Tuesday 19 May 2020

DIY Parabolic mic part II - the mic

Initial mic tests
I made initial tests of the rig with a few mics. Ideally, a mic should be sensitive with low self-noise while being light and small enough not to unbalance the dish. I had high hopes for my LineAudio CM3, one of the smallest and nicest sounding cardioids around, but it didn't sound right inside the tubular mount and with the xlr plug attached was still a touch big and heavy. The best was one of my EM172 'clippy' lav style mics which weighed nothing and fitted neatly in the tube even with its bumblebee furry on. However the cable was a bit short and I didn't buy a Sound Devices MixPre-3 recorder to just run mics on plug-in power. A bit of googling turned up a simple circuit to convert EM172s to run on 48 volt 'Phantom' power, and the possibility of wiring two or three in parallel to improve the signal to noise ratio.

triple capsule viewed through handle

Capsule circuitry and assembly
I bought three capsules from micbooster.com with short cables already soldered on (safer, as the capsules are easily killed by excessive heat when soldering). I didn't bother with matched capsules as I figured any differences would be evened out as the outputs are summed by the parallel wiring. Once the capsules arrived I rigged up a breadboard and 1meg potentiometer to get my head around the 'simple p48' circuit to step the 48 volts of phantom power provided by the recorder down to the right voltage for the capsules.

Although there is plenty of info around online this post is about the best summary, with plenty more info in the comments too:
From reading it seemed about 7.5-8V was ideal, but I couldn't find any info on resistor values with multiple capsules. Using the breadboard and pot it worked out that to deliver ~7.5V across three capsules I needed a 10K resistor. Using the the MixPre-3 to deliver the power while wearing headphones was instructive; at lower voltages I could hear the sensitivity drop off, whilst at 10V and above (I got up to about 14V at one point, not intentionally!) the capsules seemed to get hyper sensitive and shrill, with the noise floor jumping up.

simple p48 components

With the components sorted I could build the cable and the mic. The 4.7uF electrolytic capacitor and 10K resistor fit in an XLR plug, I then added a bit more work by adding a 3.5mm jack and socket halfway to sit in the parabolic handle and avoid always having a cable dangling from it. Alternatively, a potentially neater but heavier option would be to put the electronics in an XLR socket installed inside the handle. For the mic itself, I first covered each capsule in heatshrink and then used sticky copper sheet to both hold them together and shield them from interference. A bit of electrical tape and heatshrink then beefed it up.

Initial tests were very encouraging. I was concerned there might not be an obvious difference, but it was clear as day. With hindsight, I was combining three improvements in one hit. Firstly, providing 7.5V to each capsule meant a higher quality signal - improved sensitivity and lower noise. Then by running three capsules in parallel I was summing three of these improved signals. Finally, the p48 XLR conversion meant I was now using the class-leading Kashmir preamps on the MixPre-3. Coupled with the physical signal boost from the parabola the whole rig was starting to sound like a very capable instrument.

Mic body
The mic needs to be supported at the focus point of the dish with a clear 'view' of the dish. After trialling a couple of pieces of PVC plumbing pipes with various shaped apertures I bought a piece of aluminium tubing and laboriously drilled and filed it to the shape seen above - giving each capsule a big 'window'. The challenge is to have as little material as possible near the caspules while providing a strong support. The ends of the tube were then plugged with foam with a socket cut for the mic capsules. Layers of electrical tape shimmed the tube to a snug fit in the dish holder.

Wind protection
Rycote sell kits of their acoustic baffle and fur for DIY wind protection. Sewing the two layers into a simple tube was all that was required, with an o-ring to prevent it sliding off. The final item was a spandex cover from parabolicmicrophone.co.uk. This is a bit tight but hopefully the elastic will loosen over time. The combination is very effective and in use the limits are more to do with the wind physically shaking the dish.

Handling noise
In my experience parabolics are very sensitive to handling noise - the Telinga I've used was just as bad as mine. Even if the mic is well isolated from the handle, the smallest vibrations from the handle pass into the dish which then amplifies them like a loudspeaker cone. Soft foam on the handle helps but the best trick is finding a comfortable grip and not to move your hand at all - easier said than done. As generally with nature recording with sensitive handheld mics it also helps if you can breathe silently and control your tummy rumbles!

Sound quality
I've got a few mics in my drawer now, but this one is by far the best performer for bird song. On-axis sounds get a significant volume boost with off-axis sounds not unduly distorted. Compared to my Audio-Technica AT4073A shotgun mic I get a much stronger and more directional signal with lower noise, at the expense of some dish 'colour'. Even professional models are reputed to 'colour' the off-axis sound in this way as the waves are reflected off the dish, with Schoeps providing an eq plugin to help correct the sound. Without direct A-B comparisons I can't say if mine is any better or worse than commercial offerings but the colour is definitely detectable if you're listening for it.

Sample recordings - coming soon

Sunday 10 May 2020

DIY Parabolic microphone pt.1 - the dish

 Recording bird song and other natural sounds has boomed in popularity over the last few years, particularly in the recent period of Covid-19 lockdown with many having more leisure time and reduced noise pollution. For many novice recordists the first thing they seek out is the audio equivalent of a telephoto lens. Unfortunately that doesn't really exist. Sounds waves travel as modulations in pressure, making them so pervasive that all microphones therefore hear sound from all directions at once. This is why noise pollution is such a pain - you can't frame sounds out of shot like in a camera.
 on location at Stanage

Traditional directional microphones (aka 'Shotgun' or hyper-cardioids) work by housing the mic capsule in a long tube with many precisely positioned slots. These attenuate the off-axis signal by phase cancellation, leaving the on-axis sound waves unaffected. But because most are aimed at recording human voices within short range (one or two meters, typically held on a boom just out of shot) to pick up bird sounds well requires the most sensitive, low noise models which are not cheap.

