🦇

Bat Frequency Detector

This bat frequency detector lets you watch high-frequency activity scroll across a live ultrasonic spectrogram and use a heterodyne listen mode that shifts captured energy down into the audible range — the same technique a real bat echolocation detector uses. The tool reads your microphone’s true Nyquist ceiling at runtime and tells you, plainly, how much of the bat band it can actually reach.

⚠ The honest part first: a normal browser microphone CANNOT detect most bat calls. Bat echolocation spans roughly 20–120 kHz, but a standard mic samples at 44.1 or 48 kHz, so by the Nyquist limit it can only capture sound up to about 22–24 kHz — and built-in mics usually roll off hard above ~16–18 kHz before that. Only the very lowest components of a few low-frequency species (e.g. some Tadarida or Eptesicus search calls dipping toward ~20–22 kHz) might brush into range. For genuine bat detection you need a USB ultrasonic microphone sampling 192–384 kHz or a dedicated hardware bat detector. This tool is honest about that and shows your real limit. Nothing is recorded or uploaded.

Microphone

Idle — press Start to allow your microphone and see the live high-frequency spectrogram.

🔍 Microphone capability — press Start to measure

Your microphone’s sample rate sets a hard ceiling (Nyquist = sample rate ÷ 2) on the highest frequency it can capture. We read it the instant you start and tell you whether any part of the bat band is reachable.

Sample rate—

Nyquist ceiling (max capturable)—

Reachable share of 20–120 kHz bat band—

Spectrogram band

FFT size

Heterodyne tuning

Tuned to 20.0 kHz — energy near this frequency is shifted down to a soft audible tone.

Listen volume

Heterodyne output level. Start low.

Scrolling spectrogram: time flows left→right, the vertical axis is frequency (Hz/kHz), brightness is relative level (dBFS, uncalibrated). The dashed cyan line marks the heterodyne tuning; the red dashed line marks the Nyquist ceiling — nothing above it exists in the data.

Strongest high-frequency peaks

The clearest tones detected above the noise floor, with how they relate to bat calls. Remember: anything genuinely from a bat would have to sit below your Nyquist ceiling to appear here at all — in practice most readings near the top of the range are electronics (CRT/SMPS whine), insects, or aliasing, not bats.

#	Frequency	Rel. level	Note
Start the detector to list high-frequency peaks.

Heterodyne listen mode

A real heterodyne bat detector mixes the incoming ultrasound with a tunable oscillator so a call near the tuning frequency drops to an audible warble. This tool does the honest equivalent in software — but it can only shift energy that is already within your Nyquist ceiling. If your mic tops out at ~22 kHz, tuning to 20 kHz can reveal a 20–22 kHz source as a low tone, but a 45 kHz pipistrelle call simply is not in the captured signal and cannot be recovered.

Heterodyne off. Press “Heterodyne listen” while running, then tune near the top of your reachable band.

Safety: the heterodyne output is a synthesized audible tone played through your speakers/headphones. Keep the listen volume low to start — high-frequency tones at high level are unpleasant and can be fatiguing. The microphone is never routed to your speakers, so there is no feedback path.

How to Use the Bat Frequency Detector

1. Start and read your ceiling. Click Start and grant microphone access. The capability panel immediately shows your sample rate, your Nyquist ceiling (the highest frequency physically capturable), and what share of the 20–120 kHz bat band that ceiling can reach. For nearly all built-in and USB consumer mics the answer is “almost none”.

2. Watch the spectrogram. High-frequency activity scrolls left to right. Switch the band to “High band” to zoom the display onto 10 kHz–Nyquist where any reachable bat energy would appear. The red dashed line is your ceiling — treat the empty space above it as a hard wall, not silence.

3. Try heterodyne listen. Press Heterodyne listen, set a low volume, and drag the tuning slider toward the top of your reachable band. Energy near the tuning frequency is shifted down to a soft audible tone. This is genuinely useful for hearing 15–22 kHz sources (CRT whine, pest deterrents, some insects) that you may not hear directly — just understand it cannot conjure calls that were never captured.

4. For real bats, use real hardware. See the hardware guide below. A USB ultrasonic microphone (192–384 kHz) or a dedicated heterodyne/frequency-division bat detector is the only way to reliably detect and identify echolocating bats.

Understanding Your Results

The Nyquist ceiling is the whole story

The Nyquist–Shannon sampling theorem says a digital system can only represent frequencies up to half its sample rate. A 44.1 kHz microphone tops out at 22,050 Hz; a 48 kHz mic at 24,000 Hz. There is no setting, no FFT size, and no software trick that exceeds this — energy above the ceiling is either filtered out or folds back (aliases) into the captured range as a false, lower-frequency artefact. The tool reads audioContext.sampleRate at runtime and draws that ceiling so you can see exactly where the wall is.

