Archive · 11 Aug 2026
India's first open-access ecoacoustic dataset captures wild soundscapes
A long, long time ago, a man named Thomas Edison invented a machine that could record sounds.
A few years later, someone used a similar machine to record a bird called the white-rumped shama at a zoo in Frankfurt, Germany.
People think that was the very first time a bird was recorded.
Back then, rich people liked to collect records with birdsong on them.
Now, fast forward to the year 2025.
A group of scientists and nature lovers in India started recording sounds from nature all over the country.
They collected almost one hundred hours of recordings.
You can hear birds, frogs, insects, bats, reptiles, and even sea animals.
The recordings were put together into a special free collection called an ecoacoustic dataset.
It is the first one made in India that anyone can use to teach computers to recognize Indian animals by sound.
The earliest known bird recording featured a white-rumped shama recorded at Frankfurt zoo in the late 19th century.
In 2025, ecologists and enthusiasts began recording sounds across India's key ecosystems, gathering nearly a hundred hours of audio.
Birdsong dominates the soundscapes, alongside calls of insects, frogs, bats, reptiles, and marine animals.
The project produced India's first open-access, crowdsourced ecoacoustic dataset.
The dataset is intended to train machine learning models to identify Indian species, which previously had access only to data from the Global North.
- Who
- A collective of ecologists and enthusiasts in India; historically, Thomas Edison and early birdsong recordists.
- What
- The creation of India's first open-access, crowdsourced ecoacoustic dataset of wild soundscapes.
- When
- Recording began in 2025; the earliest known bird recording dates to the late 19th century.
- Where
- Across key ecosystems in India; the earliest known bird recording was made at Frankfurt zoo.
- Why
- To enable machine learning models to identify Indian species, filling a gap caused by training data limited to the Global North.
- Dataset
- India's first open-access, crowdsourced ecoacoustic dataset
- Recording scope
- Nearly 100 hours across varied terrain in key Indian ecosystems (2025)
- Captured species groups
- Birds, insects, frogs, bats, reptiles, and marine animals
- Earliest known bird recording
- White-rumped shama at Frankfurt zoo, late 19th century
- Historical trigger
- Thomas Edison's invention of the talking machine
- Primary use
- Training machine learning models to identify Indian species
- Previous data limitation
- ML training data restricted to the Global North
- Participants
- Collective of ecologists and enthusiasts