To test this feature, visit your live site.

Use Audiomoth with an external hydrophone

in Configuration Support

Is this hydrophone H2d Hydrophone (aquarianaudio.com ) compatible with Audiomoth v 1.2.0?

31 comments

31 Comments

Mats Amundin

Apr 29

Oki, that explains it! I am always using dB, so I just assumed it was dB. Thanks for the fast clarifications!!

Mats Amundin

Apr 29

OKi, got it! :-) The gain steps in the AM make me bewildered. You say the Low-Medium-High should give 4.3dB-15dB and 30dB, respectively. However, Hill et al. 2019 AudioMoth_A low-cost acoustic device for monitoring biodiversity and the environment, report much lower gain steps: 27.2dB-30.6dB and 32dB, respectively. And Lapp et al. 2023 Quantitative Evaluation of the Performance of the Low-Cost AudioMoth Acoustic Recording Unit, report -14.2dB for Low and -5.7dB for Medium, relative to the High setting. What are the correct values..? And yet anotherquestion: Jay Barlow told me that the bit depth at 384kS/s was only 12, compared to 16 with the other settings. Is this correct,and if so, what are the consequences for the recordings? Thanks!!

Alex Rogers

Apr 29

Replying to

The first preamp has a fixed amplitude gain of 11 and then there is a two-stage amplifier with software-controlled gain as per the table below (from https://github.com/OpenAcousticDevices/Application-Notes/blob/master/Using_HydroMoth_to_Make_Underwater_Recordings/Using_HydroMoth_to_Make_Underwater_Recordings.pdf).

These are direct amplitude ratios, not dB. We try to avoid using dB, as in our experience, only electrical engineers understand them, and even then it is easy to mix power and amplitude. The ADC is 12-bit at 384kHz and is oversampled at lower frequencies - 13-bit at 192kHz, 14-bit at 96kHz, up to a maximum of 16-bit. The software removes any DC bias in the final ADC and amplifies the signal so that a full 16-bit range of -32768 to 32767 corresponds to -1.25 to 1.25V at the ADC input regardless of sample rate.

Version 1.1.0 of the AudioMoth hardware omitted the first x 11 amplifier as the MEMS microphone itself had a higher gain.

Edited

Mats Amundin

Apr 27

As I have described in an earlier post, I am working on an underwater setup for the Audiomoth, based on an earlier porpoise detector housing. It has a piezo-ceramic hydrophone with a frequency response suitable for recording dolphin and porpoise sonar clicks. I followed advice from Jay Barlow at NOAH and put a 12 kOhm resistor between signal and ground and connected the external jack detector to ground. I use an Aquarian Scientific PA-4 preamp, and it works fine in the high frequency range at lower gain levels but loses 11 dB from 30kHz to 150kHz. Of course this can be compensated for in the post-process, but it affects detection of weak clicks due to lower S/N ratio. I read in the Forum that you were working on a new external preamp for the Audiomoth – is there now a suitable one for my purpose? If so, what are the dimensions – the space in the housing is limited (diameter 67mm and length 80mm; most of this space is taken by the AM and an extra 3xAA battery casing).

Another possible issue is a suspicious “wrapping down” in the spectrogram. I recorded a pinger, with main frequency output in 50-85kHz range, but with some components between 100-120kHz, at a sampling rate of 250kS/s. The top spectrogram below is this AM recording, and the one below the same pinger (there is some variation in the pings) recorded with a Reason hydrophone using a NI USB-6351 A/D converter sampling at 333kS/s, and a custom-made software on my laptop.

As you can see, there are down-sweeping tones in the AM recording which are not there in the laptop recording. Is this an aliasing effect? As far as I have understood there is no anti-aliasing filter in the AM, right? I tried to remove this by applying a lowpass filter, with cutoff at 150kHz, but this did not work, possibly because it is not inserted before digitalization.

Any suggestions on how to deal with this?

Thanks for any advice!

mats

Alex Rogers

Apr 27

Replying to

We're quite a way from having a final design for the external preamp and are addressing freshwater applications with low-cost pieze hydrophones first so there won't be a product in the immediate future. Check the voltage that the AudioMoth input ends up sitting at when the hydrophone is connected. The first stage of the AudioMoth input applies a gain of x11 but does no DC removal so you need to be sure that - 0 <= bias_voltage -/- (signal_voltage x 11) <= 3.0 - to avoid the risk of the input signal clipping.

The wrapping around is the harmonics of the signal aliasing back to lower frequencies. For normal recordings, AudioMoth samples at 384kHz and then averages multiple samples to produce lower sample rates. This is effectively an anti-aliasing filter but it is only first-order and has a weak cut-off compared to a hardware anti-aliasing filter before the ADC. You would need to apply a low-pass filter on the front-end to remove the aliasing.

