BioSonic — Nature Speaks. We Translate.

NATURE SPEAKS. WE TRANSLATE.

The most accurate bat sound analysis available.

BioSonic uses image recognition on spectrograms to identify bat species, filter noise, and separate feeding and social calls — automatically.

Try 15 GB for free

USED BY LEADING CONSULTANCIES

1. The problem with your AutoID

FALSE DETECTION

Why it keeps missing those faint calls

SCENARIO B — Real Barbastelle hidden behind grasshopper noise

BARBAR echolocating behind grasshopper noise. Kaleidoscope sees only the most loud part of the sound (many times, that’s noise) — BioSonic finds the bat using image recognition on the spectrogram.

KALEIDOSCOPE PRO

Echo pattern lost under noise floor — missed

— NOISE ONLY — No bat detected (Barbastelle missed)

BIOSONIC

Echolocation pattern locked under noise floor

✓ BAT: BARBAR · 78% confidence · Social call

SCENARIO A — Grasshopper noise, no bat present

Same noise. Different conclusion.

KALEIDOSCOPE PRO

Sees a busy spectrogram → assumes bat

⚠ BAT DETECTED: BARBAR (Barbastelle) · Confidence: Low

BIOSONIC

Trained on 3.5M noise files · full spectrogram, full context

✓ NOISE — No bat detected.

2. How AI beats your current AutoID

HEAD-TO-HEAD VALIDATION

Independent experts confirm BioSonic's accuracy

Two independent studies — one from UK sensor platform Wilder Sensing, one from Denmark's leading bat expert — plus a direct comparison against Kaleidoscope Pro on the same dataset.

THIRD-PARTY TESTING

Wilder Sensing · Geoff Carss + Annabel Jeffries

98.7%

F1 ACCURACY

“We tested a range of existing solutions using real acoustic data, and BioSonic clearly stood out, achieving 98.7% accuracy (F1 score) and giving us full confidence in the results we’re bringing to the platform.”

BIOSONIC VS KALEIDOSCOPE PRO

Same audio, same ground truth · Grassland + Paludiculture dataset

BioSonic scores 98.7% F1 vs Kaleidoscope Pro’s 73.0% F1 — a 25.7 point gap. Kaleidoscope produces 17× more false positives and 16× more false negatives.

BIOSONIC F1

98.7%

Precision 97.8% · Recall 99.6%

F1 GAP

+25.7pt

BioSonic outperforms Kaleidoscope by nearly 26 percentage points

KALEIDOSCOPE PRO F1

73.0%

Precision 61.4% · Recall 90.0%

METRIC COMPARISON

Higher is better on all three · bars start at 50% baseline to highlight the gap

PRECISIONBioSonic +36.4pt

BioSonic

97.8%

Kaleidoscope

61.4%

RECALLBioSonic +9.6pt

BioSonic

99.6%

Kaleidoscope

90.0%

F1 SCOREBioSonic +25.7pt

BioSonic

98.7%

Kaleidoscope

73.0%

BIOSONIC

Confusion Matrix · Accuracy 99.6%

PREDICTED BAT

PREDICTED NO BAT

ACTUAL BAT

716

TRUE POSITIVE

3

FALSE NEGATIVE

ACTUAL NO BAT

16

FALSE POSITIVE

4,579

TRUE NEGATIVE

KALEIDOSCOPE PRO

Confusion Matrix · Accuracy 93.8%

PREDICTED BAT

PREDICTED NO BAT

ACTUAL BAT

440

TRUE POSITIVE

49

FALSE NEGATIVE

ACTUAL NO BAT

277

FALSE POSITIVE

4,503

TRUE NEGATIVE

WHERE KALEIDOSCOPE PRO BREAKS

Every dot is one error

FALSE POSITIVES (Wrong "bat" calls)

Cost: investigator time wasted reviewing non-bat audio, inflated species counts

BIOSONIC

16

KALEIDOSCOPE

277

Kaleidoscope: 17.3× more

FALSE NEGATIVES (Missed bats)

Cost: protected species overlooked, incomplete EIA, compliance risk

BIOSONIC

3

KALEIDOSCOPE

49

Kaleidoscope: 16.3× more

Same data, very different outcomes. BioSonic delivers 98.7% F1 with only 19 errors across 5,314 predictions. Kaleidoscope Pro misses 10% of true bats and produces 17× more false calls — meaning consultants spend hours reviewing false detections that BioSonic never flags, and still miss protected species that Kaleidoscope fails to catch.

Source: Table 1 — Wilder Sensing validation study, Grassland + Paludiculture combined datasets (strict matching; Kaleidoscope Pro NoIDs not rewarded as true positives).

