DIY stereo parabola

Martin, R., Honold, J. (2016) DIY stereo parabola. https://avesrares.wordpress.com/2016/04/28/diy-parabola/

Ever since we published our recommendations of sound recorders for birders we have been receiving questions on how to further improve sound quality. Coming from a mobile recorder, there are two main options, namely shotgun microphones and parabolic microphones. Shotgun microphone prices start from 350€ (e.g. Sennheiser ME66 + K6). They have a convenient size for handling and transport but they do not offer true amplification but rather block/weaken incoming sound waves from the sides through interference. In contrast, a parabolic microphone truly amplifies sound by concentrating sound waves in a focal point. Furthermore, it is much more directional. It is, however, unwieldy and expensive: until recently, prices of commercial kits started around 1000€ (stereo). The best known producers of parabolas are Telinga, Dodotronics and Wildtronics.

We wanted to try building our own parabola: in the last 2 years we spent a lot of time developing a parabola recording setup and discussing the design with other recordists. We went through different versions and kept on optimising the setup to our liking. Meanwhile Dodotronic has started selling one of our setups commercially.

In this post we want to share our experiences to help you build your own parabolic microphone. Several friends have already successfully copied the design and so can you with little money and time.

Figure 1: The finished parabolic microphone with an Olympus LS-3 mounted on the handle

Expenses and effort

Building your own parabola including a mic will set you back about 100€, given you already have the tools: a soldering iron, a jigsaw and some glue. No special skills are required and time and effort required are low, too (that is, if you don’t screw up). The hardest part will be forming the dish itself. For this, you need further tools like a drilling machine and a big oven. If you don’t have, it is certainly possible to ask a workshop for help. Otherwise you can simply buy one and only do the mic yourself.

The microphone

The microphone is the main part of the recording setup. In theory you could use just about every recording device you can think of in the parabola. Building a dedicated mic setup for the dish will be much more efficient in terms of handling and sound quality though.

Our setup is made up of four Primo EM-172 which cost ~10€ each. They are available from Frogloggers (USA) (BE CAREFUL: we got a feedback, that an order in September wasn’t delivered, they didn’t answer to mails and the money was lost) or Micbooster (UK). The ones sold by Frogloggers have an „acoustically transparent water-seal“ which might be useful (beware, it does not make it a hydrophone). Micbooster also offers a high quality plug with cable.  We use 30ppi foam for wind protection which is usually used in aquaristics (check ebay). A thin sheet of fleece covering the front side of the dish gives further wind protection.

Microphone setup – theory

Let’s cover the theoretical basis first. The aim is a parabola microphone for stereo recordings. (For the sake of clarity: The bird sounds you focus your parabola on will not have a stereo effect in the recording. That is because the amplification and therefore volume will be the same in both channels, given you aimed the dish correctly. The stereo effect instead stems entirely from the differences in background sounds between the left and right channel. See figure 2 &3 below)

Figure 2: A parabolic microphone records the background sounds in stereo (yellow = left channel, turquiose = right channel). The  volume of the focused sound is the same in both channels and therefore lacks stereo impression.

Listen to this Barred Warbler, recorded with the above mentioned setup, as an example of a stereo recording:

Barred Warbler (Sylvia nisoria), 03.06.2014, Altai, Kazakhstan: There is a Thrush Nightingale (Luscinia luscinia) to the left and another one to the right. A Corncrake (Crex crex) starts singing at 5s from the right. Listen for the Hoopoe (Upupa epops, the calls of a Red-backed Shrike (Lanius collurio) (right), a Common Cuckoo (Cuculus canorus) (24s onwards, left) and a fly-by Yellowhammer x Pine Bunting hybrid (Emberiza citrinella x E. leucocephalos) (1:oo min). At about 1:03, the Barred Warbler starts a song flight and you can hear the clapping wings.

The sound waves need to be divided between the left and right channel to get the stereo effect. A disk does the job and splits the amplifying area of the dish in two halves. As a negative side effect of this splitting each channel only receives half of the sound waves. That means without additional measures you would lose half the amplification (see figure 3).

Figure 3: The disk separates the dish in two halves.

