WILDLABS logoWildlabs.net : The conservation technology network

WILDLABS logoWildlabs.net : The conservation technology network

Group

Acoustics / Feed

Acoustic monitoring is one of our biggest and most active groups, with members collecting, analysing, and interpreting acoustic data from across species, ecosystems, and applications, from animal vocalizations to sounds from our natural and built environment

discussion

Standard Threshold for "Habitat Use" in Avians?

Cortney Dean

7 March 2025 9:05pm

Hi all!Background: I did bioacoustic sampling at 12 sites, roughly 18 days of recording at each site, over the course of a season (both birds and bats). In reviewing our species...

7 1
Walter Zimmer

9 March 2025 7:47am

Is that not a statistical question and depending in the classification performance of the SW (Birdnet), which itself may depend on the species and the variability of the call? 

I, myself, have no practical experience with Birdnet, but use a commercial product that derives from Birdnet and found in about 1000 days, for example,  the following tail statistics (22 x 1), (12 x 2), (6 x 3), (5 x 4) bird detections. In total, 223 different species were classified, with black birds being on top with 149000 detections. Overall, there seems to be a nearly exponential decay of different bird detections, which to me shows a problem with NN-based classifications, as there seems to be no clean separation between 'signal' and 'noise', as one can find in traditional signal processing. 

From these statistics (histogram) I would conclude that in the end there is no meaningful threshold, as there is no meaningful 'misclassification floor' that could be used to define a threshold.

Consequently, you may have to consider context, inspect every suspect snippet and let human decide. 

Tessa Rhinehart

11 March 2025 2:44pm

Hey Cortney, this is such a good question! 

First off - I'd be very careful about interpreting the number of BirdNET detections of a species as a vocalization index for that species. For example, let's say a recorder captures a lot of Species A, which sounds like Species B,  but Species B isn't present at that site. You will often find that BirdNET (or any other sound ID model) reports many detections for Species B and Species A, making it seem like both species are present. So, the number of detections of a species at a site isn't necessarily a more reliable indicator of species presence at that site.

One alternative to consider would be to verify species presence at each site and each day, then use the number of days that species was present as a threshold for habitat use. What threshold to use depends on the species - we've used three consecutive days of presence for some songbirds, for instance.

You could also remove dates when the species is expected solely to be migratory. However, this can get tricky... For example, I've observed that Cerulean Warblers that breed at a particular site will arrive there a week or two before migrants pass through at nearby non-breeding migration hotspots.

Doing a day-by-day analysis of presence enables you to more easily check machine learning detections than if you were checking for the number of detections for each species. You can verify the presence of the species at each site and day by listening to the highest-scoring clips for each day for each species. 

If you have 18 days of recording * 12 sites and you listen to the top 5 highest-scoring clips per species and day, that's a maximum of 1,000 five-second clips per species. In my experience, reviewing that amount of clips takes me only a few hours using a Jupyter notebook like this one:

See full post
funding

BirdCLEF+ 2025

Alex Rood

12 March 2025 2:22pm

Species identification from audio, focused on birds, amphibians, mammals and insects from the Middle Magdalena Valley of Colombia.

2
See full post
discussion

Sea turtle Bioacoustics Project

Sam Jamil

17 February 2025 8:16pm

Hi everyone, I have an opportunity to work with Green sea turtles and Olive ridleys off the coast of Sri Lanka, this will be my first bioacoustics conservation project. I would...

5 0
Claudia Lacroix

7 March 2025 12:47pm

Hi Sam, I  did my master's on hatchling turtle vocalisations and their role in nest emergence behaviours (currently under review for publication). I recorded nest emergence behaviour in-situ using microphones and camera traps. I worked with snapping turtles, but the methods could be quite useful. I would be happy to share my thesis if that would be helpful.

there are a few sea turtle papers that describe hatchling vocalisations but not many experiments testing hypotheses for these vocalisations. 

here are some papers that could help you get you started:

Shoot me a message if interested in chatting more :)

 

M Zahirul Islam

7 March 2025 1:32pm

Hello Sam ...great work, would like to see the paper when it comes online. I would like to know about the device....Bests Zahir

See full post
discussion

Vertual Fence

doug hemingway

27 January 2025 6:26am

I have 240 hectares of old-growth forest in Far South East Coast NSW Australia.I tried to build a feral fence but that got all too hard.The site is on a peninsular so much easier...

