More on Microphones by Michael Williams, (www.williamsmmad.com)   △   <   ∧   >

8.4 - WIND NOISE PROTECTION

I make no apology for basing the majority of this section, and with his permission, on the excellent paper^[14] presented at the 91st AES Convention, New York in 1991 by Jorg Wuttke, Technical Director of Schoeps (1979 to 2007). As Mr Wuttke makes clear from his presentation, there is no perfect solution to reduce interference caused by the noise generated by wind – it is always a trade-off between efficiency and side-effects. Wind-noise is a combination of a general linear air-flow and, local turbulence either generated by surrounding physical obstacles and/or by the windshield structure itself, or by the wind protection casing of the microphone.

There is however a major difference between the sensitivity of a pressure microphone to wind and that of a pressure-gradient microphone as shown in Figure 75.


Figure 75 - comparative amplitude of wind noise interference
for a pressure microphone (Schoeps MK2) and
a pressure-gradient microphone (Schoeps MK41)
© frequency response curves by courtesy of Schoeps

The magnitude of interference will depend basically on the type of microphone being used - pressure microphones will be less effected than pressure-gradient microphones. The bi-directional microphone will be the most effected, with the interference due to wind decreasing with hypercardioid, cardioid and hypocardioid directivities – the magnitude of interference being proportional to the bi-directional component in the directivity equation. So in very bad wind conditions it is always advisable to start by choosing omnidirectional or hypocardioid microphones. The other directivity choices are a matter of the degree of wind interference. Three basic types of wind shield exist for sound recording situations

• the synthetic foam ball
• the spherical or cylindrical basket shell covered with a fine mesh or synthetic fur
• the ‘softie’ – a combination of an extremely open plastic ‘pan scrubber structure’ covered with a synthetic fur covering

The synthetic foam ball shown in Figure 75 is made from an open cell foam with the microphone housing cut into the mass of the foam. This is a cheap practical solution but with limited efficiency protection against wind noise, and really only adapted to pressure microphones.

The basket-and-mesh windscreen shown in Figure 76 can create a high degree of wind protection – the protection being proportional to the diameter of the windshield basket and the density of mesh. It is the best wind protection solution for pressure- gradient microphones. However with a high degree of protection, standing wave resonances inside the frame of the basket will somewhat degrade the frequency response curve, and low frequency directivity will also be affected.


Figure 77 shows the relative efficiency of the synthetic foam windshield (W5) as against the fine basket-and-mesh windshield (W20) when used on a pressure microphone. The wind noise rejection is remarkably similar in amplitude, whereas the comparison between the frequency response curves in Figure 78 leaves no doubt that the synthetic foam windshield produces less interference with the high frequency response - a small amount of high frequency lift will correct the high frequency roll-off. This HF correction can be achieved either with an electronic high frequency equalisation circuit or a parametric filter, or by using a microphone capsule where high frequency lift has been integrated into the microphone design so to obtain a flat energy response (directivity index of 1 throughout the frequency range) as shown in Figure 40 in section 8.2.1.


Figure 77 - Comparative Amplitude of Wind Noise Interference
for a Pressure Microphone (Schoeps MK2)
With and Without Windshields W5 and W20
© frequency response curves by courtesy of Schoeps

Figure 78 - Comparison of Frequency Response of an Omnidirectional Microphone (Schoeps MK2)
with No Windshield, and W5 and W20 Windshields
© frequency response curves by courtesy of Schoeps
On the other hand Figure 79 shows the relative efficiency of the fine mesh basket windshield (W20) as against the synthetic foam windshield (W5) when used on a pressure-gradient microphone.


Figure 79 - Comparative Amplitude of Wind Noise Interference
for a Pressure-Gradient Microphone (Schoeps MK4)
With and Without Windshields W5 and W20
© frequency response curves by courtesy of Schoeps
For pressure-gradient microphones, there is absolutely no doubt as to the advantages of the basket and mesh windshield over the synthetic foam windshield, as shown in Figures 79 & 80.


