Predict the noise radiated by multirotor UAVs — consumer quadcopters, industrial inspection drones, delivery UAS and eVTOLs — from blade passage frequency (BPF) and a spherical-spreading distance model. Adjust rotor size, RPM, thrust and observer geometry to see tip Mach number, 1 m SPL, observer dBA and regulatory exceedance update in real time.
Parameters
Drone class
Loads representative values for the selected category
Number of rotors N
Rotor radius R
cm
RPM
Thrust / rotor T
N
Observer distance R_obs
m
Observer angle θ
°
0° = directly below the aircraft, 90° = horizontal
Results
—
Rotor tip speed (m/s)
—
Tip Mach number
—
BPF (Hz)
—
1 m SPL (dB)
—
Observer dBA
—
Exceedance (dB)
—
Drone, sound field and observer
The quadcopter at the centre radiates concentric sound waves. Colour encodes local SPL (green = quiet, red = loud); the observer's dBA appears in the bottom-right corner.
Spherical-spreading distance attenuation (point source) and directivity correction L_dir. A-weighting is about −10 dB for BPF in the 100–500 Hz band.
Drone (Quadcopter) Noise Prediction — BPF and dBA Distance Attenuation
🙋
Drones sound louder than they look. What is that constant buzzing actually made of?
🎓
Good question. Drone noise has three ingredients. The most obvious is the BPF (Blade Passage Frequency) tone — the periodic chop of the blades cutting through the air, at BPF = N_b·RPM/60. A consumer quad with 2 blades at 6000 rpm sits at 200 Hz; that is exactly the "buzz" you hear. On top of that you get broadband noise from the turbulent boundary layer and tip vortices, and the higher-order harmonics from blade thickness and loading. All three overlap, but the ear locks onto the BPF spike.
🙋
200 Hz is supposed to be low-pitched, right? Why does it still carry so far?
🎓
Right — and 100-500 Hz happens to be the band where the A-weighting curve treats sound as loudest. Low frequencies also lose less energy to the atmosphere and pass through walls more easily. That is why the noise lingers after the drone disappears from view. Drag the observer distance from 1 m to 50 m: under the point-source model the level drops only 6 dB per doubling, so 50 m gives just −34 dB vs 1 m. A residential night-time limit of 45 dB is easy to bust.
🙋
So what do designers actually change to make a drone quieter?
🎓
The most powerful lever is the tip Mach number M_tip. Noise rises with M_tip to the 4th-6th power, so dropping from 0.4 to 0.3 buys you 5-10 dB. In practice this means "big rotor, slow RPM". DJI Mavic 3 sits at R = 13 cm × 6000 rpm → M_tip ≈ 0.24, giving 65 dBA at 1 m on hover. The Joby S4 eVTOL uses R = 1.5 m × 1500 rpm → M_tip ≈ 0.45. Beyond that, blade tip shape (winglets, sweep), phase-synchronised rotors (DJI Phase Shift) and ducted rotors (Lilium Jet) all help.
🙋
eVTOLs are those "flying taxis" Uber kept talking about. Can they really be quiet enough to land in a city?
🎓
That is the core challenge of Urban Air Mobility (UAM). A traditional helicopter does 90 dBA at 100 ft on take-off — far too loud for a residential vertiport. Joby targets 65 dBA at 100 ft, roughly 1/16 the acoustic power of a helicopter, achieved through ducts, large tilt-rotors and low RPM. Real-world acceptance probably needs under 60 dBA, and the FAA NPRM / EASA Special Condition VTOL discussions hover around this range. Residential delivery drone adoption will likewise be gated by whether they can hit 45 dBA at night.
🙋
Is the "exceedance" stat in this tool measured against that 55 dBA number?
🎓
Yes — 55 dBA is the interim threshold the tool uses for residential delivery. With the default consumer drone at 1 m you get 60 dBA, exceedance +5 dB. Move the observer to 30 m and dBA drops to about 30 — clearly safe. The full ICAO Chapter 10 / FAA Part 36 rules are more nuanced (take-off, flyover, sideline separately), but this simple comparison is usually enough for an early feasibility check.
FAQ
A spinning rotor radiates a tone at the blade passage frequency BPF = N_b·RPM/60. For a consumer drone this is typically 100-400 Hz, right where human hearing is most sensitive. A signal that contains a strong pure tone sounds harsher than broadband noise of the same dBA — so much so that ISO ratings add a "tonality penalty". This is the main reason a small quadcopter still sounds loud from far away. The tool visualises the BPF and its contribution to dBA.
Under a point-source spherical-spreading model, sound pressure level drops by 6 dB each time the observer distance R doubles: SPL(R) = SPL_1m − 20·log10(R). A hover noise of 70 dB at 1 m becomes 58 dB at 4 m, 46 dB at 16 m and 30 dB at 100 m. Ground reflection, atmospheric absorption and wind shift this further; a typical residential night-time limit is around 45 dB.
Noise grows roughly with the 4th-6th power of tip Mach number M_tip ∝ Ω·R, so commercial drones keep M_tip ≤ 0.5 and low-noise designs target 0.3-0.4. DJI Mavic 3 sits around 0.24, Joby S4 eVTOL around 0.45. Raising M_tip from 0.3 to 0.4 alone adds 5-10 dB, so the standard low-noise recipe is "large rotor, slow RPM", which also improves hover figure of merit.
