In a recent conversation among the team we were discussing what the thinking was behind the apparently common practice of using a spaced pair of cardioid mics, often facing inwards, as a pair of overheads on drums? If the idea of a pair of overheads is to capture a stereo picture of the whole kit, rather than be cymbal mics, then exactly what is this arrangement trying to achieve and where does it come from?

How Stereo Works

You can take any two different signals, pan them left and right and call the results “stereo” but this isn’t really stereo, it is twin track, two channel playback. Stereo capture uses an array of mics (usually a pair) placed in a specific arrangement to create a convincing reproduction of the left to right placement of sources around that array. There are rules about this, though these are often misinterpreted as the only way to do stereo - they aren’t, but they are good ways and they are codified so as to be repeatable. Some are very specifically defined like NOS or ORTF, others less so, like AB or Decca Tree.

So there are lots of “right” ways to do stereo. There also some “wrong” ways which, while they might sound good (and if it sounds good then it probably is good) strictly speaking are not “stereo”. A good example is the various “pseudo-stereo” two mic techniques often used to capture acoustic guitars. A ribbon and a condenser on the neck and body might sound nice panned away from each other but they aren’t “proper stereo” - they might well sound better than a “proper” stereo array would in a particular application but it’s important to understand the difference.

Stereo Cues

Any two correlated signals panned left and right will give the impression of width but believable stereo is created by the brain in response to positional cues. These include time of arrival differences to our ears, very specific timbral differences and level differences, which combine to create an approximation of the information the listener would use to interpret where a sound was coming from if they were actually there.

The difference between a pseudo-stereo effect and stereo capture is that an impression of vague “width” isn’t the same as a specific stereo position, though for sounds in the centre of the stereo image the difference is less notable.

How Should You Capture Drums In Stereo?

Like so many things it depends, but if you are trying to create the impression of listening to a drum kit then a suitable stereo array at head height some distance in front of the kit would be the “correct’ way to do it. No-one suspends themselves 7 feet over a drum kit to listen to it! However the sound of a recorded drum kit in popular music has, for a long time been a little removed from the reality. Width in particular is exaggerated, to the point where this isn’t wrong, any more than recording a grand piano from the perspective of someone with ears three feet apart with their head under the lid is “wrong”. Anything that ‘works’ is fine but I do return to my question - Why do people use spaced cardioids on drums?

What is Wrong with Spaced Cardioid?

Stereo relies on the creation of a convincing phantom centre image. How does sound appear to come from between the speakers when there is no speaker there? It’s a good question and one we probably don’t consider closely enough.

Our brains interpret sounds arriving from two speakers as coming from between them if the positional cues match what we would expect to hear if the sound were actually coming from that position. If not, we hear anything from a poorly defined “stereoness” all the way through to hearing two speakers both sides of us, rather than a continuous panorama between them. This is often referred to as stereo imaging.

The positional cues captured by conventional microphones (let’s not get into dummy heads etc.) are level and timing. Depending on how the mics are set, we will capture only level differences (e.g. XY, MS, Blumlein), only timing differences (spaced techniques using omnis - e.g. AB or Decca tree) or a combination of them both (e.g. ORTF or NOS).

This is an exact science and can get complicated. Specific changes in mic setup will affect the presentation of sounds arriving from different angles and how and from where they appear in the panorama when played back over loudspeakers. While we can often play fast and loose with this when recording narrow sources like individual players, these things make a big difference with wider sources. This is what classical engineers spend their time thinking about all day because an orchestra is wide and this stuff really make a difference!

To understand how it makes a difference here’s the maths: When there is no time difference between the loudspeaker signals, then a level difference of 15dB between the right (+15dB) and the left (0dB) will move the phantom image all the way to the right. If the right signal arrives 1.1 ms earlier but there is no level difference, then the signal will be heard as coming from the right loudspeaker.

So depending on how wide your sound source is and how far you are from it, there are mic placements that will be “right”, with the sound source properly filling the space between the speakers, and placements that are “wrong”, with the sound source not being adequately accommodated in the space between the speakers, by being either too narrow or too wide, if sounds are too wide then everything on the edges gets piled up on top of each other.

The term used to describe this is the SRA or Stereo Recording Angle. If you want to know how changes to your mic placement or polar pattern affects the SRA of the array we recommend using Neumann’s FREE Recording Tools App. What you’ll see, if you play with it, is that to maintain an SRA of 90º with spaced parallel mics you need a spacing of 72cm. Less than 51cm and the SRA rises over 180º, meaning at typical distances the kit will sound narrow on playback (which might be a good thing, drumkits aren’t 30 feet wide after all).

Because in this example the mics are parallel the polar response isn’t very significant. If the mics are angled then cardioid mics will of course contribute level differences. If the mics are as close as possible we get XY, with the mutual angle between the mics and the polar pattern being the things that affect the SRA. If the mics are moved apart, we move into near coincident arrays like ORTF where both mutual angle and distance between mics both contribute.

What Are Drum Overheads Then?

