in the centre is paramount, then MS may be worth considering – subject to the important points about sound quality below.

Reason for using an MS mic setup #2. If practicality and portability are important. Think about the single-shank stereo mics that are portable enough for field recording: Rode NT4, Audio-Technica BP4025, Shure VP88, Sennheiser 418, Audio-Technica BP4029, Sanken CSS5. Apart from the first two, these are all MS mics. Equally if you’re doing stereo with two separate mics, MS has an advantage over other setups in terms of portability, because in this configuration the mic bodies are in line, so you can fit them into a single windshield. With full size microphones, most other stereo techniques require separate windshields for each mic because of the way the mics have to be positioned, at an angle to each other or spaced apart. Having two windshields is quite a lot to cart around, and requires stereo mounting bars or dual mic stands. An MS setup is much more portable for a single person moving around on the hoof, which is what most field recordists do. You can mount one mic on top of the other using back-to-back clips, put them in a single Rycote windshield blimp, and hold that in your hand or mount it on a boom.

The classic Sennheiser MKH30/MKH40 pair works this way. Wildlife recordist Geoff Sample once said that he uses a 30/40 pair pretty much exclusively, and his recordings are magnificent. You can also use other mics in the mid position, such as the MKH50 which has a tighter super-cardioid pattern, or a shotgun like the MKH60, though I’ve never used either of these. These mics are very expensive, but if you’re on a budget you can hire them. I’ve loaned 30/40 pairs from the Warehouse in Edinburgh and Glasgow before. Richmond Film Services in London also have them.

The other MS set up I’ve used and loved is my current main rig: two Schoeps CMC6 bodies, one with an MK41 super-cardioid mid capsule for the mid, and the other with an MK8 figure of eight capsule for the side. Again, superb sound quality, very expensive, but can be hired. To my ears they sound even better than the Sennheiser 30/40, though that’s very subjective. The Senns sound a bit darker and very slightly compressed; the Schoeps just sound utterly transparent to me.

Unlike the Sennheisers the Schoeps are small enough to fit into a Rycote mono windshield. Rycote sell a larger windshield for stereo pairs, which they recommend to give maximum wind protection, but it’s massive, like carting around a melon. The mono shield, whilst hardly small, is a bit more manageable on the hoof, and seems to work OK. The Schoeps are also very light, noticeably moreso than the Sennheiser 30/40.

This portability argument only applies if we’re talking about full sized mics. There are lots of people who argue that for field recording full sized mics are a waste of time, and that miniature mics are the ticket. Ian Rawes of the London Sound Survey, for instance, swears by small mics because he needs to be discrete in busy urban environments. A lot of field recordists use miniature DPA 4060 omni mics, which are fantastic quality, light as a feather, can be head mounted, or binaurally mounted, or spaced apart in a windshield or on a wire coathanger (a Chris Watson technique). So you absolutely don’t need MS to have a portable setup; it’s just a more portable than other options with full sized mics. For what it’s worth, I love my miniature DPA omnis, but I find the Schoeps MS rig more versatile, and easier to handle. With the DPAs the cables always seem to get tangled up, you can’t unplug the mics from the cables, and as they’re omnis they have to be used as a spaced pair for stereo, which doesn’t work so well if you have something important in the centre of the stereo field.

Reason for using an MS mic setup #3. If mono compatibility is required with no compromise of quality. MS is often noted as being highly mono compatible, because summing to mono simply cancels out the side signal, leaving the mid. Again this is slightly misleading, because other coincident setups such as XY and Blumlein pairs are also mono compatible, so the advantage isn’t unique to MS mic’ing. But an MS mic setup might have the edge in sound quality over XY when summed to mono. Adding together audio picked up off axis by two cardioids in an XY setup is likely to result in poorer quality than summing an MS signal, which leaves just the on-axis single mid mic. If mono compatibility is really important to you then this might be a factor.

The final, very important issue with MS mics is how they sound. Lots of caveats are needed here. Sound quality is so subjective, capsules from different manufacturers will vary massively, and as I have made clear above it all depends on what you’re recording and what you want. So there are no hard and fast rules, and no MS ‘sound’ as such. However, based on personal experience, I will say two things.

