Our highlights from the International Bioacoustics Congress 2025

The International Bioacoustics Congress (IBAC) happens every two years. It’s one of the scientific meetings valued by many of us in the Bioacoustic AI network, because it brings together many different people who focus on animal sound – how to analyse it, how animals produce it, and why they do!

The 2025 edition of IBAC featured talks and posters from our Bioacoustic AI Doctoral Candidates. In this article, BioacAI co-investigator Pavel Linhart gives an insight into his highlights.

Posters and a talk from the BioacAI team: Minkyung Kwak, Vincent Kather (co-authoring with Ines Nolasco), Ilaria Morandi, Yasmine Benhamadi, Pavel Linhart.

Reconnected

My last IBAC was 16 years ago in Lisbon in 2009 as a beginning PhD student. Unfortunately, I missed all the subsequent conferences due to lack of funding or time. Despite being away from IBAC for many years, returning in 2025 felt like coming home. I reconnected with friends and colleagues I’ve known from other events and past or ongoing projects, and it was a pleasure to meet many new faces as well. The week included some 75 talks and about 160 posters, a huge increase from 2009. My summary below is highly subjective and selective, but I hope it captures some highlights and the unique spirit of the meeting.

IBAC is a… community

One of the hallmarks of IBAC is its single-track format, with no parallel sessions. This ensures that everyone experiences the same talks, fostering a strong sense of community and maximizing opportunities for interaction and learning from one another. While this limits the number of oral presentation slots, it redirects many fascinating contributions to poster sessions making it an integral part of the conference. It is truly special to see MSc students and senior professors stand side by side at their posters, as well as share the main stage for talks. This creates a relaxed and welcoming atmosphere. It even allowed for the sharing of powerful, intimate messages that would be difficult to hear at many other conferences. The talks themselves were of high quality, with many essentially serving as mini-plenaries, summarizing years of research. This makes IBAC an excellent occasion for introducing students to a wide range of bioacoustic topics while also helping experienced researchers catch up on advances across diverse areas.

At the core: acoustic perception and production

Acoustic perception and production remain at the core of bioacoustic research. Three of five plenaries (Peter Madsen, Natasha Mhatre, Benjamin Judkewitz) focused largely on morphological and neural mechanisms of sound production. It was fascinating to learn about morphological structures evolution created to make sounds (e.g., Peter Madsen); what tactics echolocating animals might use to adjust sensing depending on prey distance (e.g. Adam Smith, best talk); how possible function of bat tragus can be tested and rejected (Karsten Vesterholm); or, how scales might help in hiding moths from bat sonar (Ella Kennedy). With the variety in the morphology of sound production organs and sounds that animals can make, I wonder if the efforts to classify unknown sounds into broader taxonomical categories (e.g., is the sound mammal, insect, bird, frog, etc…) can succeed. We have listened to cheetah chirps sound like birds or very mechanical like sounds assumed to be narwhal calls. 

Chaos reigns, in animal sound research

Nonlinear phenomena (NLPs) – such as frequency jumps, biphonation, sidebands, deterministic chaos – received significant attention in talks and posters. Vocalizations with different types of NLPs are now analyzed separately and in increasing detail. New tools were developed allowing synthesis of different NLPs for playback stimuli and testing their possible functions. Jacob Dunn presented that primate vocal tracts are morphologically diverse and produce rich NLPs, while David Reby showed that humans may have simpler vocal tracts for speech efficacy but still use NLPs widely in non-speech vocalizations. Several other contributions described NLPs in different taxa and types of calls and showed how they are related to emotional states (e.g. Kasia Pisanski. Mathilde Massenet) and health of animals (e.g., Livio Favaro).

Tools: robots, AI, microphone arrays, and more

It is always interesting to hear about the development of new tools and how they might help us discover new facts about animals or help to protect them. In particular, marine bioacoustics seems to provide many clever solutions, probably because the marine world is harder to reach and study. It was rewarding to hear about marine gliders monitoring right whales in real time and slowing ships in shipping corridors to protect whales from collisions (Katherine Indeck). It was fascinating to see a robot continuously freezing movement and listening to underwater sounds in order to navigate the seabed to collect sounds and habitat images of coral reefs (Aran Mooney). Acoustic localization is increasingly used to identify sound sources within habitats or groups of animals, and AI tools are becoming more powerful, with transfer learning and active learning enabling large pre-trained models to be tailored to specific species and locations (e.g., Vincent Dumoulin, Ben McEwen, Giordano Jacuzzi) while hopefully remaining small enough for field deployments (e.g., Lorène Jeantet) and reducing their environmental footprint.