4 0
Akiba

28 January 2025 12:38am

Hi Doug. 
We are located in Melbourne and work with Bush Heritage in large reserve monitoring. Would love to chat and hear more about the issues you are having to see if we can help out with it. Let me know via private message and we can set up a meeting. 
Akiba

doug hemingway

28 January 2025 8:41am

I will be back in Melbourne Thursday.

I would like to send you the Florer and Forner study done in 2010 to bring you up to speed.

I'm new to this and can't find the attachment key.

Can you send an email address?

Doug

See full post
discussion

Has anyone tried Whombat?

Adrien Pajot

19 February 2025 12:03am

While reading my emails I came across this paper:  In the context of a personal project (NBM) I might be interested in using it. Has anyone tried it?

6 1
Jocelyn Stalker

25 February 2025 3:57pm

I just started using this software a few days ago and so far it has captured and kept my attention more than other programs such as Raven or Kaleidoscope. I have only been using it to annotate, but I'm really excited about its ability to test machine learning models. I'm currently developing frog call recognizers using BirdNet. I don't yet know of anyone else using a Whombat + BirdNet workflow (and for neither bats nor birds, nonetheless!) but I'd encourage more biologists to try it out.

See full post
discussion

Bat Identification Tools Comparison

Stefan Zeeman

16 May 2023 3:32pm

Hello, does anyone have any experience comparing the results of Bat identification software for large batch processing? I would be processing significant amounts of audio...

5 0
Stefan Zeeman

8 June 2023 1:38pm

Hi Stuart, Thanks for your response, and nice to e-meet someone behind the BTO acoustic pipeline! I have been testing out BTO quite a lot recently, and have started comparing results to Kaleidoscope this month, and have been manually checking classification results against raw audio. Will keep you updated with any findings if of interest.

Is there a minimum frequency recording that needs to be taken when uploading the BTO acoustic pipeline? I'm also interested in small mammals, and suspect that 96 KHz should be fine, rather than 192/384 when recording for bats.

Carly Batist

22 June 2023 6:57pm

Another option for you could be Arbimon - it's free and does not require any coding/programming knowledge! 

Sheila Holmes

24 February 2025 3:20pm

As someone who is looking for a relatively simple solution for multiple collaborators (all with very little expertise in bats or acoustics) in a large consortium, I am curious as to the results of your comparison. I have been looking at both BTO and Kaleidoscope, though my previous experiences with Kaleidoscope were that it was far from perfect. However, it may be more cost-efficient. 

See full post
discussion

Audiomoth with solar panel in desert for anthropogenic noise and wildlife limitations and configuration

Tom Bruce

21 January 2025 12:18pm

Hi all,I am a new user to passive acoustic monitors generally, and was looking at purchasing an audiomoth. I have a couple of questions that I hope the community here may have...

12 0
Richard McMillan

24 February 2025 4:09am

Hi Tom

Saffron do a retrofit for the Sng Metre Mini and Micros (Versions 1 and 2). We have them doing a 12 month survey in very hot and dry conditions here in Australia. Will not touch them until October 2025, and they are run by a solar power pack.  

Read "Retrofitted Solar Power Pack - Saffron Aid" on Saffron Aid - Empowering Local Communities and Building Animal and Environmental Resilience in the Indo-Pacific

Retrofitted Solar Power Pack - Saffron Aid

Expand the life of your longitudinal study by increasing the recording time of your Song Meter Micro or Mini. Have your recorder retrofitted to take power from our Solar Pack. Do not fit any batteries into the recorder all power is supplied by the Solar Pack's inbuilt lithium battery. With a 512GB SD card set

Saffron Aid - Empowering Local Communities and Building Animal and Environmental Resilience in the Indo-Pacific • 23 Oct 2024
Lars Holst Hansen

24 February 2025 7:05am

Hi Tom!

I have no experience with deployment in extreme heat or solar panels with Audiomoth.

If you go with AudioMoth, I would consider the new injection molded case for the AudioMoth Dev version: https://github.com/OpenAcousticDevices/Application-Notes/blob/master/An_Injection_Moulded_Case_for_AudioMoth_Dev/An_Injection_Moulded_Case_for_AudioMoth_Dev.pdf

This case is prepared for external power which I believe will be hard to fit reliably with the standard case.