Now a physical effect used by a special sort of microphone, the boundary microphones or pressure zone microphones, comes into play: If sound waves hit a non-absorptive boundary, the signal of the direct and reflected signal add up and double sound pressure, thereby increasing the peak gain by 6dB [1]. This pretty much makes up for the loss due to the split by the disk. But this amplification starts only above the critical frequency which depends on the size of the disk [QQQ] (see figure 4). In contrast to the parabolic dish, however, amplification is constant over frequency [2].

Figure 4: The minimum frequency of amplificaition of the boundary microphone effect depends on the size of the disk. If you don’t match the size of the disk and the parabola there will be two frequencies where the amplification increases strongly.

 

To make use of this effect, Telinga builds a stereo setup with microphone capsules facing the disk [3]. There is only a small gap between the capsule and the disk which is effectively the PZM system. On one hand this means that it records less noise from the sides and is less prone to wind disturbances. On the other hand the sound waves cannot reach the microphone capsules directly (cf. figure 5) [1], the stereo effect is less pronounced.

 

Figure 5: Different positionings of the microphone capsules and their effect. Left: Microphone capsules facing towards the disk. Right: Microphone capsules within the disk and facing outwards. Blue lines are sound waves, red lines are the microphone diaphragms.

 Our first setup therefore had outwards facing microphone capsules which were integrated in the disk (take care that the microphone diaphragm is at the same level as the surface of the disc). This results in a stereo parabola that rivals the amplification of a classic mono parabola (own observations). The stereo effect is nice due to the outwards facing capsules but wind can be problematic (turning the disk parallel to the the wind helps). See „The microphone“ for measures against wind rumble. The described microphone setup has been adopted by Dodotronics in their stero parabola (available here).

If you would like to take recordings that are focused on one individual but still give a  good impression of the surroundings in stereo, this is the setup for you. (Keep in mind: If you would like your recordings to match the human hearing as well as possible, a parabolic microphone is not the way to go and you should consider a SASS. But this is an entirely different approach and we won’t go into detail for this type of microphone. See here and here for further, especially DIY information on SASS).

After some time with this setup we thought up a different design to tackle the wind problem and increase the amplification by a further 6dB [4]. This required adding another disk to the setup. The additional amplification comes at a cost though: the stereo effect is much weaker. Use this setup if you want to isolate the sound in focus, e.g. for analysing the recording, but still with an audible stereo effect.

Figure 6: Both setups in comparison. Left: One disk with capsules that face outwards. Right: Two disks, with capsules facing towards the parabola (here downwards). For sizes of the disks see figure 4. The disk is rectangular here but it might as well be round.

The microphone setup – construction

We used 3mm thick plexiglas sheets for the microphone setup. They are cheap (search ebay for „plexiglas sheet“) and easy to shape. All you need is superglue, a jigsaw and a drilling machine.

To get the desired stereo effect we need at least two mic capsules (left and right channel). Wiring capsules in parallel improves the signal-to-noise ratio. We chose to use two Primo EM-172 mic capsules per channel which achieve the described improvement regarding noise while still being able to run on plug-in-power provided by the recorder (at least the Olympus LS-3, LS-5, LS-10, LS-11 as well as the Sony M10 provide enough PIP, so no external power source is needed). The capsules of each channel must be as close to each other as possible to avoid interference through phase shifts.

Figure 7: Wiring of the capsules to the plug

 

The parabola – theory

The parabolic dish is responsible for the gain of the signal. Its shape follows a parabola, mathematically speaking. Transferring the 2-D parabola to 3-D makes it a paraboloid. For sound recording two characteristics of a paraboloid come in handy: 1. sound waves reflected by a parabolic shaped object are reflected and meet in one point, the focus. 2. The distance sound waves travel from the source to the focus is always the same, no matter where it gets reflected on the parabola (-> no phase shifts). The shape of the parabola is defined by its focal length (diameter does not affect the shape). A longer focal length increases the distance between dish and microphone, thereby increasing exposure to wind and overall size of the setup. In turn, the longer the focal length, the lower the distortion [5]. You can use the attached Excel-file to calculate the shape of your parabola depending on your desired focal length (keep in mind to use the same scale on both axes in the file to see the right shape).

