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

article

In case you missed it... (no.2)

Stephanie O'Donnell

27 October 2023 1:20pm

Five #tech4wildlife people, projects and updates that caught my attention this month. This issue is a naturetech, biodiversity startup edition! Featuring a living map of the biodiversity startup scene, is nature data...

1
See full post
careers

Acoustic Monitoring Specialist

Mandy Holmgren

26 October 2023 7:32pm

The Institute for Bird Populations (IBP) seeks a California-based acoustic monitoring specialist to collect, manage, and process avian acoustic data from multiple research and conservation projects across California...

0
See full post
Link

Soundscapes and deep learning enable tracking biodiversity recovery in tropical forests

Stephanie O'Donnell

20 October 2023 3:38pm

New paper in Nature Communications from Jörg Müller et al. using BAR-LT recorders and CNNs to track biodiversity recovery. Study shows that #soundscapes 🎙🎶 and deep learning are powerful tools for tracking biodiversity recovery in tropical forests

1
Read "Soundscapes and deep learning enable tracking biodiversity recovery in tropical forests | Nature Communications" on Nature

Soundscapes and deep learning enable tracking biodiversity recovery in tropical forests | Nature Communications

Tropical forest recovery is fundamental to addressing the intertwined climate and biodiversity loss crises. While regenerating trees sequester carbon relatively quickly, the pace of biodiversity recovery remains contentious. Here, we use bioacoustics and metabarcoding to measure forest recovery post-agriculture in a global biodiversity hotspot in Ecuador. We show that the community composition, and not species richness, of vocalizing vertebrates identified by experts reflects the restoration gradient. Two automated measures – an acoustic index model and a bird community composition derived from an independently developed Convolutional Neural Network - correlated well with restoration (adj-R² = 0.62 and 0.69, respectively). Importantly, both measures reflected composition of non-vocalizing nocturnal insects identified via metabarcoding. We show that such automated monitoring tools, based on new technologies, can effectively monitor the success of forest recovery, using robust and reproducible data. Cost-effective biodiversity monitoring through time is important for evidence-based conservation. Here, the authors show that automated bioacoustics monitoring can be used to track tropical forest recovery from agricultural abandonment, suggesting its use to assess restoration outcomes.

Nature
discussion

Is there a good resource for learning bird calls?

Stephanie O'Donnell

20 October 2023 12:13pm

This question popped up on Twitter (or x?) and has had some interesting replies! I thought I'd start it as a discussion here too, and not limit it to australia, as it feels...

2 0
Stephanie O'Donnell

20 October 2023 12:19pm

Some of the recommendations Kylie has been getting: 

Stephanie O'Donnell

20 October 2023 12:21pm

A few more! 

https://twitter.com/D0CT0R_Dave/status/1715113765644140728https://twitter.com/D0CT0R_Dave/status/1715114095614238851 https://twitter.com/jason_mingo/status/1715130636749730120
See full post
discussion

Arbimon - Jobs are not working

Jennifer Lisa Insupp

11 October 2023 8:31am

Hi guys,I have a quick question: The jobs on Arbimon are not running for a week now, I can't find a support site on Arbimon. Does someone have a suggestion, why the jobs won't...

3 1
Marconi

12 October 2023 4:22pm

Hi Jennifer,

Thanks so much for notifying us about this! We have a support page here and you can contact our user support team directly by clicking the green 'question mark' icon in the bottom right corner of the Arbimon page. We can absolutely look on the backend to see what is stalling your jobs, we'll just need a bit more information (e.g., project name, job type, playlist size, etc.) to do diagnostics. I can reach out to you directly for that info - what's a good email for you?  

See full post
discussion

Song Meter Micro experience?

Nick Gardner

27 October 2022 6:33pm

Hi everyone, First off, what a great community this is! I only recently found out about it, and it seems like a game changer, especially for early career folk like me.I was...

24 1
Nick Gardner

11 May 2023 4:45pm

Hi Lucy, 

Sorry! I was in the field by the time you replied, and have not been back long. I realise it's been months now, but are you still planning on using the SM Micros? If so, please have a look at my thoughts in the new comment on this thread. I would think in Norway, the stridulating insect issue would not be as bad as in the tropics. Depends on what you want to use them for. Battery life is still definitely an issue though, and even more so in low temperatures. Good luck, let me know how you get on, or if you went with a different recorder.