INDEPENDENT VALIDATION

Southern Jylland — Validated by Jan Drachmann Study

Five detectors, 11 species, one conclusion: BioSonic reaches identical ecological conclusions as manual analysis — same detector ranking (which monitoring locations had the most bat activity), same patterns — while catching more rare and protected species.

✨ NATTER’S BATS INSIGHT

Going through tens of thousands of files by hand, it’s easy to miss a few. Jan found 11 Natter’s bat calls — BioSonic found 100. On review, all 100 were validated as correct. That’s the payoff of AI on large datasets: nothing slips through.

Detector Ranking — Identical Results

D4 > D5 > D2 > D1 > D3 — same order whether counted by hand or by AI

Manual BioSonic

391

461

D1

1,156

1,688

D2

94

118

D3

23,527

25,514

D4

2,370

3,505

D5

Common Species

BioSonic detects more Pipistrel (+203%) and Trold (+16%)

Manual BioSonic

Dværg

13,556

10,669

Trold

11,433

13,248

Pipistrel

2,257

6,834

Rare & Protected Species

Separate scale — these are the ones that matter for Annex II compliance

Manual BioSonic

Vand

112

221

Frynse

11

100

Dam

0

Brun

40

41

Syd

55

67

BioSonic Finds More (%) — Per Species

Positive = BioSonic found more; Negative = Manual found more

+16%

+203%

+97%

+809%

+3%

+22%

-21%

0%

-100%

Dværg

Trold

Pipistrel

Vand

Frynse

Dam

Brun

Syd

Skimmel

Nyctaloid

Myotis

Comparison by Phonic Group

Grouped by echolocation type — differences shrink dramatically at group level.

Manual BioSonic

27,246

30,751

Pipistrellus

186

321

Myotis

106

108

Nyctaloid

log scale

Pipistrellus: Dværg + Trold + Pipistrel · Myotis: Vand + Frynse + Dam + Myotis sp. · Nyctaloid: Brun + Syd + Skimmel + Nyctaloid sp.

% of Total Detections per Species

Species with large % differences make up tiny fractions of total.

Manual % BioSonic %

49.2%

34.2%

Dværg

41.5%

42.5%

Trold

8.2%

21.9%

Pipistrel

0.4%

0.7%

Vand

0.04%

0.3%

Frynse

0%

Dam

0.2%

0.1%

Brun

0.2%

Syd

0%

Skimmel

0.04%

0%

Nyctaloid

0.2%

0%

Myotis sp.

Danish	Scientific	Phonic Group	Manual	BioSonic	Diff	Diff %	Expert Comment
Dværg	P. pygmaeus	Pipistrellus	13,556	10,669	-2,887	-21.3%	Redistribution within Pip group
Trold	P. pipistrellus	Pipistrellus	11,433	13,248	+1,815	+15.9%	“More or less the same”
Pipistrel	P. nathusii	Pipistrellus	2,257	6,834	+4,577	+202.8%	Key migration species — likely noise recovery
Vand	M. daubentonii	Myotis	112	221	+109	+97.3%	Confirmed correct, missed in noise files
Frynse	M. nattereri	Myotis	11	100	+89	+809.1%	“They’re more or less all correct” — expert verified
Dam	M. dasycneme	Myotis	0	0	0	0.0%	Agreement — no detections
Myotis sp.	Myotis sp.	Myotis	63	0	-63	-100%	BioSonic classified to species — “no problem”
Brun	N. noctula	Nyctaloid	40	41	+1	+2.5%	Near-perfect match
Syd	P. kuhlii / E. serotinus	Nyctaloid	55	67	+12	+21.8%	Close match for rare species
Skimmel	V. murinus	Nyctaloid	1	0	-1	-100%	Single call — statistically negligible
Nyctaloid sp.	Nyctaloid sp.	Nyctaloid	10	0	-10	-100%	BioSonic classified to species level
Total			27,538	31,180	+3,642	+13.2%

CUSTOMER QUOTE

Independent bat expert, Denmark

Jan Drachmann

Independent Bat Expert, Denmark

Manual: D4 > D5 > D2 > D1 > D3

BioSonic: D4 > D5 > D2 > D1 > D3

Identical ✓

“Detector 4 is the most important place for bats. The ranking is the same — both your and my analysis agree: D4 > D5 > D2 > D1 > D3. The ecological conclusions are identical.”

WHAT BIOSONIC DOES

BioSonic. One place for all your bat analysis.

Every feature your team needs to analyse bat sound at scale — AI detection, spectrogram review, team collaboration and report-ready visualisations in a single platform.

