Making the Most of Bird Sounds

Posted byJoss CarrPosted inBiological recording, Blog, Virtual Symposium, Wilder SensingTags:, , , , , , , , , , , , , , , , , , ,

Learn how to plan, analyse, and report high-quality bird surveys using sound monitoring in this ecoTECH webinar with Wilder Sensing.

Bioacoustic monitoring is becoming an indispensable tool for bird surveyors, ecologists, and conservation professionals – but how do you ensure you’re using it to its full potential? This blog will guide you through best practices for planning, running, and analysing bird surveys with bioacoustics, with a focus on high standards and practical application.

We’ll explore:

  • What bioacoustic methods are – in a nutshell, what is bioacoustic (or, passive acoustic) monitoring and how does it compare to conventional methods?
  • Considerations when designing your bioacoustic survey – Highlighting the things that you need to consider and justify before deploying your recorders.
  • Practical advice on deploying recorders – how to choose what model of recorder you need, how to position them within habitats and in relation to one another, and how to optimise your fieldwork while still keeping the recorders discrete.
  • Choosing the right recording settings – explaining how to adjust your settings to maximise data quality whilst reducing data volume.
  • Planning for analysis and storage – raising awareness of how vast biacoustic datasets can be. Have a plan for analysis, storage, and reporting!
  • Understanding analysis and storage software solutions – highlighting what a good software platform needs to deliver for effective bioacoustic surveys.

This event is part of a webinar series delivered by the Biological Recording Company and Wilder Sensing with support from Wildlife Acoustics and NHBS.

About the Speakers

Geoff Carss

Geoff Carss is the founder and CEO of Wilder Sensing and Ethos Wilder. Growing up in Botswana and Northumberland he really enjoyed getting close to nature.

After obtaining multiple degrees in Geology, Geoff started a successful career as a Technologist, Entrepreneur and Environmental Advocate. He also served on the Board of Directors in a number of technology startups, and as trustee for the Bristol Avon Rivers Trust.

After observing the continued decline of biodiversity in the UK and around the world, Geoff started the Wilder Sensing project in order to bring the latest technologies to the environmental debate. His vision is to enable a more evidence based approach to addressing our environmental challenges.

James Bell

James has a BSc(Hons) in Climate Change from the University of East Anglia and a MSc in Applied Wildlife Conservation from the University of the West of England. During his studies, James developed a keen interest in emerging technical solutions to both the climate and biodiversity crises. Outside of his studies, James has long since loved birds and is rarely too far from a pair of binoculars. Bringing together cool tech and birds, Wildersensing was a perfect fit for James who joined in April 2025.

As part of the team, James helps to get new users set up with the software, as well as provide ongoing support to customers. James also works closely with the technical team to help feedback user experiences to continually improve our platform.

Q&A with Geoff and James

1. When spacing out recorders, do you need to consider how wind might affect the sound carry distance?

James: Yes. Whilst passive acoustic recorders typically only capture sound events within 50 metres, the general recommendation is to position your recorders at least 100 metres apart. This is to prevent double counting a single sound event in exceptional circumstances, such as sound being carried beyond 50 meters by wind, etc.. Additionally, it is also a good idea to avoid positioning recorders in exposed areas where there’s often strong, blustering wind (for example, exposed moorland). You will still get a few detections from recorders in such settings, but there will also be many calls drowned out by the wind which you’ll miss. Opt for a sheltered location, if at all possible.

2. Would positioning a recorder low to the ground lead to a reduction in audio quality?

James: It doesn’t lead to a reduction in audio quality, but it does decrease the detection likelihood for species. Having it positioned higher off the ground – e.g. at our recommended height of 1.5 metres – strikes a good balance between logistical ease and detection.

3. In the presentation, James mentioned that bioacoustics monitoring is not appropriate for all species. What species is it not appropriate for?

