hn.today

A million orthopteran singers from a year of continuous field recorders

biorxiv.org4 points1 comments
Screenshot of A million orthopteran singers from a year of continuous field recorders

A large-scale, unsupervised acoustic analysis identifies and organizes insect songs by exploiting the inherently repetitive structure of orthopteran calls. Continuous recordings from 27 autonomous units over 13 months (November 2024-November 2025; 175,896 recorder-hours) were processed to detect 10.9 million individual singers in five-minute windows and link them within nights and days into 911,005 tracks. Of those tracks, 103,802 persisted for two hours or more (71,770 nocturnal, 32,032 diurnal), yielding 335,882 hours of tracked song. The method segments distinct singers in short time windows, stitches them across changing carrier frequencies and syllable rates as conditions vary, and aggregates them across days and sites into a searchable catalog of singers and song types.

Analysis places singers on a plane of carrier frequency versus syllable repeat rate where individuals fall on discrete, temperature-sliding lines; 45 song types are defined (37 trilling, 8 chirping), each represented by dozens to thousands of singer-nights. Syllable rate increases by about 3% per degree Celsius and carrier frequency by about 1.5% per degree within nearly every type, confirming and extending classical temperature-dependence findings. Published reference recordings and species-level data (e.g., Anaxipha) map onto the same type lines at expected temperatures. The pipeline produces robust, high-resolution data for studying temporal, spatial and physiological drivers of insect acoustic behavior.

Read on biorxiv.org1 comments on Hacker News

Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.

More in Science

The daily digest

Today's best Hacker News stories, summarized and screenshotted, one email a day.