Also, it has room for a GNSS/GPS unit which can help mitigate time drift (very useful for long deployments) AND be used to precisely schedule recordings in relation to sunrise and sunset.

Read more about the GPS possibilites here: https://github.com/OpenAcousticDevices/Application-Notes/tree/master/Using_AudioMoth_GPS_Sync_to_Make_Synchronised_Recordings 

&

https://github.com/OpenAcousticDevices/Application-Notes/tree/master/New_Support_for_GPS_Synchronisation_in_the_Standard_Firmware 

In any case, I would also ask @alex_rogers for direct advice on your plans. 

Cheers,

Lars

Ryan Smith

24 February 2025 11:36am

Highly recommend the Saffron Aid modified Song Meter Mini for a professional deployment. 

If you want to go down the audiomoth route it makes sense to look at options for the AudioMoth Dev. @Lars_Holst_Hansen made a great comment on this application above.

On another note, we are about to release a recorder that might be retrofit-able at Smith Robotics. So keep an eye out for that.

Equally, what is the real reason for the solar, as the SD card might be the limiting factor in some deployments. There are a few 4G options on the horizons fairly soon I think!

See full post
discussion

Small mammal detection

Sarah Dalrymple

21 February 2025 3:43pm

Is anyone (other than BTO's Acoustic Pipeline) working on detection and classification of ultrasonic vocalisations of small mammals? We're looking at automatic detection of...

3 0
Ryan Smith

21 February 2025 10:51pm

BTO definitely has the best classifiers out there at the moment. I know of a few people attempting to improve their small mammal sets, but none are great yet.

What hardware are you using? 

Sarah Dalrymple

23 February 2025 10:25am

Thanks Ryan, we've been trying the BirdNeT and BatDetect2 models using Acoupi on a pi and it seems to work well, considering it's running on a pi... A small mammal model that worked as well would be great for live alerts to rodent incursions on offshore islands, but it sounds like it's not something that's possible at this stage.

Ryan Smith

23 February 2025 4:15pm

I think its possible. I spoke to Stuart at the BTO a while ago about this. It's just not available as an off the shelf solution yet. If you can get real time outputs on the pi (or even an esp32), it would not be difficult to add the 4G or wifi part to it!

See full post
discussion

Bats detectors

Nabil Amor

17 February 2025 9:04pm

Hi everyone, can you recommend reliable bat detectors for fieldwork? I'm focusing on bat (Chiroptera) research and need a detector suitable for species identification and...

0
See full post
discussion

Bass/vibration measurements

Nikhil Whitaker

29 January 2025 4:34am

Hi, I work in a conservation organization, and we're looking to monitor noise from a nearby (20 meters away), very low, 20-40 hz analysis of bass/vibrations. I was recommended to...

5 0
Nikhil Whitaker

13 February 2025 12:17pm

@Thomas, @ Harold, thanks a lot for your feedback. Clearly there appears to be no cutting back in costing if quality is to be expected. I was hoping for cheaper solutions, but to record at a professional quality level looks like we have to go in for the more expensive options.  Thanks again, best regads, Nik. 

Nikhil Whitaker

13 February 2025 12:19pm

The point was to record vibrations, not air noise, thanks for explaining the technical issues with the HOBO unit we were thinking of, clearly a few minutes of data would not help. 

Constance Woodman

13 February 2025 3:34pm

You can hack a 6-inch underwater speaker onto a recording unit, and swap it from being  speaker to a microphone. They are, on paper, the same thing.  That frequency range doesn't work will with small piezo surface (itty microphones). If you want to get fancy, you can add the resistors and capacitors to make a band-pass filter, so you don't pick up quite as much thunder and passing vehicles.

This is a fairly standard audio problem that a good audio engineer can assist with.

I've detected low frequency bird wingbeats with some low-power analog wizardry in the past. But you will need dedicated audio engineering time.

See full post
discussion

List of bioacoustics software

Tessa Rhinehart

1 July 2020 3:23pm

Edit: Since posting this over 4 years ago, we've moved it to its own GitHub repository and associated website. If you have any suggestions for software...