Figure 80 - Comparative Amplitude of Wind Noise Interference
for a Pressure-Gradient Microphone (Schoeps MK4)
with and Without Windshields W5 and W20
© frequency response curves by courtesy of Schoeps
The absolute levels of wind noise in the frequency response curves (Figures 77 to 81) have not been calibrated with respect to the microphone frequency-response curves. However relative wind noise levels from one curve to another are valid, as are the relative levels of the frequency response curves of the microphone. For a pressure-gradient microphone, there is absolutely no doubt as to the advantages of the basket and mesh windshield over the synthetic foam windshield as shown in Figure 80, but of course the trade-off is evident in the irregular high frequency response curve, low frequency roll-off and degradation of directivity pattern as shown in Figure 81. Unfortunately this is the price to be paid for the high efficiency wind protection. Using the W20 windshield - below 200Hz the microphone directivity becomes almost omnidirectional, above 5kHz the directivity becomes hypocardioid, and parallel attenuation at 90° is only valid over a short frequency range from about 400Hz to 2.5kHz.


Figure 81 - Frequency Response of a Pressure-gradient Microphone
at 0°, 90° and 180° for Schoeps MK4 with W5 and W20 Windshields
© frequency response curves by courtesy of Schoeps
Wind noise is essentially low frequency disturbance, so high-pass electronic filtering would also seem to be a suitable remedy. This is certainly the case for very low velocity wind interference and therefore small excursions of the microphone diaphragm due to wind. However higher wind-flow velocity can easily be more than 100 times the acoustic signal particle velocity, which means that the diaphragm enters into non-linear mode or can even be pushed to the limits of physical excursion. These high level signals invariably overload the pre-amplification circuitry producing drastic signal clipping. These distortion components are impossible to filter electronically. It is therefore important to reduce wind impact very much at source and use electronic filtering only to reduce any residual low frequency wind noise.

The synthetic foam wind protection ball can be improved for pressure-gradient microphones by creating a small air space between the microphone capsule and the surrounding foam. This combines the advantages of the synthetic foam with the basket- and-mesh technique, however protection will remain considerably less than with the basket-and-mesh design. The synthetic foam protection has the advantage of less internal resonances within the air space - the foam will in fact absorb some of these resonances. Any high-frequency roll-off caused by the foam can be easily compensated with electronic equalisation or microphone capsule equalisation. However for higher wind velocities and for pressure-gradient microphones the basket structure is the only solution. Multiple screens have shown to be only marginally effective ^[15] – a combination of three screens being the absolute maximum needed. However it has been shown ^[15] that, within limits, size is the important feature of a good windshield - the magnitude of protection is proportional to diameter of the basket-and-mesh structure. There is a similar advantage in size with respect to the pressure-gradient directivity pattern, the larger windscreens will cause less interference with the basic directivity pattern in the low frequencies.

The use of an outer mesh made from a synthetic fur has considerably improved the performance of many of the windshield designs but again, the frequency response is still affected but to a lesser extent compared with the simple mesh design. Air flow due to wind will generate noise on the surface of a windshield, due to friction at the boundary between the moving air and the stationary windshield surface. This boundary friction noise can be reduced by the use of a synthetic fur material in place of the mesh screen (example in Figure 76 - the Schoeps windshield W20R1 and in Figures 82 and 83 – Rycote windshields), with the added advantage that the fur surface will tend to dampen some of the resonances within the basket container. The fur fibre is attached to a material mesh base, which acts in a similar way to the fine mesh of the traditional basket-and-mesh windshield.


Figure 82 - a Rycote Windshield/Windjammer (15cm in diameter)
with a Suspension for a Schoeps M/S Microphone Pair
(Cardioid MK4 and Bi MK8)
© photos by courtesy of Rycote

In very high wind velocities the smaller windshield does not give adequate protection, even with a fur covering, so the larger fur covered windshield diameter as shown in Figure 77 (15cm) is really a necessity. Measurement of microphone characteristics using some form of windshield protection is not a simple process - the problem being how to generate wind without producing extraneous noise – the generated wind must have both air flow velocity and turbulence in order to simulate real natural conditions. Mr Wuttke in his AES paper ^[14] explains some of the difficulties in both creating these conditions and making measurements with the custom made radial ventilators, which are used to generate wind disturbance in a laboratory.

