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

8.2 - DIRECTIVITY

8.2.8 - HIGHLY DIRECTIONAL MICROPHONE SYSTEMS

The Parabola

The parabolic reflector, contrary to the rifle mic, is almost always sold as an accessory, and can be mounted with practically any suitable microphone. It has one big disadvantage in that it is certainly more cumbersome than a rifle mic, being about 40cm to 50cm in diameter. However the big advantage is an enormous increase in gain in the higher frequencies (about 20dB at 10kHz). The price to pay for this gain is that the frequency response is by no means flat, it has a 6dB/ octave upward slope throughout the whole of the frequency range up to about 10kHz. For sound effects and bird song etc, this response curve is without importance and indeed the high gain of the system is a definite advantage. For speech pick-up where intelligibility is important this 6dB/octave slope again is no problem. But for quality speech recording, equalisation has to be used to get a natural sound, and this may introduce unwanted side effects such as increasing the low frequency level of ambient sound – but at least we are able to choose the best compromise equalisation/ambient sound rejection.

How does it work? Again back to basics - all sound arriving on axis of the parabola will be reflected towards a single point called the focus - this is normally where we place the microphone as shown in Figure 66.


Figure 66 - The Parabolic Reflector

Again the best demonstration of this system is using small wave propagation in a water bath or with computer animation. The lower frequency limit of the parabola is when the diameter of the parabola is smaller than the half wavelength – at this limit the system becomes more and more omnidirectional. To compensate for the directivity at low frequencies it is possible simply to roll-off the response with a high-pass filter. Any type of microphone can be used for sound pick-up. Figure 67 shows the frequency response using an omnidirectional microphone and a cardioid microphone. The cardioid has a reasonably smooth response with a 6dB/octave slope, whereas the omnidirectional microphone shows typical comb filtering response – the microphone sums the direct incident wave and the reflected wave - out-of-phase conditions produce dips in response, in-phase conditions reinforce the sound level.


Figure 67 - cardioid and omnidirectional microphone
with and without a parabolic reflector

Even though these measurements were made before the days of computer assisted display, I think it is evident that not only is the frequency response curve smoother with the combination of a cardioid microphone and the parabolic reflector, but low frequency rejection is also much better compared with an omnidirectional microphone. This is not just the effect of the parabola but also due to the roll-of in low frequencies with a pressure-gradient microphone. Some high frequency roll-off is also present probably due to the fact that the microphone was not exactly at the focal point of the parabola. Indeed deliberately placing the microphone outside the focal point is also a technique used to equalise the frequency response in the high frequencies when a more natural sound is required. You can produce varying degrees of high frequency roll-off as the microphone is positioned further and further away (a few centimetres) from the focus. But this is obviously at the expense of high frequency gain.

Another partial equalisation technique is sometimes used where the inner surface of the parabola is covered with a certain thickness of absorbing material such as felt. This surface of felt can also help to reduce handling noise by partially damping transmission of sound along the surface of the parabola. The problem of microphone suspension systems in general will be treated in the next section. Handling noise is slightly different to microphone suspension in that the sound of manipulation of the parabola is radiated by the surface of the reflector itself. Some form of insulation between the surface of the parabola and the hand-holds is necessary – the simplest being to suspend the parabola shell with elastic suspension attached to an annular tube or hoop as shown in Figure 68. Wind noise on the surface of the parabola is on the other hand very difficult to eliminate without completely enclosing it in a windshield!


Figure 68 - elastic suspension of parabola shell with microphone clip arm

There are a few commercially available parabolic reflectors which are made out of transparent plastic to enable alignment with the sound source by looking through the reflector. However the sound pickup alignment is very critical and some form of ‘gun- sight’ is preferable.

Here is a do-it-yourself anecdote for those who want to try and build their own parabolic reflector as shown in Figure 68. The main problem to solve is how to build the mould in the form of a parabola. My original approach was to mould a lens form as near as possible to the shape of a parabola, and to sand-off any excess material (which in the end proved unnecessary). You will need a 200 litre metal oil drum which is closed with a lid – when the lid is removed the remaining open drum top is usually reinforced with a rigid tubular structure. The second and more difficult item to be found is a sheet of approx 1mm thick pure rubber – I used a rubber sheet which is normally supplied to hospitals as an bed-sheet. It is important that the rubber be pure with no additives so the elasticity is not reduced. The next item is about 10kg of ‘Plaster of Paris’, and some wire or string.

Stretch the rubber over the top of the drum, but not too tightly, just flat - the fixing with the wire or string of the rubber round the drum must be very tight. Mix the plaster into a liquid and QUICKLY pour the plaster onto the surface of the rubber (plaster sets very quickly, but dries very slowly). It would be advisable to experiment with a small quantity to get the right consistency – the plaster must pour like a liquid, but excessive water will make the plaster fragile and will prolong the drying process. Under the weight of the plaster, the rubber will sink leaving more space to continue pouring. Fill the remaining space until the plaster is level with the edge of the drum. If you have not fixed the rubber well enough you will end up with a mess in the bottom of the drum (as happened to me - you feel a bit foolish!). Leave the plaster to dry for a long time – after a few days you will be able to remove the parabola lens, and after a month or two the plaster should be completely dry. Imagine my surprise when I found that the shape of the moulded ‘lens’ was exactly a parabola – no sanding was needed - the form produced was probably nearer to a caternary curve which is itself very close to a parabola. You can cut out a template to check the profile yourself.

When dry (absolutely dry), cover the surface with beeswax or equivalent (used in the manufacture of fibre-glass boats as a release agent) – I made my own boat out of fibre- glass around this period. The next stage, as with making a fibreglass boat, is to put on the first gel-coat layer. When this is hard, mould on a thickness of about 5mm of fibre- glass – the same as you would use to repair the fibre-glass body-work of a car. The best is to leave the fibre-glass to harden in bright sunlight. The shell will then split from the plaster mould without needing any other manipulation. This is the moment when you realize whether you have applied enough beeswax or not! If not, back to square one and start again. If strong sunlight is not available, then you can always force slats of wood in between the mould and the fibre-glass until it lifts off – careful not to damage the mould if you want to mould a number of other shells. With various student groups, I manufactured quite a number of parabola shells in the ‘old days’! I leave the construction of the microphone clip system to your imagination – it is not quite as easy as it sounds.

Sighting is best achieved with a 20mm tube moulded through the surface of the shell and aligned along the axis. This can be done at a later stage when the system is working, picking up some high frequency source at a distance and then moulding the tube in line with the source by sighting through the tube. I think you will realise by now that the parabola is one of my favourite hobby-horses (called ‘Violon d’Ingres’ in French) – it has certainly proved its usefulness on many a outside film set. When filming the best approach is to use a general ambiance mic on a boom arm and train the parabola mic onto any important action sounds – this adds dynamics to the sound effects track. Dialogue needs quite heavy equalisation for it to be acceptable. Once we even went to the extreme of mounting two parabolas at the ends of a 2 metre tube, slightly angled outwards, to do some experimental stereo sound recording of birds!