Geoff: Those which don’t vocalise very much! Or, in other words, species which are easier to see than to hear. Another consideration is that bioacoustic recorders are fixed point devices – they only record in a specific spot. If you’re surveying a large area for a species which may only use small patches, and have limited budget for bioacoustic recorders, a traditional moving transect survey may be more appropriate.

4. Is it possible to scan through sonogram datasets on BirdNET to quickly to isolate and identify unusual calls?

James: Yes. After running your data through Wilder Sensing, you can look at the species list generated and then interrogate in more detail the interesting and unusual species. Within the Wilder Sensing platform, you can go to the specific section/s of the sonogram which the machine learning algorithm tagged as representing that species. This human validation step is essential for unusual records.

Geoff: As an example of this, on Countryfile one customer shared their story of interrogating an audio file in which the tags went ‘Lapwing, Lapwing, Lapwing, Lapwing, Lapwing, Peregrine Falcon’ and then silence! Going beyond the simple species list, then, the customer was able to pull out a behaviour (Lapwings going silent when a Peregrine Falcon enters the area) and validate the presence of both species as correct.

5. Is it possible to report misidentifications back to BirdNET in order to improve the accuracy of the identification system?

James: We work closely with Cornell University, who manage the machine learning algorithm behind BirdNET. When reporting such instances, we can isolate correctly identified high quality clips for each species which Cornell can then feed back into the training algorithm. As an example, in the past, many users reported having a certain phrase in a Robin’s call misidentified as a Hawfinch. By reporting this back to Cornell, they were able to iteratively improve the algorithm such that this misidentification has now stopped happening as often.

6. Can you analyse the Doppler shift in frequency of bird calls to triangulate the flightline heights of passing calling birds?

James: This is a super interesting and complex topic. The short answer right now is no.

Geoff: We did some related work at Wendling Beck in Norfolk looking at tracking the direction of migratory bird flocks using multiple microphones stationed apart from one another. This works fairly well in principle – you first pick up calls on microphone 1, then, say, 10 seconds later on microphone 2 which is 100 meters away, then on microphone 3 and so on. That just looks at direction of movement though, not flightline height. I’ve also been involved in work using clusters of synchronised microphones to estimate the direction and distance of bird calls, focusing on skylarks specifically. Again, this worked in theory, but it is too complex to be commercially viable yet.

7. When deploying bioacoustic recorders in public areas there is presumably a risk of recording conversation between members of the public. Is this an issue for data protection laws?

Geoff: The passive recorders just record – they don’t analyse. Furthermore, once uploaded to our website, we don’t analyse the audio files for human sounds. There are, however, some conversations which are unavoidably accidentally recorded, and even if not analysed, these do end up in data storage. For customers who are especially scrupulous about complying with GDPR, we recommend deleting the files from storage once they’ve been analysed and interrogated. But that does of course mean the files can’t be re-analysed in the future. It’s a trade-off.

8. Do you do recorders for bats as well?

Geoff: To clarify, Wilder Sensing do not create or sell audio recording devices – we sell a software platform which works with audio recording devices to analyse the audio files and harvest ecological data from them. We work with various companies that sell audio recording devices – Wildlife Acoustics, AudioMoth, Frontier Technology, for example. And whilst we’ve spoken exclusively about birds today, we do also provide analysis software for bats. It’s a totally different context for bats with different considerations, though. If you’re interested in this I suggest getting in contact with us directly at info@wildersensing.ac.uk.

9. How long would it take to process one hour of audio data using BirdNET?

Geoff: We’ve recently completely re-written how Wilder Sensing behaves from a processing perspective. We’re now able to utilise ‘virtual machines’ to cope with increased demand – we get 50,000–100,000 files a day through the system some days! Thanks to this, processing times are way down. It now takes about 30 seconds to analyse a file with one hour’s worth of audio. So, large datasets are now generally processed in a few hours. Processing time is also more or less the same irrespective of whether or not are ‘interesting’ species detected.