In 2000 at the 109th AES Convention^[26][27] Chris Woolf and Oliver Prudden presented two papers describing a technique for accurately measuring and displaying graphically the efficiency of windshields in real outside wind conditions, using real-time comparison between an unprotected reference microphone and the protected microphone within the windshield system. This approach would seem to offer a reliable alternative to the laboratory wind machine. Perhaps the fact that Chris lives in Cornwall in the extreme SW of England also assures a relatively frequent supply of windy conditions in which to make the measurements!

Up to now we have discussed the effect of wind on the microphone capsule. Wind noise is also generated by the air-flow friction on the surface of the microphone body. It is therefore important to surround the whole of the microphone (capsule and body) with a wind protection shield. A small spherical windshield will protect the microphone capsule but offers no protection against wind-noise on the microphone body, which at high wind velocities can be considerable. One solution, used with the usual tubular form of microphone, is to enclose the whole of the microphone in a reasonably long windshield, as shown in Figure 78.


Figure 83 - The Long Rycote Windshield WS4 (10cm in diameter)
and the Windjammer WJ4 = WS4 with Synthetic Fur Covering
© photos by courtesy of Rycote
The wind protection should also include a few centimetres of the cable coming from the microphone plug. It is easy to determine the sensitive areas that have to be protected, simply by stroking the surfaces whilst listening on headphones to the amplified output, for instance on a portable mixing desk or recorder. If you can hear the sound made by touching the microphone body and the cable, then you will need to protect these surfaces from wind noise. Another way of protecting the microphone body from wind noise in less violent wind conditions, is the use of the ‘wind spoiler’ which covers the exposed body of the microphone as shown in Figure 84.


Figure 84 - the Rycote Wind Spoiler mounted on a ’Softie’ Windshield
for a Short Rifle Mic
Unfortunately the Rycote windspoiler is no longer available, however it is possible to make a do-it-yourself replacement with any cellular foam tubular covering, normally used to insulate water pipes against the frost. Do not try using this type of foam as a windshield simply because it has a closed cellular structure, each cell is like a bubble with a thick skin, whereas the foam used in a foam windshield has an open cell structure allowing the sound wave to pass through the openings in the foam to the microphone capsule, with a minimum of attenuation.

The ‘Softie’ is the third type of windshield in current use, being really a cross between the fur covered basket and the synthetic foam windshield. The synthetic fur outside covering of a softie is maintained in shape by a very open cellular structure, similar to a wire mesh pan scrubber as shown in Figure 80, except of course it not made from wire! The advantages of the synthetic fur outside covering are maintained without the problems of diffraction and resonances associated with the internal rigid basket structure. It does however have one small disadvantage in that it is not suitable for very high wind velocities as the general structure which supports the outer fur covering is not as rigid as the basket and mesh windjammer shell. It does however offer efficient wind protection in the lower wind speeds found in every day operational conditions, and it is a much cheaper alternative to the basket-mesh-fur windjammer.


Figure 85 - The ‘Pan Scrubber’ Interieur of a ‘Softie’
[14] 1992 – "Microphones and Wind", by Jorg Wuttke, Schoeps, Germany.
1991 - 91st AES Convention, New York; 1992 - JAES Vol 40 No10, pages 809-817

[15] 1961 – "Experimental Determination of the Effectiveness of Microphone Wind Screens" by John C. Bleazey, RCA Laboratories, USA JAES Vol 9 No1, pages 48-54

[16] 2000 – "Windnoise Measurement using Real Wind" by Chris Woolf and Oliver Prudden, Rycote Microphone Windshields, UK, 109th AES Convention, Los Angeles – Preprint 5269

[17] 2000 – "A Display Technique for Evaluating the Disturbance of Microphone Response Patterns"
by Chris Woolf and Oliver Prudden, Rycote Microphone Windshields, UK.
109th AES Convention, Los Angeles – Preprint 5180