So what kind of stereo array are overheads consisting of a pair of cardioids 5 feet apart angled in, to face the snare? I ask because this is something I see a lot and I’ve never understood it. At 5 feet apart (I’m approximating lots of examples here) the SRA between a pair of parallel cardioids is 40º. That means that most of the kit is either in the left of the right speaker with very little in between.

By using cardioids and angling them in on the snare you end up with little meaningful positional contribution from the level differences introduced by the polar pattern. A coincident array works because the centre of the stereo image is off axis by the same amount for both mics and as sounds move away from the centre they move into the most sensitive axis for one mic while moving further into an area of reduced sensitivity for the other. The opposite happens if the mics are facing towards each other and the timing differences and level differences are likely to work against each other!

Spot Mics Or Stereo?

So what is happening here? It might be that these mics are being used in a way which is closer to being “cymbal mics” rather than overheads, being used to capture the sound of the cymbals as spot mics in the same way as the tom mics are being used to capture the toms. If that is the case then that’s fair enough, though the examples I’m thinking of have tended to try to cover the whole kit and are even sometimes used in addition to a close mic on the ride.

This does introduce possibly the biggest factor to this subject, which is that usually these mics are being used in combination with close mics on the kit and the “everything out to the sides” sound these overheads will create, don’t leave us with a hole in the middle of the kit because that hole is amply filled by the close mics - Maybe people do it this way because the stereo image doesn’t matter if you’re constructing it from close mics and just want a “stereo-ey” haze of cymbals around your close mics. Seems a pity though!

Low Ceiling

So why do it “properly”? it depends on how you approach your drum recordings. If you have a great sounding kit in a good sounding room then I’d suggest capturing the sound faithfully, would be a priority. If however, you’re recording in a small, domestic space (and plenty of people are) then you probably want to favour close mics and get the rest of the sound of the kit with as little ‘room sound’ as possible because of your, less than ideal, room.

Cardioid Ubiquity

In this case using cardioids for overheads is potentially beneficial, allowing you to mitigate to some extent the influence of the low ceiling you’re probably recording under - a high ceiling is the single best thing you can do for your drum recordings after getting the right player and drums, with mics and preamps coming a long way down the list compared to having a high ceiling.

Another suggestion is that cardioid mics are the default. Multi-pattern mics are more expensive and fixed omnis are more unusual - pretty much everyone has a pair of cardioid condensers so it’s no surprise that, while omnis might arguably be a better choice for spaced pairs, if you only have cardioids, you’re going to use cardioids!

Centre Line

However something which is so often overlooked is the centre line of the kit. It’s all about where things appear between the speakers, so why do so many people set their overheads either side of the kick drum? Assuming you want the kick and snare in the centre of your mix then the centre line of a drum kit runs through them. So many people don’t do this and as a result the snare is off to one side in the overheads.

Practicalities

In a smaller spaces, my preference is to use cardiods in XY over the snare, correctly oriented for the kick/snare centre line. I have a few reasons for this. The main one is practical - you can mount an XY array on a single stand, very useful when space is tight. Some people see XY as a bit of a “boring” choice but I’d prefer to think of them as reliable. They are very difficult to get wrong and in a less than ideal space their forward (or downward) facing character helps favour the drums over the room.

So where does this spaced cardioids facing inwards thing come from? To check I wasn’t imagining it I checked with the team and beyond. Friend of the blog Mike Exeter agreed:

“I agree about the live thing - that is a good observation. There they are covering the cymbals without attention to image. I think people see and copy.”

We’ll look the issue of placing overheads like they do in live sound in the next paragraph, but what about the centre line Mike?

“Centre line definitely Kick and Snare. Good luck!”

Thanks Mike!

So one suggestion is that people are doing this because it looks right. For many people, opinions on where to put mics is based on what they have seen and in a live sound situation mics are often put up in precisely this position. This isn’t necessarily wrong, the mics are being used for a different purpose and in a very different environment. There is a lot of common ground between recording and live sound but there’s a great deal that’s different. Taking the idea of how something looks slightly further, the “live sound” overheads arrangement is visually symmetrical - something which shouldn’t be overlooked. It’s natural to favour symmetry, we’re hardwired to like it. The only thing is a drum kit isn’t really very symmetrical…

Don’t just take it from us. George Massenburg agrees:

“Most engineers put the overheads in exactly the wrong place. Many young engineers, and some old engineers too, think that stereo miking of a set starts with mics left and right over the cymbal positions. My feeling it that they’re not listening. I don’t know what they’re hearing but if you first imagine that probably the most important role of a rock and roll or R&B set is to have a very strong kick and snare centre image, it follows that your overheads want to be equidistant from your kick and snare so you have a balanced centre image.

If you look at them and try to go with intuition, what you think it should sound like, you’re making a huge mistake.”

What Do You Think?

George Massenburg and Mike Exeter seem to concur that there might be too much ill considered overhead placement going on. Do you use the live overhead placement in the studio? If so why do you use it? Is there something we’re missing? Share your thoughts about overhead placement. One thing I know for sure is that if you ask five engineers how to set overheads over a drum kit you’re probably going to get at least five different answers!