Cheers,

Nick

Panji Gusti Akbar

26 September 2023 5:55am

Hi everyone. I am suuuuuper late on this discussion but I just finished 6-month fieldwork deploying 18 SM Micro over 70 different locations in Java, so I thought I would like to share my opinion on this device. So my work is basically installing each recorder in one location for the duration of 7-14 days before moving it to a different location, with battery and silica gel changed in between every deployment. Here's my thought:

Waterproofing - The first 2 months of the fieldwork was the wettest season I have ever experienced in the field. Extreme rain almost on a daily basis. Most of the time the recorder was working just fine except for the 2nd week of deployment when I found the inside a bit wet, but it was probably because I forgot to ensure that both lids were clicking.  The recorder still works fine after this deployment, and this was when I decided to change the silica gel before the next deployment.

Sound quality - As mentioned by others. there's an annoying peak at 6kHz. I did not originally plan to do soundscape analysis so it wasn't a big problem at the beginning, but now there is some discussion on doing so. Other than this I found the quality to be just adequate.

Battery Life - I am using alkaline 1,5 volt (non-rechargable) battery and get ~120 hours of recording, with a sample rate of 32khz. Suitable for my method, but certainly not for longer deployment.

Safety - My field site is full of a poacher who does not like to be detected, and the rangers warned me so many times about the camera trap that was destroyed or stolen by these people. I used steel cable and a small carabiner-style padlock which I later realized wouldn't stand a chance against a few hits of the machete, but luckily no recorder went missing--mostly because I tried to hide it as far as possible from the trails and cover our tracks. The small size of SM Micro makes it easy to hide among foliage (be careful on contact noise between leaves and recording), and the fact that it is colored green and has no external light also greatly helps (unlike Swiftone. Seriously, who thought that white is a perfect color for an expensive device that will be left in the forest for a long time?)

In conclusion, it is a cheap and easy-to-get device (compared to the mythical Audiomoth) for species presence detection in a short time like I did, but certainly not for longer deployment or more intricate bioacoustic analysis. 

Wildlife Acoustics

6 October 2023 9:08pm

Hi everyone, 

I wanted to provide an update on behalf of Wildlife Acoustics regarding some of these comments in this thread. As some of you may know, we have just announced the second generation of our Song Meter Micro. Along with 33% longer battery life due to an expanded battery tray (which now holds 4 AAs instead of 3), the Song Meter Micro 2 will also have a flatter frequency response. Though we don’t have a firm release date at the moment, we are currently looking at an early 2024 release, with pre-orders likely to start at the end of 2023. Detailed frequency response charts and specs will also be posted to our site soon. In fact, we have built an anechoic chamber in our new office and will be doing detailed testing with a host of new precision instruments. For more info on the Micro 2, check out our current product page. Lastly, thanks all for your open and honest feedback. It allows us to make our products better and better for you. 

Anechoic Chamber

See full post
careers

Conservation Innovation Manager

David Will

27 September 2023 11:08pm

Island Conservation's Innovation Team is dedicated to developing innovative, data-driven tools to increase the scale, scope, and pace of island restorations around the world. As the Conservation Innovation Manager, you...

1
See full post
discussion

Soil Ecoacoustics - viable?

Stephanie O'Donnell

12 September 2023 1:29pm

Hi everyone, A colleague is keen to get a gauge on the potential of soil ecoacoustic tech. They are asking - it appears to be quite simple and practical tech, is it viable...

7 0
Stephanie O'Donnell

19 September 2023 9:03am

Awesome, thanks a lot carlos! I've shared with the person who asked me. 

if you're up for it, I also may steal you for a variety hour spot on this at some point, as now i'm curious to hear more! 

See full post
article

In case you missed it... (no. 1)

Stephanie O'Donnell

12 September 2023 11:49am

Five #tech4wildlife people, projects and updates that caught our attention this month. An AI supported bear early warning system, a project that's connecting indigenous communities with high speed internet, exploring...

1 0
Lars Holst Hansen

13 September 2023 6:23am

Hi Steph! Nice with this "sporadic series"! The bear alarm in China reminds me of the Bear Repeller by the Hack the Planet guys:/Lars
See full post
event

Audiomoth online conference

Stephanie O'Donnell

5 September 2023 11:46am

For the second time, you can join an online conference dedicated to the little green bioacoustics device known as Audiomoth.