Why most bat calls are simply absent

Even where a mic could theoretically reach ~22 kHz, two more things work against you. First, the bat band starts around 20 kHz and runs up to ~120 kHz, so the overwhelming majority of call energy is far above any consumer ceiling. Second, built-in laptop and phone microphones, plus any OS high-pass filtering, attenuate steeply above ~16–18 kHz — a roll-off you can verify by running a mosquito frequency hearing test to find the personal upper limit your own ears (and mic) can reach. The practical result: a bright streak high on the spectrogram is almost always electronics or an orthopteran insect (cricket, katydid), not a bat. If you want to actively sweep through the high-frequency band looking for a specific source, the frequency sweep detector lets you scan the full range methodically.

What the heterodyne tone actually represents

The heterodyne output is a level-following audible tone driven by how much energy sits near your tuning frequency — a faithful software analogue of a hardware heterodyne detector, but bounded by the same ceiling. The mixing process is the same principle that field bioacousticians use in handheld detectors: multiply the incoming signal by a local oscillator, then low-pass filter to extract the difference frequency. It is an honest down-shift of captured energy, not a measurement of any specific bat species. Treat it as a listening aid for sounds at the edge of human hearing, and treat any species inference as a guess unless you are using a true ultrasonic microphone capable of full-spectrum bat echolocation recording.

Bat Call & Hardware Reference

Typical echolocation call bands by species

Published-typical values for a few well-studied species. Bat call frequency varies with behaviour (search vs approach vs feeding-buzz phase), habitat and individual, so these are ranges, not fixed tones. The right-hand column shows whether any part of the band could ever reach a 48 kHz mic (Nyquist 24 kHz) — for almost every species the answer is no.

Echolocation call frequency bands (kHz). FM = frequency-modulated sweep; CF = constant-frequency; QCF = quasi-constant frequency.
Species	Call type	Typical band	Within ~24 kHz ceiling?
Big brown bat (Eptesicus fuscus)	FM sweep	~20–65 kHz (search F_min ~22–30 kHz)	Only the lowest tail (~20–22 kHz), rarely
Mexican free-tailed bat (Tadarida brasiliensis)	Shallow FM / QCF	~20–75 kHz (search ~20–30 kHz)	Only the lowest tail (~20–24 kHz), rarely
Little brown bat (Myotis lucifugus)	FM sweep	~40–80 kHz (peak energy ~45 kHz)	No
Common pipistrelle (Pipistrellus pipistrellus)	FM/QCF	peak ~43–49 kHz (~45 kHz)	No
Greater horseshoe bat (Rhinolophus ferrumequinum)	CF	~82 kHz constant-frequency component	No

Sources: big-brown / little-brown call figures — Animal Diversity Web and Surlykke & Moss, Echolocation behavior of big brown bats (2000); free-tailed search-phase ~20–30 kHz — published Tadarida brasiliensis echolocation studies; common pipistrelle ~45 kHz and greater horseshoe ~82 kHz CF — UK/European bat-group echolocation keys. Values are typical published ranges; treat as approximate.

Hardware that actually detects bats

To capture real echolocation you need a microphone and digitizer that sample far above the audio range. Common options:

USB ultrasonic microphones sampling 192–384 kHz (Nyquist 96–192 kHz) feeding bat-analysis software — the practical entry point for full-spectrum recording.
Heterodyne bat detectors — a tunable handheld that mixes ultrasound down to audio in real time (like this tool, but with a true ultrasonic mic); cheap and great for live listening, narrowband.
Frequency-division detectors — divide the incoming frequency by a fixed factor so the whole band drops into audio at once; broadband, lower fidelity.
Full-spectrum / time-expansion detectors — record the true ultrasonic waveform for spectrogram analysis and species ID; the standard for surveys.

A normal phone or laptop microphone is none of these. This page is the honest, free bat detector online that shows you your real ceiling and lets you listen to the top edge of human hearing — not a substitute for dedicated hardware.

Frequently Asked Questions

Can this actually detect bats with my phone or laptop mic?

Almost never. Bat echolocation spans roughly 20–120 kHz, but a normal microphone samples at 44.1 or 48 kHz, so by the Nyquist limit it can only capture sound up to about 22–24 kHz — and most built-in mics roll off hard above 16–18 kHz before that. Only the very lowest-frequency tail of a few species (some big brown or Mexican free-tailed search calls dipping toward ~20–22 kHz) might brush into range, and even then usually too faint to read. For genuine bat detection you need a USB ultrasonic microphone (192–384 kHz) or a dedicated hardware bat detector.

What is the Nyquist ceiling and why does it matter so much?

The Nyquist–Shannon sampling theorem says a digital audio system can only represent frequencies up to half its sample rate. A 44.1 kHz mic tops out at 22,050 Hz; a 48 kHz mic at 24,000 Hz. No FFT size or software setting can exceed this — energy above the ceiling is filtered out or aliases back into the captured range as a false lower tone. The tool reads your microphone's actual sample rate at runtime and shows the real ceiling so you can see exactly where the hard wall is.