Type certification is being driven by ICAO Chapter 10, EASA Special Condition VTOL and the FAA NPRM. Joby and Lilium target 65 dBA at 100 ft (≈30 m) during take-off and landing. For residential delivery drones, ground-level 55 dBA is often the threshold for approval. This tool uses 55 dBA as the baseline and reports the exceedance, which is useful for early-stage regulatory feasibility.
Real-world applications
Consumer / cinema drones: DJI Mavic- and Autel EVO-class drones produce about 65 dBA at 1 m while hovering — roughly conversational level. Outdoor shoots become comfortable once the operator is more than ~5 m away. For indoor shoots or close-range residential filming, use this tool to build a noise map from observer position, drone distance and altitude so that complaints can be anticipated.
Delivery drones (Amazon Prime Air, Wing, etc.): Because they fly over homes, the de-facto certification bar is 50-55 dBA at ground level. Set the typical 50-100 m altitude and an ~80° observation angle (almost directly overhead). Run the tool with RPM and thrust corresponding to both empty and fully loaded payload to check that the worst case still clears the limit.
eVTOLs / flying taxis (Joby, Lilium, Vertical): Vertiport noise during take-off and landing is the gating issue for UAM. Joby targets 65 dBA at 100 ft for take-off and 45 dBA at 1000 ft on cruise. Use R = 30 m, θ = 30° (climbing departure) and sweep rotor radius / RPM to study the low-noise design space.
Industrial / inspection drones (Matrice 350, Skydio X10): Bridge, powerline and plant inspections often happen at 5-10 m, so worker hearing protection becomes the real concern. Above 70 dBA an 8-hour shift typically needs hearing protection. Use this tool to estimate dBA at the operator's position to plan duty cycles and PPE.
Common pitfalls
The first trap is confusing dB with dBA. dB is the physical sound pressure level; dBA includes the A-weighting that approximates human loudness perception. A BPF near 200 Hz carries an A-correction of about −10 dB, a 20 Hz subsonic tone about −50 dB, and a 1 kHz tone 0 dB. When an eVTOL paper says "65 dB", always check whether it is before or after A-weighting. Every regulation (55, 65 dBA, …) is A-weighted. This tool applies a fixed −10 dB A-weighting for simplicity, but a production analysis must weight the BPF spectrum properly.
The second trap is over-trusting the point-source model. SPL(R) = SPL_1m − 20·log10(R) is the ideal spherical spread, valid only when R is much larger than the rotor diameter (typically R > 5·D_rotor). Near-field positions (R < 1 m), ground-bounce-dominated locations and the near field of ducted rotors can introduce 5-10 dB errors. The model also ignores atmospheric absorption (significant at high frequency), wind refraction and multi-path reflections off buildings. Field validation should be cross-checked against ISO 3744 (hemi-anechoic) or ISO 7196 (long-range outdoor) measurements.
The third trap is the intuitive but wrong belief that "more rotors = quieter". You do reduce the per-rotor thrust, but in this formula (SPL_1m ∝ 10·log10(T·N)) what matters is the product T·N, so for the same total thrust the difference between 4 and 8 rotors is only a few dB. In real machines rotor interference (non-synchronised rotors spread the BPF) and the airframe weight penalty often dominate. The high-leverage moves remain "bigger rotors, slower RPM", optimised tip geometry and phase-synchronised rotors (DJI's Phase Shift Technology). Rotor count is a secondary tuning knob.
How to Use
Enter the number of rotors (typically 4 for quadcopter, 6 for hexacopter, 8 for octocopter).
Input rotor radius in centimeters (DJI Phantom 3: 15 cm, Freefly Alta X: 48 cm).
Set rotorational speed in RPM (consumer drones: 4000–8000 RPM; industrial heavy-lift: 2000–3500 RPM).
Specify thrust per rotor in Newtons (e.g., DJI M300 RTK per motor: ~380 N at hover).
Simulator calculates tip speed, blade passage frequency, and far-field SPL at 1 m reference distance.
Worked Example
DJI Matrice 300 RTK quadcopter: 4 rotors, 17.5 cm radius, 6500 RPM hover, 385 N thrust per motor. Rotor tip speed = 11.9 m/s (Mach 0.035), BPF = 433 Hz (4 × 6500/60). Predicted SPL at 1 m: 82 dB. Observer at 50 m urban site receives 62 dBA (accounting for A-weighting and spherical spreading). Exceeds typical residential limit of 55 dBA by 7 dB.
Practical Notes
Delivery operations near 60 m altitude: reduce SPL by ~16 dB due to distance; prioritize low-RPM, large-diameter rotors (e.g., 40 cm radius at 3000 RPM) to shift BPF below 200 Hz, away from human sensitivity peak.
Industrial inspection missions over construction sites (already 85 dBA ambient): verify compliance with local airspace noise ordinances; octocopters distribute thrust across 8 rotors, reducing individual rotor loading and broadband noise floor by 3–5 dB versus quadcopters.
Tip speed should not exceed Mach 0.6 (near 200 m/s on large rotors); supersonic tips trigger shock-cell radiation and rapid noise rise; monitor A-weighting dBA metric since rotor fundamental and harmonics concentrate near 500 Hz–2 kHz.