1 1
Rokshana Bushra

24 October 2023 7:18pm

Hello! Is it possible to get a recording of this conference? I am working on a conservation project using Audiomoth and I think the material covered in this conference will be...
See full post
Link

Estimating California Spotted Owl population with bioacoustic survey

Ellie Warren

29 August 2023 1:55am

News about a population survey of California Spotted Owls using 1,600 autonomous recording units in the Sierra Nevada region. Machine learning made analyzing this acoustic data possible at this large scale for the first time.

0
discussion

Building a new recording device for sound localization

Kim Hendrikse

20 June 2023 11:57am

Hi all, Nice group! I'm building an acoustic monitoring device that I wanted to share as it may be of interest to some people and also I'd like to know of any requirements that...

27 0
Kim Hendrikse

12 August 2023 10:26am

Thanks! Yes, I'm using the libraries from opensoundscape and it was Tessa who pointed me to wildlabs :) I'll publish the source code to the sound capture code this weekend in github with minimal instructions on configuring the rest of the OS till I get time to write an article about it which won't be till I've built 5x articles and done a convincing field test for a motivating article. Note, although I learnt how to configure the time synchronization with PPS from other articles on the net some steps were missed out that were vital to getting a fast and good sync, I'll be publishing these. Specifically, the start order of the gps versus chrony needs to be reversed from the default otherwise the PPS signal is not being communicated via shared memory to chrony which ruins the convergence to accurate time. Along with allowing the chrony time server daemon to step the clock so long as the alignment has > 1 ms of difference.



Simulations with opensoundscape suggest really, really good accuracy with the soundfinder algorithm is possible. When we get closer to new years eve I'll get really good test data from all the illegal fireworks that get set off here.



Once the capture side of things is sorted I'll move on to finding a good inference algorithm that performs well on the raspberry pi and then to working on a central server to correlate all of this automatically with the goal to making something that sends out an alert to a google maps link as to the location of the source, that will be fun :)



Oh and actually, I have to finish the install code for my other project that links computer vision algorithms to a state management and alerting system. This runs on jetson series SBCs. I've ported it all to the new Jetson Orin series computers but haven't yet finished writing the install code because I got distracted with this project about 4 months ago. The computer vision alerting state machine thing is here, and it's also useful in the fight against poachers. The new Orin series from nvidia allow the highest scoring open source computer vision algorithms (based on the microsoft coco dataset) to be integrated, such as yolov6, which results in significantly less false positives. And the state management facilities of my app provide you with a large amount of flexibility to reduce this even further.

https://github.com/hcfman/sbts-install
Kim Hendrikse

13 August 2023 3:28pm

As promised, I've pushed the code to github today, the URL is:



 

Note, this is somewhat of a premature announcement (to keep my promise) because I still need to add a bunch of documentation and how to install the prerequisites, recommended way of deploying etc.



However, for the curious, it is a very simple program. And so long as you install jackd2 and the prerequisites to the build script (./build.sh) it will run on whatever system you have and accurately align the sound files to the system time, along with the tracking file.



Much of the work surrounding this project was working out how to setup the GPS synced system time correctly. There are several guides on how to do this, however, when I tried to run this on a system that was disconnected from the network I found that there were important steps omitted in the guides that were needed to provide fast and accurate syncing in the absence of any network. I'll be adding these steps to this repository over time and write up a nice article about it once I've done the field tests. With the changes I made when installing the time sync changes the current system will sync to sub microsecond accuracy within a minute from when 5 satellites are locked.

Kim Hendrikse

24 August 2023 12:52pm

I've tested a Petterson ultrasonic microphone with a sample rate of 384khz and it works out of the box with this software as well. Would be cool to sound localize some bats.

See full post
Link

Acoustic Monitoring and AI to Recognize Elephant Sounds & Gunshots

Ellie Warren

9 August 2023 11:11pm

This article discusses Cornell's bioacoustics work with AI tools to recognize both elephant "voices" as well as gunshots. The elephant rumbles analyzed in this work are almost imperceptible to the human ear, but can be clearly detected with technology.

0