How does the heterodyne listen mode work, and what can it really do?

A hardware heterodyne bat detector mixes incoming ultrasound with a tunable oscillator so a call near the tuning frequency drops to an audible warble. This tool does the honest software equivalent: it shifts energy near your tuning frequency down to a soft audible tone whose level follows how much energy is present. The catch is that it can only shift energy already within your Nyquist ceiling — it can reveal a 20–22 kHz source as a low tone, but it cannot recover a 45 kHz pipistrelle call that was never captured. It is a listening aid for the top edge of human hearing, not a way to beat the sample-rate limit.

What frequencies do common bats echolocate at?

It varies by species and behaviour. The big brown bat (Eptesicus fuscus) sweeps roughly 20–65 kHz with search-phase minimums around 22–30 kHz. The Mexican free-tailed bat (Tadarida brasiliensis) ranges ~20–75 kHz with search calls near 20–30 kHz. The little brown bat (Myotis lucifugus) sweeps ~40–80 kHz with peak energy near 45 kHz. The common pipistrelle (Pipistrellus pipistrellus) peaks around 45 kHz, and the greater horseshoe bat (Rhinolophus ferrumequinum) uses a constant-frequency component near 82 kHz. Only the lowest tails of the lowest-frequency species ever approach a consumer microphone's ceiling.

If I see a bright high-frequency streak, is that a bat?

Probably not. Because almost no bat energy reaches a consumer mic's ceiling, a strong tone near the top of the spectrogram is far more likely to be electronics — CRT/monitor whine around 15.7 kHz, switched-mode power supply noise near 20 kHz — or insects such as crickets and katydids, or an aliasing artefact folding down from above Nyquist. Genuine bat identification requires a true ultrasonic microphone and full-spectrum analysis.

Is my audio recorded or uploaded?

No. The microphone signal is analysed live in your browser to draw the spectrogram and drive the heterodyne tone, then discarded — nothing is recorded, saved, or transmitted. The microphone connects only to an analyser node and is never routed to your speakers, so there is no feedback path. The microphone is released the moment you press Stop or close the tab.

Can I use this tool to detect ultrasonic pest repellers or CRT monitor whine?

Yes — this is arguably where a normal laptop or phone microphone is most useful with this tool. Plug-in ultrasonic pest repellers typically emit tones between 15 kHz and 22 kHz, and old CRT monitors produce a horizontal-scan whine near 15.7 kHz; both fall within or just below the Nyquist ceiling of a 44.1/48 kHz mic. Switch the spectrogram to “High band” mode and use heterodyne listen tuned to 15–22 kHz to detect these sources clearly — something a consumer mic genuinely can do, unlike true bat echolocation.

What do insect and bat sounds look like in the 15–22 kHz range, and how do I tell them apart?

Many orthopteran insects — crickets, katydids, and some cicadas — produce species calls that extend up to 20–22 kHz, sometimes appearing as sustained, nearly constant-frequency bands on the spectrogram. Bat echolocation sweeps are brief (<100 ms), frequency-modulated chirps that slope steeply from high to low frequency and repeat at call intervals of 50–500 ms. If you see a steady horizontal stripe it is more likely an insect or an electronic source; a brief downward-sloping streak that repeats rhythmically is structurally more like a bat call — though at this low frequency it could still be an insect. Without a true ultrasonic microphone, species identification is not possible.

What is frequency-division detection, and how does it differ from heterodyne?

Frequency-division detectors divide the incoming ultrasonic frequency by a fixed integer (commonly 10 or 16) so the entire bat band drops into the audible range at once — broadband but lower resolution. Heterodyne detectors (like this tool’s listen mode) mix the signal with a tunable local oscillator so only energy near the tuning frequency becomes audible as a warble — narrow-band but high fidelity for the chosen frequency. A third approach, time-expansion, records the true ultrasonic waveform and plays it back slowed by 10x. This tool implements software heterodyne, which lets you sweep the tuning slider to probe different parts of your reachable band. You can also compare the raw frequency content using the FFT frequency analyzer to complement heterodyne listening.

Does increasing the FFT size let me see higher frequencies?

No — FFT size affects frequency resolution, not the frequency ceiling. A larger FFT (e.g. 8192 vs 2048) divides the 0–Nyquist range into more bins, making closely spaced tones easier to separate, but it does not extend the spectrum above the Nyquist limit. The tradeoff is temporal resolution: an 8192-point FFT at 48 kHz takes about 170 ms of audio per frame, so fast-sweeping bat calls may smear across multiple bins. For bat work, 4096 is a reasonable balance; use 2048 if you want sharper time resolution for studying call structure.

🔊