A parabolic dish has special characteristics concerning the amplification of sound waves: the amount of amplification depends on the frequency of the sound waves as well as the diameter of the dish. Let’s consider the influence of the dish diameter first.

Formula 1: Formula for the calculation of the amplification G of a parabolic dish with diameter d, frequency f, sound velocity c and efficiency eA

Given the formula we can now determine the minimum frequency f_m at which a parabola gives an amplification greater than 1. The larger the diameter, the lower the minimum frequency of amplification (see figure 8).

Figure 8: Minimum frequency of amplification as a function of dish diameter. To receive the minimum frequency f_m, we set G=1, c=343 m/s and eA = 0.5 (which is a realistic value for a normal parabola, cf. [5].

So the amplification of your parabola will start at this frequency. Below this there is no gain. This does not mean that you cannot record sounds below this minimum frequency; they will only not be amplified.

Given a fixed diameter, the amplification is still not independent of frequency. The amplification increases with increasing frequencies (see figure 9).

Figure 9: Theoretical amplification of parabolic dishes with c = 343 m/s and eA of 0.5. Different colors correspond to different dish diameters d: red = 60cm, orange = 50cm, pink = 40cm, blue = 30cm, black = 20cm. The dashed line represents a gain of 1.

Figure 10: Logarithmic representation of the gain as a function of frequency. The logarithm makes the distortion look less bad than figure 9. Again colors correspond to different dish diameters: red = 60cm, orange = 50cm, pink = 40cm, blue = 30cm, black = 20cm.

This relatively stronger amplification of higher frequencies gives a parabola its typical sound. Usually, the higher frequencies of calls and songs don’t carry far. A parabola can save these frequencies even from a distance, making birds sound closer than they really are. On the other hand this emphasis on high frequency amplification can be perceived as harsh, especially in very close recordings.

If you want to attenuate this effect, you can try displacing the microphone a bit away from the focus point. This attenuates the higher frequencies stronger than the lower frequencies [6]. Alternatively, you can use an equalizer which further attenuates noise in the higher frequencies.

The parabolic dish – production

This will probably be the hardest part. You can find bowls that approximately have the shape of a parabola. Or you can buy one from Telinga or Dodotronics but they come at a price. We wanted to build one ourselves.

We decided to use a diameter of 37cm. This was the maximum size we could manufacture easily and which we still considered handy and portable. To save weight we decided to use plastic. With plexiglas you can get a transparent dish which is convenient for aiming.

At first, we produced an original mould from wood. If you use a turning lathe, you can shape it with an accuracy of more than 1/10mm. Next you need a wooden ring with the diameter of the mould + twice the thickness of the plastic (here 2 x 0.8mm) to which you attach the plastic sheet with screws. We used plastic sheets from the local hardware store with a width of 1m (usually used to cover up broken windows, sold by meter).

Figure 11: Original mould of the parabolic dish (below) and the wooden ring with the attached plexiglas sheet. After heating the sheet together with the plastic can be deep-drawn over the mould.

Figure 12: The wooden original mould for the parabolic dish. Pores and cracks were filled up with a surfacer. The red line marks the border at which the deep-drawing is stopped.

Now you can deep-draw the plastic. To do so, mount it on the ring, put it on top of the mould and place it altogether in the oven (see figure 11). First you need to find out the right temperature; for the material we used it was 90°C. Too much heat „burns“ the material and blurs it – if temperatures are too low the plastic will not be moldable (Alternatively a gas burner does the job, but it is almost impossible to heat the plastic evenly). Once the right temperature is reached, the ring will be pulled with some force above the mould. After a few seconds of cooling out it’s done. Now all you need to do is cut out the parabola. This is a standard deep-drawing procedure (a method also used by Dodotronics).

If the plastic sheet is thin enough, the dish can be rolled up for transport. It was possible for example with a thickness of 0.8mm with the material we used.

Mounting the microphone and the handle

The handle of a drilling machine makes a comfortable handle (Lukas P. and Steve K. had the idea to use the handlebar grip of a bicycle). There are several options for mounting the microphone to the handle. One of the simplest is to use a threaded bar or screw to which you attach the microphone on one side, lead it through the dish with flat washers on both sides and a counterthread incorporated in the handle. The cable can either be led around the dish or through a small hole in the dish.

