Minimizing mic noise

Microphones are designed to be heard, but they should ideally be translators of sound rather than originators. Some self-noise is inherent in every microphone, such as the jostling of air molecules against the diaphragm or the thermal noise in the electronics of an impedance converter. However, these contributions — usually just a continuous bland hiss — will be very low compared with the wanted sound. What irritates listeners far more is the much higher levels of erratic, impulsive noise that are often mechanically transmitted to the microphone.

Microphones detect sound by sensing the movement of a light diaphragm relative to the fixed, massive reference of a back plate or a magnet. However, such movement can also be generated if the reference moves — for example, if the microphone or its cable is struck — while the diaphragm is held still by inertial forces (such as the mass of air resting against it). Unfortunately, no microphone can distinguish between the first type of movement, usually caused by a sound we want to record or broadcast, and the second, which is the result of unwanted vibrations.

Obviously, the aim should be to keep the microphone body completely free from external, mechanically transmitted noise, but that is not so simple. Microphones have to be physically supported in space and are most often connected to the outside world via a cable — and both routes provide a pathway for mechanical noise. The cable is a highly significant conduit, and using thin, flexible cable and a properly anchored decoupling loop can reduce the noise it transmits. Dealing with the support is trickier.

Springs and masses

The efficacy of mic suspension at reducing unwanted vibrations is dependent on the frequency of the vibrations. Handling noise, which includes that transmitted through poles and stands, tends to have a spectrum that is tilted heavily toward low frequencies. As the frequency increases, it becomes progressively harder to move the physical mass of a microphone body; thus, handling noise rarely has much content above a few hundred hertz unless the microphone is very small and light, such as a subminiature personal type. However, the low frequencies are also the hardest ones to eradicate.

A typical mass/spring system can isolate suspension. It consists of an oscillating mass exhibiting compliance (the ease with which the mass is set in motion) and damping (which is the dissipation of energy from the system). Mic suspension complies with the rules of physics associated with such a system. One of these is that given the fixed mass of a microphone, the lower the frequency of the handling noise, the further suspension has to be able to wobble to give a particular degree of isolation.

To isolate effectively from low-frequency handling noise, ideal suspension should have high compliance and damping. This ensures that the microphone can move many centimeters, but will be brought back gently to its original position with minimal overshoot (perfectly damped). This sounds great except that practical high-compliance suspensions can also set up the microphone to wobble up, down and sideways rather wildly and make it difficult to control on a boom pole, for example. Fortunately, microphones are not as sensitive to handling noise on every axis. The one axial to the diaphragm, Z, is far more important than the other two, X (sideways) and Y (up and down), so sophisticated suspension can exploit this to give both good isolation and good control.

Importance of resonance

Every spring system has a natural resonant frequency, and at the point of resonance, a stimulus will be accentuated rather than damped. This fundamental frequency also includes harmonics that will repeat the same behavior at higher frequencies, though in an increasingly damped fashion. Because resonance is rarely identical on each axis, microphones can frequently buck and yaw rather vigorously near these particular frequencies. However, at above roughly three times the resonant frequency, the fairground-ride behavior calms down, and it is possible to make a suspension isolate with reasonable efficiency. (See Figure 1 on page 17.)

The resonance point is important in defining how suspension functions and needs to be as low as possible. Ideally, it should be less than one-third of 20Hz (the lower limit of human hearing), so the mount is isolating at the lowest frequency we can hear. With large, heavy microphones, this is feasible (though not always easy). With lighter microphones, it gets progressively more difficult; the spring element has to be much more compliant to match. With the small, light designs beloved of location recorders, getting the fundamental below 20Hz can take some ingenuity.

Practical considerations

The biggest problem is combining this high compliance with sufficient control to prevent the microphone from wobbling around so vigorously that it starts to generate extra handling noise, becomes uncomfortable or distracting for artists underneath, or even crashes into the basket of a windshield. Being able to tailor different degrees of compliance into each axis and to dampen any oscillations heavily so the mount doesn't “ring” are key factors. (Ringing here is used in the sense of long-term oscillation after an energy impulse, not necessarily to mean producing a bell-like tone).

Another problem is nonlinearity. Rubber bands, cushions or diaphragms are extremely common in suspensions, but their behavior under tension is severely nonlinear. As with stretched guitar strings, they become harder to move the tighter they get — the compliance reduces drastically — so the resonant frequency rises sharply with displacement. This means that the suspension works dramatically differently for small or large movements.