10. How do you see this technology developing in the next 3–5 years?

James: I would anticipate that as global sound libraries become larger and more training data are made available, models will likely develop in two major ways: (1) improvements in global coverage and performance; and, (2) the capacity to identify call types rather than just species. Right now, for instance, models perform best in North America and Western Europe because that’s where the largest training datasets have been made available. Models tend to not perform as well elsewhere where training datasets are smaller. As global sound libraries get larger, species coverage and model precision will likely improve. Similarly, as training datasets become larger, it introduces the potential for labelling and training on specific call types – e.g., flight calls, alarm calls, fledgling begging, etc.. This could open up fascinating insights on behaviours and site use.

11. What projects and partners do you see Wilder Sensing working with over the coming years?

Geoff: When we started Wilder Sensing three years ago, our main partners were large organisations like the Wildlife Trusts, the RSPB, and the Wildfowls and Wetlands Trust. Over the past two years we’ve seen a huge increase in demand from a diverse mixture of other partners – regenerative farming ventures, solar farm developments, water companies, etc. – who are really interested in looking at cost-effective ways of monitoring wildlife on the sites they are responsible for. We also have a large number of professional ecological consultancy firms using our platform to complement their site-based survey methodologies. Hopefully this diverse mixture of partners will continue into the future!

Useful links

Literature References

  1. Fairbairn et al. (2026) ‘BirdNET can be as good as experts for acoustic bird monitoring in a European city’: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0330836
  2. Gibb et al. (2018) ‘Emerging opportunities and challenges for passive acoustics in ecological assessment and monitoring’: https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.13101
  3. Metcalf et al. (2022) ‘Good practice guidelines for long-term ecoacoustic monitoring in the UK’: https://ebaker.me.uk/pubs/2023_02_Guidelines.pdf
  4. Teixeira et al. (2024) ‘Effective ecological monitoring using passive acoustic sensors: Recommendations for conservation practitioners’: https://conbio.onlinelibrary.wiley.com/doi/10.1111/csp2.13132
  5. Metcalf et al. (2021) ‘Optimizing tropical forest bird surveys using passive acoustic monitoring and high temporal resolution sampling’: https://zslpublications.onlinelibrary.wiley.com/doi/10.1002/rse2.227
  6. Vasile and Merce (2025) ‘Acoustic bird monitoring – advantages and disadvantages’: https://jhfb.ro/index.php/jhfb/article/view/166
  7. Walston et al. (2025) ‘Ecovoltaic solar energy development can promote grassland bird communities’: https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2664.70208
  8. Wood et al. (2021) ‘Survey coverage, recording duration and community composition affect observed species richness in passive acoustic surveys’: https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.13571 
  9. Ziegenhorn et al. (2026) ‘Comparing acoustic and visual monitoring methods for assessing biodiversity and distributions of Arctic-breeding shorebirds’: https://besjournals.onlinelibrary.wiley.com/doi/full/10.1111/1365-2664.70300

Wilder Sensing ecoTECH blogs

  1. How Can We Use Sound to Measure Biodiversity: https://biologicalrecording.co.uk/2024/07/09/bioacoustics-1/
  2. Can Passive Acoustic Monitoring of Birds Replace Site Surveys blog: https://biologicalrecording.co.uk/2024/09/17/bioacoustics-2/
  3. The Wilder Sensing Guide to Mastering Bioacoustic Bird Surveys: https://biologicalrecording.co.uk/2024/11/26/bioacoustics-3/
  4. Bioacoustics for Regenerative Agriculture: https://biologicalrecording.co.uk/2025/03/31/bioacoustics-for-regen-ag/
  5. AI-powered Bioacoustics with BirdNET: https://biologicalrecording.co.uk/2025/07/08/birdnet/
  6. Making the Most of Bird Sounds: https://biologicalrecording.co.uk/2026/03/11/making-the-most-of-bird-sounds/

Event partners

This blog was produced by the Biological Recording Company in partnership with Wilder Sensing, Wildlife Acoustics and NHBS.

More for environmental professionals

Check out our events on Eventbrite

Published by Joss Carr

Junior Naturalist at Biological Recording Company.

Leave a comment