Mats Amundin

Apr 29

Replying to

Oki, thanks! I measured the input voltage, and it is 2.4V. I recorded a short sequence in air, and there was no DC offset. I wonder about the 3.0V max: with my setting, with the 12kOhm resistor between signal in and ground, and considering the earlier comment by Edwin Houwertjes (pasted below) I would have supposed that the max should be less than 3V, since the signal moves around the new bias voltage 2.4V. I.e. the positive half of the signal would max at 3-2.4V=0.6V, giving a total of 1.2Vpk-pk. Am I totally wrong here..?

Alex Rogers

Apr 29

Replying to

Yep. a DC bias of 2.4V should be okay. It will saturate with an input signal greater than about 110mV peak-to-peak as the next stage of the amplifier is applying a gain of 11 about the existing DC bias.

Edited

Mats Amundin

Jan 09

Hi again, the reason for my earlier question about the H2d hydrophone is that I want to put my AM in an underwater casingwith room for more batteries and connect it to an external hydrophone. This casing has a custom-made ceramic hydrophone which I connected to the input terminals on the AM PCB, via an Aquarian Audio PA-4 preamplifier (powered by a separate 9V battery). I have soldered a bridge between the AM input ground and the jack detect terminal, and inserted a 12kOhm resistor between ground and signal input, as suggested by Jay Barlow in an earlier note. This worked fine when I tested with the internal batteries. I recorded 120kHz “clicks”, similar to harbour porpoise sonar clicks, generated by a fishfinder. Then I went to the next step by replacing the internal batteries with an externa 9V battery cassette (6 AA batteries), connected to the + and – terminal on the AM PCB (the internal batteries were removed). I first tested this setup using the AM live app, with the AM setup as a USB Microphone. This worked fine, also capturing wav on the uSD card. Then I did the final test, trying to record on the uSD card. The AM Firmware Basic was installed without problem, and the AM config seemingly was OK as well. However, when I turned the switch to Default, there was no flashing by the red LED, and there was nothing recorded in the uSD card. I did the External SRAM check, and the green LED flashed both in Default and Custom. When I switched to Default with the uSD card not inserted properly, only the green LED flashed.

What may be wrong, and can it be fixed? Thanks in advance,

mats

Mats Amundin

Jan 09

Replying to

Oki, I'll try, even though it is very tiny pads...

Anyway, I am so happy that the PCB seems to be OK! :-)

Thanks so much for your swift assistance, Alex!!

Alex Rogers

Jan 09

Replying to

Just to confirm, are you connecting the 9V to the battery + & - terminals or the GPIO + & - pads? It should be the battery terminals.

Mats Amundin

Jan 09

Replying to

I have soldered them to the now 2x 3AA pack to the battery terminals :-)

Alex Rogers

Nov 27, 2023

Yes. It works very well with the H2d Hydrophone.

Edited

Mike Bennett

Jul 21

Replying to

Reading all the previous comments and bumping up for update. The previous poster asked about using the HTI-99-UHF for porpoise clicks.

https://hightechincusa.com/products/hydrophones/hti99uhf.html

When ordering this hydrophone, there are choices for frequency response filters (2Hz-250kHz), and gain (-40dB to 44dB), so what would be the ideal choice considering in the AM software there are also filters to be applied? Frequencies of interest 125-150kHz.

I would think it best to filter out the unwanted frequencies at the hydrophone, before they get recorded, then filtering much larger files.

Alex Rogers

Jul 21

Replying to

Generally, filtering in post-processing is always better as you get to revisit any decision. The only exception is if you want to exclude frequencies that would otherwise saturate the input. For example, if you want to exclude high-amplitude low-frequency signals which saturate the input and prevent the recording of low-amplitude high-frequency signals. Without direct experience of particular hydrophones and their use in a specific application, it's hard to offer any specific advice.

Edited

Mike Bennett

Jul 21

Replying to

But at high sample rates that means going through lots of raw data and uSD cards! I suppose I am on my own here with no specific choice, even though I gave you a specific hydrophone. Other discussions in these forums seems to favor an HTI hydrophone v an Aquarian.

I wonder if anyone has tried the newer Aquarian A5 that mounts to a bulkhead? The sensor may be too close to the bulkhead thus creating a vortex flow. And the A5 bandwidth upper limit is 10kHz.

After my experiments, I will gladly share my results. It does seem that not too many researchers that have used and still use AudioMoths post their results here. Bummer, indeed.