The second option is to use a parabolic dish to focus the sound from one direction. The precise curve of the parabola bounces all the sound waves from one direction onto your centrally mounted mic. There are a couple of complications - the size of the dish affects both the amount of amplification (bigger is better), and the frequency range - the wavelength has to be smaller than the dish. Dishes sized to be portable - around 500mm diameter is popular - amplify typical bird song frequencies while lower frequencies like traffic hum are unaffected. The links below have more info.

A few years back I went on a sound recording workshop with Chris Watson and had the chance to use his professional parabolic mic.  The combination of a Telinga Universal dish and handle and Sennheiser MKH 8020 is reputed to be the ultimate rig for field recording. While it was way out of my budget at £1800 the dish and handle was surprisingly simple (the handle appeared to be from a drill) and it did look possible to have a stab at making my own.

I already own a hammer drill with an auxiliary handle and so, figuring I'd be unlikely to need both the drill and the mic at the same time, I was off to a good start. A trip to B&Q turned up a sheet of 25mm MDF and couple of plumbing parts that looked promising. Time to think about the dish.

The focal length can be short, with the focus well inside a deep dish, or long, like a satellite tv receiver with a flatter reflector and the focus well in front. There are some benefits to having the mic closer to the dish, such as wind protection and less phase-induced attenuation of certain frequencies. Most commercial mics share the classic '1/A' design where the dish is fairly open and the focus is level with the rim, creating a diameter four times the depth.

This template should print at the correct size (scale to 100%) or make your own at https://www.desmos.com/calculator/htfdxbgqdt

50cm, 1/A template

Dish material
Commercial dishes are made from either Polycarbonate (Telinga) or PETG (Dodotronic). Both of these can be easily obtained in thin sheet (ebay) and thermomoulded. DIY projects often use fibreglass, which is much heavier and said to colour the sound. I went for 0.8mm PETG as the moulding temperature is lower (90-160oC, depending on relief). Polycarbonate requires higher temperatures, and baking for many hours before moulding is also recommended to remove any moisture. Whether this is for solely cosmetic reasons I'm not sure, but it sounded beyond my DIY resources. The disadvantage of PETG is it a bit less robust and the lower moulding temps make it more liable to distortion if accidentally exposed to heat. In the Uk that isn't likely to be an issue but storing it in the car in summer will probably not be a good idea.

Making a mould
This is likely to be the stopper for most DIYers. Unless you have access to a lathe you are going to struggle. Luckily my Dad is a keen woodworker and has a lathe in his shed, and before finalising the design a quick call confirmed the size and weight he could handle. I made the mould from layers of 25mm MDF, five layers giving 125mm depth which in the standard 1/A shape gave a diameter of 4X the depth = 500mm. The calculated diameters for each layer were drawn on and cut out with a bandsaw. To reduce weight we also cut the middle out of the three biggest layers, and then aligned and glued all five together with a few dowels for added strength. A final MDF ring was then cut to hold the sheet being moulded, with an inner diameter of roughly 510mm. With a profile template and my Dad's skill we soon had the shape tolerably close. I later did some work with filler and sandpaper to tweak it further.

 finished mould ready for forming

Ideally, thermomoulding would be done with an oven and a vacuum former. While these can be made at home - using patio heaters and aluminum foil to make big ovens, and vacuum cleaners etc - it involves a lot of work and expense for a single dish. I went for a more redneck approach using our gas hob and some elbow grease. I first cut a 600mm circle from the PETG sheet with a stanley knife, removed the protective films and attached the PETG disc to the MDF ring with a staple gun. This I held a foot or so above the gas hob, keeping it moving in the hope of heating evenly. The first attempt went badly - I didn't get it hot enough to mould to the full depth. Then it was harder to heat as it wasn't flat and the PETG got too hot. You'll know when it gets too hot as it stretches easily, even under it's own weight, quickly getting paper thin and crinkly and only good for the bin. I learnt my lesson and round two went better, pre-heating the mould so the sheet didn't cool so fast, and using a heat gun to further warm the perimeter for the final section. However the edge was still a bit thin, and I cut it too short making for a dish which looked ok but was a bit floppy. By the third attempt I actually ended up with something which looked good and held its shape well. If I was to try again I'd make a thicker MDF outer ring to hold the sheet without bending.

 Dish cooling after moulding - note staples

After leaving a few minutes to cool, I pinned the centre of the dish to the mould and cut the excess plastic from around the perimeter, then sanded it smooth. Then it was just a case of carefully cutting a 40mm hole in the centre, screwing the handle and pipe-fitting together, and suddenly it looked like the real thing.

 cutting the centre hole

For initial tests I used a stereo pair of Primo EM-172 from micboosters.com. These small omni mics have low self-noise and are cheap. With the right mic at the focus and the left mic kept outside the dish it was simple to compare directionality and gain whilst recording, and in also by switching channels check for any 'colour' introduced by the dish and/or mic mount. It sounded promising, so I proceeded with building a bespoke mic - see section II.


Thorough blog, the main source of information for this project

Good overview from Klaus of Telinga

Discussion of stereo parabolics

A bit dated, but a detailed technical paper (pdf)

Commercial models:
https://parabolicmicrophone.co.uk/ Cheap, basic UK model
http://www.telinga.com The professional choice
http://www.dodotronic.com Mid-priced offering
http://www.wildtronics.com US-based manufacturer (also has some good info in the articles section)