New approaches

Two entirely new designs for suspensions have appeared in recent years that have rethought the problem of keeping microphones cushioned and quiet, as well as have introduced some alternative solutions to simple bits of rubber. Both can tailor the compliance on each prime axis, and both have sufficiently linear action to give effective isolation for the large displacements that occur at low frequencies.

The French OSIX design, derived from research originally funded by Stefan van den Burg and now handled by Cinela, uses a wire spring curved in a hoop shape that crosses over itself with the microphone inside the hoop. (See Figure 2.) The single coil can bend forward and backward, enabling a high Z compliance. For up, down and sideways movements, the loop distorts, which requires more force and makes the suspension stiffer in these directions.

Metal springs have little self-damping, so the OSIX has to provide separate damping elements in the form of rubber bands twisted around the fixing points. This arrangement dampens all movements that require the spring to twist at its anchors, but is less effective at dealing with twisting in the wire itself. This means that there is some degree of translation of movements from one axis to another, though the OSIX is far from weak in this respect.

As a classic design of spring, mass and damper, the suspension is best tuned to some degree, so OSIX products are marketed for individual microphones. They are essentially inflexible in this respect, and this, coupled with the complex assembly, makes them moderately expensive. They also suffer from a degree of fragility — the delicate spring is easily bent — but they are highly regarded by users.

The other design, researched and developed at roughly the same time in the UK, is the lyre suspension, so-called because the elastic element has the shape of the classical instrument. (See Figure 3.) Although the overall outline of the lyre is roughly circular like the OSIX, it has a different operating principle. Z compliance is governed largely by the long path length of the recurved arms, which allow a high degree of Z-axis movement, while X-axis compliance is much reduced; to move in this plane, the arms of the mount would have to tighten their sideways curl. Y-axis movement is also restricted, because both pairs of arms have to move simultaneously to raise or lower the central clip.

The lyre's shape determines the compliance rather than the material used, but the latter is important for the damping. Unlike the OSIX, the lyre can have integral damping, which simplifies the design and tends to calm any translations of movements from one axis to another. The lyre's use of a moldable shape memory material brings two other benefits: It is virtually unbreakable and enables the addition of an integrated clip for the microphone, minimizing assembly costs. This makes the design suitable for a wide range of microphones. It also allows the design to be scalable across a wide range of sizes, because these plastics retain their strength and toughness even when very thin, whereas rubber bands and wire springs soon become excessively delicate.

Future of suspensions

Development in suspensions has been an extremely slow process. There have been a tiny handful of new ideas since the Shure Donut in 1970, but nothing compared to the flood of improvements in microphones in the same period. It is likely that suspensions based on rubber bands, diaphragms and webs will be with us for a while, because they are simple to conceive and inexpensive to produce, and less inventive manufacturers and less discriminating users are unlikely to give up such designs in the short term. But both the OSIX and lyre suspensions have shown that much better techniques are now available and have set a higher level of performance that new contenders will have to reach.

Chris Woolf is an independent design engineer who has worked with companies including Rycote and Schoeps.

Performance measurements

The plots reproduced here show the performance of various types of suspension with rising frequency, as measured along the Z axis of the microphone. Note that the upturn in the response at just below 2kHz is a typical artifact of the fast Fourier transform measuring system used; it was left in only to prove that these are real plots.

In each case, the black trace is the performance with no isolation; a shaker is used to rattle the microphone directly. The red trace shows what happens when the shaker does the same thing but via the suspension. The blue shading gives a quick visual indication of the benefits the suspension gives, and the 50Hz notes at how low a frequency the mount manages to isolate. Notice that the red trace always rises above the black at very low frequencies, showing that the mounts do indeed amplify movement under these conditions.

Figures a, b, c and d show the same microphone mounted in diaphragm, web, band and lyre mounts, respectively. The mounts used are typical ones, but without any attempt to optimize performance.

Figures e and f show similar plots using a compact microphone in OSIX and lyre mounts. The microphone is much smaller and lighter, which makes isolation more difficult. The performance of both mounts is broadly the same, but note that the OSIX has a significant extra mass (65g as opposed to 40g), which helps to lower the resonant frequency — something the lyre suspension does not need.