AI bat sound analysis

Leading accuracy for bat classification across the UK and Northern Europe.

7+ species per file

Multi-label detection — every species captured, never collapsed into one label.

Feeding & social call split

Classify behaviour — know what the bat is doing, not just that it passed.

Integrated spectrogram viewer

Integrated spectrogram review — verify any detection in one click.

Skip high-confidence files

The AI auto-approves, letting you focus on the rare calls below confidence thresholds.

Comments & collaboration

Comment on tricky files — ask colleagues for a second opinion without leaving the app.

Heatmaps & automatic weather data

Automatic weather integration and heatmaps — report-ready visualisations.

SEE IT IN ACTION

Spectrogram Viewer

Review detections, comment & collaborate

Species Heatmap

Detection density per location on satellite map

Activity Over Time

Total calls by species across monitoring period

Transect Mapping

GPS route with bat detections along the path

BAT ACTIVITY OVER TIME

Feeding, social calls & species activity across the night

SPECIES ACTIVITY OVER TIME

Each species has its own emergence & activity window

SOCIAL & FEEDING CALLS

Both spike sharply right at sunset

ACTIVITY HEATMAPS · BEHAVIOR × LOCATION

Feeding Activity where bats are hunting

Feeding buzzes cluster over wetlands, forest edges and coastal habitats — BioSonic identifies prime foraging zones for EIA planning.

Social Call Activity where roosts are

Social calls cluster tightly near roosts — BioSonic pinpoints roost locations within 500m, critical for windmill placement decisions.

INTERACTIVE BATMAP

BUSINESS CASE

327 hours saved per 2 TB — £21,255 back in the budget.

Real comparison from an ecology consultancy in south England, measuring time spent on sound analysis between Kaleidoscope Pro and BioSonic on the same 2 TB dataset.

TIME SAVED

66%

per 2 TB of sound analysis

HOURS SAVED

327h

out of 495h total manual effort

SAVINGS PER 2 TB

£21,255

327 h × £65/h

COST PER 2 TB

Kaleidoscope Pro (495h × £65)£32,175

BioSonic (168h × £65)£10,920

£21,255 saved per 2 TB of manual review cost

ECOLOGIST FEEDBACK

What bat workers think about BioSonic

Lars Michael Nielsen

Biologist, DMR

"It's seldom wrong, it's a pleasure using it and it's making my job easier."

Peter Sivertsen

Biologist

"What sets BioSonic apart is their exceptional accuracy, both in species identification and in filtering out noise. In this, they're truly at the top."

Luke Waddison

Ecological Consultant, WNIC

"Since joining BioSonic our analysis time has reduced significantly, freeing us up to take on different tasks. The system is a pleasure to use and the team are excellent in their support."

James Shipman

Senior Ecologist, Author, Jacobs

"This is all cutting edge stuff and the way bat work will go! I look forward to seeing it go strength to strength."

André Dabolins

CEO, EcoFauna

"Going through 100K files can take me 14 days. With BioSonic I review them in 2 days."

Geoff Carss

Wilder Sensing · + Annabel Jeffries

"We tested a range of existing solutions using real acoustic data, and BioSonic clearly stood out, achieving 98.7% accuracy (F1 score)."

Jacob Højgaard

PhD Ecology, NLABT

"BioSonic always impresses me with how well it detects bats, even from the faintest bat calls."

Sofia Berg

"Today we survey too short a time and risk missing key species. Together with BioSonic, we can keep track of even the most rare species."

Martin Pilstjärna

Ph.D Ecology, Impact Director, Qarlbo Biodiversity

"BioSonic provided us with excellent service throughout the data collection process and delivered well in relation to our agreement. I am more than inclined to recommend them as a provider of bioacustic data for biodiversity projects elsewhere."

Online comments on BioSonic

Unprompted, on the open internet — from ecology consultants who actually use the platform.

GEOFF CARSS · WILDER SENSING

BLOOM ECOLOGY

COWI

JAMES SHIPMAN · DMR

BIOSONIC

Let's see what BioSonic can do for your bat projects.

Start for free — or book a walkthrough.

Book a Demo Try 15 GB for free

Josef Carlson

CEO & Founder, BioSonic

josef.carlson@biosonic.se · biosonic.se

BARBAR echolocating behind grasshopper noise. Kaleidoscope sees only the most loud part of the sound (many times, that’s noise) — BioSonic finds the bat using image recognition on the spectrogram.

Same noise. Different conclusion.

Detector Ranking — Identical Results

Common Species

Rare & Protected Species

BioSonic Finds More (%) — Per Species

Comparison by Phonic Group

% of Total Detections per Species

BioSonic in Action