The focal point of our residency was a floating rig with a pair of hydrophones suspended below the surface of the water, capturing the sounds of lifeforms inside the Cosmic Pond.
The rig became known as Terraqueous II, in tribute to Carl Sagan’s repeated reference to terraqueous planets in his 1980s Cosmos television series. The word had stuck with us while developing our residency proposal, and because we were planning to explore Earth-bound water it seemed appropriate.
Our plans for Terraqueous I didn’t get much further than the back of an envelope, and featured a washing up bowl as a floatation aid. The conclusion to use a lifebuoy made far more sense aesthetically and practically. Hence, Terraqueous II.
Unlike a washing up bowl, a lifebuoy is a device associated with human survival at the threshold of an uninhabitable space. It’s a lower tech equivalent of a space suite or scuba diving kit, vital in supporting life at the edge of the “beyond”.
We decided to use a pair of hydrophones to capture a stereo sound image of life in the pond. To achieve a similar localised, spatial listening experience to that of human hearing on land, it was important to get the spacing between them correct. Taking the speed of sound in water, 1481m/s, as opposed to 343m/s in air into account, we estimated the hydrophones needed to be c.82cm apart:
- Width of average adult head, w = 19cm
- Speed of sound in air, c = 343m/s
- Ratio, r = w/c = 19/343 = 0.055
- Speed of sound in water, c = 1481m/s
- Ratio, r = 0.055
- Distance between hydrophones = c*r = 1481*0.055 = 82cm
A spaced pair of hydrophones would theoretically create a convincing stereo effect. Although precise distances of the insects from the hydrophones might be impossible to measure, relative distances and communication patterns of insects singing in the pond could be perceived.
The standard diameter of a lifebuoy is 30 inches (76.5cm), so the hydrophones needed to be suspended beyond its outer edge. A plank longer than the diameter of the ring was lashed to the buoy, and holes drilled through to allow the hydrophone cables to pass beneath. Long carriage bolts were attached to the top of the plank enabling cable winding and height adjustment of hydrophones.
Alan from ACA advised us to attach an anchor to the rig. Even the slightest breeze over the pond could move surface objects by considerable distances, and we had to avoid our rig getting blown into weeds or the bank. Alan kindly made us an anchor from a found piece of old iron and a lump of concrete, which did the job perfectly.
The JrF hydrophones have an unbalanced audio output, which is potentially problematic over a long cable run. Electrical noise can be induced in the cable, interfering with the audio signal. This was a concern as our rig would be connected by c.50 meters of submerged audio cable to the shore.
The hydrophones were isolated from the cable using two Neutrik NTE1 audio transformers to help minimise the risk of noise. These would convert the hydrophones’ unbalanced audio outputs to differential output signals, and so reduce the risk of electrical noise entering the long cable. The transformer outputs were connected in parallel with a Zobel network consisting of a 680pF capacitor and 150Ω resistor to dampen any ringing effect that can sometimes be induced in audio transformers (figure 1).
[On reflection, it may have been more appropriate to use transformers with a higher primary to secondary ratio to better match the high impedance hydrophone with the low impedance preamp input. However, the Neutrik NTE1s with 200Ω windings in a ratio of 1:1 proved perfectly adequate for our installation.]
The circuit board and connections had to be kept dry. We used a WAGO junction box, waterproof to IP68 standard; i.e. it could withstand complete immersion in water should that happen. The box allows three cables to be inserted, which suited our needs perfectly; hydrophones 1 and 2 in, and the differential audio signal out through one 5-core microphone cable to the shore. The junction box was securely slotted into three holes drilled through the plank, secured by its three connector glands as pegs.
The audio signal from rig to shore ran through c.50 meters of 5-core microphone cable. We did consider making the the rig completely autonomous, with solar charged battery and radio microphone link to shore, but decided a hard wired cable solution was most appropriate. Not only in terms of cost and assembly time, but the potential unreliability and lower sound quality of a radio link – particularly when listening to such quiet material. A Bluetooth audio link was also considered, but we didn’t want to rely on relatively unknown (to us) technology and compromised fidelity.
A lack of reliable broadband internet at ACA, typical of much of rural England in 2018, prohibited us from streaming audio from the rig via the Internet. That would also have allowed us to livestream audio from the pond via a web platform such as Locus Sonus – although we did manage to do that via a mobile phone with limited success at the final weekend event. It’s very likely that a future rig will be completely wireless if possible.
For launch and retrieval of the rig, it was necessary to attach and detach the cable easily. This called for a waterproof audio connector, but a reliable 5-pin XLR proved hard to find. Instead, we used a smaller waterproof 5-pin DIN connector on the rig.
At the shore end, the cable was connected to a small mixing desk in the observatory warm room. This provided sufficient gain for our speaker installation. We also recorded the audio from the hydrophones on a Sound Devices 633 audio recorder.
The audio signal from the hydrophones provided a discernible sense of stereo localisation, and recognisably distinct insects could be heard singing in the pond. However, the sound was accompanied by considerable electrical noise when connected to any device with an electrical earth, such as the amplifier for our speaker installation.
Although Allenheads is one of England’s most isolated villages, a power cable passes directly over the pond and this was interfering with audio signal from the hydrophones. Changes in the timbre of the noise synchronised with fluctuating demand on local power usage. This gave an interesting account of human activity in the area, although it’s not what we intended to capture.
With the cable connected with unearthed equipment, the noise wasn’t an issue. Recordings made with battery powered devices were noise free, but we still needed to fully balance the connection electrically. The differential signal in the long cable wasn’t enough. More reliable noise reduction would be achieved by connecting the common shield of the cable to earth at the rig end, increasing the common-mode rejection ratio of our circuit.
Terraqueous II v1.0 was launched in late May 2018, and left for just over a month drifting slowly around a small area of the pond as far as its anchor would allow. It continuously relayed the subaquatic sound of the pond to the warm room for whoever was in there to listen to on headphones. It was still intact, watertight and working, albeit a little weather-beaten, after more than a month.
Returning for the second part of the residency in July, we completely eliminated the noise by sinking an earthing rod (copper pipe) into the pond bed, connected to the common shield in the cable. The upgrade gave us Terraqueous II version 1.1, the final iteration of the rig which was relaunched for our closing weekend event.
The results from Terraqueous II have been very encouraging, in terms of the underwater sounds captured with the hydrophones and pubic reaction to the concept and installation. We recorded several days of continuous activity inside the pond, and those recordings have already been used in subsequent artwork.
Acoustic ecologists will certainly be interested in our recorded material, and we’re happy share the experience and materials, or collaborate with freshwater fauna listening projects using spaced pairs of hydrophones. We look forward to launching Terraqueous III sometime in the near future. Please get in touch with us here.