Bioacoustic AI: our work in IBAC

Contributions from our Bioacoustic AI network fit well into this direction. Four out of ten PhD candidates participated and presented their research, and the event was also attended by Tereza, Pavel, and Dan from the supervisors’ side. Ilaria gave a talk on individual identification limits in Yellowhammers, showing that both humans and machines (BirdNet) can reliably assign individual identity to songs up to 100 meters away. Overall, there is high demand in the community to move beyond species identification toward assigning sounds to individuals, unlocking new opportunities for acoustic monitoring (see also e.g., Sam Lapp). Three other Bioacoustic AI PhDs presented their posters.  Yasmine presented preliminary results from interviews with PAM practitioners on sustainability of acoustic sensors, future needs and challenges that need to be addressed regarding possible trade-offs between autonomy and sustainability. Vincent presented his solution to streamlined evaluation of various feature extraction models (see also previous blog). Min reported on how Yellowhammers shared initial song phrases within a locality but not across further distances and sharing of initial phrases was independent of dialect. Further, Tereza, Pavel, Ilaria, and Min stepped out of their comfort zone and took part in an animal imitation contest despite limited imitation abilities. With the support of Dan and Yasmine and other people in the audience, they tried to re-create pan-European Yellowhammer soundscape and collected some language specific transcriptions of Yellowhammer songs.

Unlock the sound

However, there are also some field-specific challenges. Despite its advantages, sound is apparently still underused as a tool for environmental and biodiversity monitoring (compared to other methods such as, e.g. camera traps, remote sensing, etc.)(Claudia Lacroix), which is particularly relevant for work within the Bioacoustic AI doctoral network. Several contributions highlighted this gap. Peter Tyack discussed the efforts by him and the International Quiet Ocean Experiment (IQOE) to include sound as an essential variable in the Global Ocean Observing System (GOOS). Other talks mentioned the lack of foundational models for sound data (but see e.g. Vincent Dumoulin talk on Perch2), and it became clear that bioacousticians are still protective of their annotated datasets. The solution, reiterated in several talks and posters (e.g., Peter Tyack, Claudia Lacroix,  Holger Klinck) is to collect high-quality, standardized data and make it available for collaborative research through databases and working groups. Such initiatives (e.g. Ellen Jacobs – Digital Twin Ocean; Isla Keesje Davidson – World Oceans Passive Acoustic Monitoring, WOPAM Project; Leonie Baier – sharing bat sounds through Xeno-Canto) will help the field make a stronger impact and ensure bioacoustics is no longer overlooked in large-scale environmental programs. But, maybe we are just less people working in bioacoustics than in other fileds. Hence, initiatives aimed at expanding the community (e.g., Tess Gridley – African Bioacoustics Community) also seems crucial to strengthen advocacy and visibility of the sound approach.

Passion, fun, and a few other tips

Throughout the conference, speakers offered inspiring and motivating advice. Peter Madsen encouraged young researchers: find an animal model you are passionate about. He demonstratied in his own case how early curiosity can steer and fuel your research career for many years to come. And obviously, sperm whales are cool animals! Other speakers elaborated on that and extended their „favorite animal“ definition to anything from 10mm transparent fish, a bush cricket, a puppy, human, robo bat, or a bioacoustics and science as a whole. Magnus Whalberg reminded attendees to have fun along the research journey, a spirit embodied by all the other co-organizers and volunteers who managed the event with humor and warmth. On stage, Nicolas Mathevon was the entertainer summarizing the series of studies on human baby cries advising: If a baby cries, do not panic – the baby is communicating, it is not that serious… 2) If you do not know why the baby is crying, do not panic – you are not alone, just get more training talking to babies! 3) If a crocodile is nearby, leave quickly.

And, yet another piece of advice from Henrik Brumm, the current IBAC president: Become an IBAC member and save money on the next conference fee (not a fake news)! Hopefully, see you in Skukuza, South Africa in 2027!

Conference webpage | Program book | Abstract book