The first is that there is a discernible ‘sound’ to the Schoeps MS setup versus other Schoeps setups, and I like it – though others might not. A couple of years ago Jonathan Prior and I rented out a load of different Schoeps capsules so we could compare different stereo setups. Comparing MS with XY, for instance, both are coincident, and in theory should record a similar size and shape of field. But I can definitely hear a difference. Subjectively, I prefer the MS, but that may because I am familiar with it. To me, the XY sounds a bit flat somehow, whereas the MS has an airiness around the sides that I like. It’s very solid in the middle, and then lighter stretching to the far left and right edges of the field. Have a listen and see (or rather hear) what you think:

Looking at the data for these mics, the figure of 8 capsule has a gentle roll off in the low end below 200 Hz, dipping to -6dB at 50Hz, which might be responsible for this slight airiness in the sides. The cardioid capsules also roll off at the low end but only to -3dB at 50Hz, half as much. More generally, I would guess that the physics of making a figure of 8 versus a cardioid are different, and both will be different from an omni, so there is bound to be some sonic influence from that, even where a manufacturer is trying to get a consistent signature ‘sound’ across their range of capsules.

(For the record, if bass response is your priority, then by far the best setup we tried was omnis in spaced pairs. Schoeps data says these are ruler flat down to 20Hz, and in our tests the low end was way more extended than any of the directional setups. There was loads of sub in there – not hyped or beefed up, just present.)

The second thing to say is that many single shank MS mics, such as the Sennheiser 418 and AT4029, tend to have a shotgun as their mid capsule, because they are designed for applications where shotguns are useful, such as film or TV. Now, again this is very subjective, but I am not the biggest fan of the sound of shotguns, especially for soundscape recording.

With a shotgun, you get strong directionality at middle and upper frequencies but that comes at a price: a certain kind of nasal, closed quality, and boominess in low frequencies, particularly noticeable indoors (though I have yet to hear the Schoeps CMIT 5 shotgun, which is designed to minimise these issues). Most shotgun mics are designed to pull voices out of cluttered backgrounds, and therefore they often work well with voices, but less so for other kinds of sources. The Sennheiser 416, for example, is an industry standard mic for TV and film work because it accentuates certain frequencies that help voices cut through, but as a result it has a strong sonic character that is far from flat or transparent. There is a great rant here about that: http://www.coreyburton.com/sennheiser.html The side capsules used in shotgun-based MS mics may not be the best either. The Sennheiser 418, for example, has a reputation for having a noisy side capsule.

Of course, lots of great field recordists use shotgun-based MS mics, because they are light, practical to carry around, they fit in a single windshield, they tend to be weatherproof, they record in stereo, and the shotgun element can be used to focus in on point sources of interest: birds, squeaky gates, water gurgling in drains, footsteps and so on. Peter Cusack told me he uses a 418 because of its practicality, despite the sound quality not being ideal. So I wouldn’t count them out automatically, but a bit of caution might be advisable if you want to record soundscapes and your focus is on sound quality. On a project some years ago I used both an Rode NT4 XY mic and a hired Senn 418 to record school soundscapes, and I was surprised at how much I preferred the NT4, given that the 418 cost over three times as much.

So if using a mid-side mic setup means using a shotgun-based, single shank mic, then for soundscape recording personally I would avoid it. And the idea that a mic of this kind gives you more flexibility afterwards to make a wider stereo image, as compared to an XY mic like the NT4, is simply wrong. The NT4 isn’t perfect – it’s quite heavy and it doesn’t cope well with humidity, moisture or wind – but as a single shank mic for ambience and soundscape type work I would choose it over any of the shotgun-based MS mics if sound quality is the main priority. The advantages of the shotgun-based MS mics over the NT4 would only be for picking out specific point sources at distance, and practical stuff like weatherproofing and weight.

That said, I’d choose my Schoeps MS rig over the NT4 for pretty much everything. It is especially good for what I would call ‘environmental ethnography’, interviewing people in situ, where I want the best possible rendition of both the voice and the wider sonic environment. As an example, here’s a clip of Peter Cusack, recorded in Berlin, using the Schoeps MS setup.

If you have comments and questions please post them at the bottom.