Figure 13: A simple mounting system for a parabolic dish using a threaded bar (A), two screw nuts (B and E), and two flat washers (C and D). Left: mounted, right: exploded view

A stabler, yet not a lot more difficult construction involves (angled) plastic tubes and plastic washers from the hardware store. Another advantage of this mount is that you can lead the cable through the tube.

Figure 14: Mount for a parabolic dish made from plastic tubes from the hardware store. Left: mounted, right: exploded view: rubberised bicycle handle (A), angled tube with thread (B), washers (C and D), thermoformed plastic tube (E). The cable can be led through the tube.

A major problem is handling noise that is transferred from your hand through the handle to the microphone itself. There are two points where you can approach this problem:

• the connection between the handle and the dish mount. Using some rubber in between can attenuate handling noise
• the connection between the dish mount and microphone mount. Figure 15 shows an example how to face this problem.

These measures help a lot in diminishing handling noise.

Figure 15: An example of a free-hanging microphone. The microphone is attached to a wire (A) by rubber bands (B). A piece of rubber at (C) can also help attenuate handling noise.

It is very handy to mount the recording device right on the handle (see figure 1). This allows you to have one hand free for the binoculars.

Figure 15: The finished parabolic microphone seen from the side. A piece of foam is used as wind protection.

The proof is in the pudding they say, so listen to some more samples taken with these setups:

Willow Tit (Poecile montanus montanus): 13.06.2014, 05:05, Hoher Ifen, Austria, male, song. Background: Common Cuckoo, Common Chaffinch, Common Crossbill, Eurasian Wren, Blackcap, Great Tit, Blackbird, Coal Tit, Song Thrush. © J. Honold

Aquatic Warbler (Acrocephalus paludicola): 30.04.2014, 20:00, Biebrizwa, Poland; male, song. Background: Common Cuckoo, another singing Aquatic Warbler, a displaying Common Snipe, Reed Bunting, Savi’s Warbler. © R. Martin

Siberian Rubythroat (Luscinia calliope): 30.05.2014, 20:18, Altai, Kazakhstan; male, song with many imitations. Background: a mountain stream, Siberian Chiffchaff, Common Cuckoo, Goldcrest, Greenish Warbler, Black-throated Thrush, Coal Tit. © R. Martin

Atlas Chaffinch (Fringilla coelebs africana): 03.03.2014, 06:45, Oukaimeden, Morocco, male, song. Background: Common Firecrest, Coal Tit, Alpine Chough, Atlas Chaffinch, Common Blackbird. © J. Honold

White-backed Woodpecker (Dendrocopos leucotos leucotos): 13.03.2015, 11:03, Rottachberg, Germany, male, calls. Background: Coal Tit, Marsh Tit, Blue Tit, Mistle Thrush, Common Chaffinch, Common Raven, Brambling. © J. Honold

Lilford’s Woodpecker (Dendrocopus leucotos lilfordi) and Great Spotted Woodpecker (Dendrocopus major major): 15.09.2014, 10:00, Durmitor National Park, Montenegro. Background: Eurasian Wren, Jay, Coal Tit. © R. Martin

Temminck’s Lark (Eremophila bilopha): 05.03.2014, 07:06, Guelmim, Morocco, male, song. Background: Greater Short-toed Lark, Red-rumped Wheatear. © R. Martin

And now enjoy building and recording with your own parabola!

Many thanks to Marco Pesente and Vicky Powys for discussing the microphones and the design.

References

[1] http://www.sengpielaudio.com/TwoDifferentBoundaryMicrophones.pdf

[2] http://www.sengpielaudio.com/UntereGrenzfrequenzbeimGrenzflaechenmikrofon.pdf

[3] http://www.naturesongs.com/parabola.html

[4] https://pantherfile.uwm.edu/type/www/audio-reports/BoundaryMicExperiments/BoundaryMicsStudy/BoundaryMicsStudy.htm

[5] http://www.w1ghz.org/antbook/chap4.pdf

[6] http://www.wildlife-sound.org/equipment/stereo_parabol/index.html