7 9
Tessa Rhinehart

12 February 2025 4:23am

Hi! There are several nice tools available that have some, but perhaps not all, of the features you're looking for. Most of these are linked to on the current version of the bioacoustics software database here - https://rhine3.github.io/bioacoustics-software/

  • PAMGuard is a classic that's more geared towards underwater monitoring. It has plug-ins to do just about anything you can imagine. Unfortunately I'm not aware of any "terrestial" equivalent
  • Probably the closest to what you're asking for can be found on cloud-based applications - in particular, WildTrax (which I have used) and Arbimon (which I haven't used)
  • There are some applications you can run on your own computer/server - one example is SoundScapeExplorer. The BirdNET GUI also keeps getting better and better

I also want to mention WhomBat as well, which is a nice and very carefully thought-out open-source audio annotation platform. 

See full post
discussion

AI and Machine Learning in Ornithology

Dr Yvette Ehlers Smith

6 February 2025 10:05am

Call for papers: Special issue on AI and Ornithology | News from National Inquiry Services CentreWe are soliciting papers on AI and machine learning in ornithological studies for...

2 1
Marissa Lamb

10 February 2025 3:28am

Hello, 

I am very interested in possibly contributing a paper for this issue, but I study waterbirds in Texas. Would this be open for review papers or meta-analysis? Or deployment of new tools?

 

See full post
discussion

Low-power acoustics systems

Sebastian MARZETTI

27 January 2025 1:44pm

Hello,I am a researcher specializing in acoustics, electronics, and artificial intelligence. Together with my team, we develop terrestrial and underwater acoustic systems with...

6 1
Sebastian MARZETTI

10 February 2025 10:42am

Hello, we do not have a website yet. We are exploring where our technology can provide the most significant value compared to existing products. We would be very interested in discussing this with you.
Sebastian MARZETTI

10 February 2025 10:49am

Hi Julia,

Thank you for your reply. Your work on connected acoustic recorders sounds very interesting, and we’d love to learn more about your application and the enhancements to Bugg. We’ll contact you soon to arrange a discussion.

Looking forward to it!

Thank you,

Best regards,

Sebastian

See full post
discussion

land mammals on xeno-canto

Willem-Pier Vellinga

1 February 2025 3:43pm

Hello all, a few weeks back we opened Xeno-canto for land mammals as well. Joining Birds, Grasshoppers, Bats, and Frogs. Four recordists prepared some sets to start out with and...

3 10
Mark Bowler

4 February 2025 8:56am

Brilliant! We'll look into archiving here. Where we have labelled data with 100's of examples of each call for each species (Peruvian Primates mostly) is that something you are looking to host? or do you prefere a more curated submission?

Willem-Pier Vellinga

6 February 2025 10:27pm

hi Mark, cool! "Labelled data" sounds rather well curated :-) XC aim to show sounds of all taxa, their complete repertoire, all of the geographic variability, at all stages of development. If there is no geographic variability, the best way of showing that is to show sounds from all over the range of course. So, don't hold back. We are not looking to turn XC into a repository of TBs of unlabeled soundscapes, but we do like soundscapes! Not sure if that clarifies anything :-) It sounds as if you might benefit from batch upload, check this out https://xeno-canto.org/help/FAQ#lotsofsounds. Or drop us a line on contact@xeno-canto.org. As an aside, for me personally XC started with sound recordings in the rainforest of Peru (along the Rio Orosa in 1996), it would be fantastic to have a good representation of mammal sounds from Peru.
Phil Mitchell

7 February 2025 5:18pm

That is splendid news! Thanks Willem-Pier!

See full post
discussion

AI for Bird and Bat Recognition

Lorenzo Girotti

7 February 2025 11:57am

Hi everyone,I'm working on a project involving the automatic recognition of bats and birds from their audio recordings in Italy, with a possibile evolution also in other european...

1 1
Carly Batist

7 February 2025 5:06pm

Hi Lorenzo,

I highly recommend the OpenSoundscapes package (developed by the Kitzes Lab at U Pittsburgh) - there are workflows to build your own CNNs there, the documentation is really thorough, and the team are very responsive to inquiries. They also have a bioacoustics 'model zoo' that lists relevant models. The Perch model from Google would be good to look into as well.

Some recent papers I've seen that might also be worth checking out -

Hope that helps a bit!

See full post
Link

acoupi: An Open-Source Python Framework for Deploying Bioacoustic AI Models on Edge Devices

Carly Batist

7 February 2025 1:39am

New paper - "acoupi integrates audio recording, AI-based data processing, data management, and real-time wireless messaging into a unified and configurable framework. We demonstrate the flexibility of acoupi by integrating two bioacoustic classifiers...BirdNET and BatDetect2"

3