TECHNICAL LIBRARY / 536
Blade-Pass Frequency: A Useful Rotor Signature, Not an Immutable Fingerprint
Blade-pass frequency gives acoustic drone assessment a useful physical reference: rotating blades create repeated disturbances that can appear as related tones. That structure can reveal more than overall loudness, but it is neither fixed throughout flight nor exclusive to aircraft. For technical evaluators, its value lies in understanding what the observed pattern supports, how conditions change it and which conclusions still require separate evidence.
ARTICLE SUMMARY
A physical signature is a reference, not a complete verdict
BPF and its harmonics help explain rotor sound without reducing acoustic assessment to loudness. Their usefulness depends on rotor state, propagation and competing sounds. Intaratep et al. (2016) document tonal structure alongside broadband and motor contributions; Al-Emadi et al. (2021) and Kümmritz (2024) show why learned-model results must be read with their data context. Credible evaluation separates detecting activity, classifying it and establishing identity or payload, rather than treating one recognizable pattern as proof of all four.
Physical periodicity
Blade passage adds structure that loudness alone cannot describe
Blade-pass frequency, or BPF, describes how often rotating blades pass a fixed reference point. It connects a sound pattern to a physical event rather than to a label learned from recordings.
Repeated blade passage produces repeated aerodynamic disturbances. These can generate a tone associated with BPF and related tones called harmonics, whose frequencies are whole-number multiples of that reference. Rotor speed and blade count determine the blade-passage repetition, while the strength of different acoustic components depends on how sound is generated. A rotor recording is not simply one pure tone: it can include broadband sound, which spreads across frequencies, and contributions from motors and interacting flows.
Intaratep et al., in Experimental Study of Quadcopter Acoustics and Performance at Static Thrust Conditions (2016), observed prominent BPF and shaft-rate tones and harmonics, together with broadband and motor noise. Overall sound pressure level summarizes sound magnitude but does not describe those relationships. Two recordings can have similar levels yet different structures. This explains the value of periodicity without proving that level is useless, that BPF is the sole informative feature or that periodic evidence always outperforms every alternative.
Operating state
The repetition follows rotor speed, not a permanent aircraft label
A physical relationship can remain valid while its observed value changes. BPF moves when rotor speed changes; flight control does not make it immutable.
Many multirotors control thrust and attitude by changing rotor speeds, and different rotors need not run at exactly the same speed. Acceleration, changes in orientation and aerodynamic loading can alter the resulting tones and their relative prominence. Intaratep et al. (2016) found slight speed differences between rotors and associated secondary peaks in their static-thrust measurements. That evidence already cautions against representing an entire multirotor as one perfectly fixed harmonic series, even before considering changing flight conditions.
Hover can provide a comparatively steady source in some circumstances, but it is not universally the easiest acoustic case. A hovering aircraft can still adjust its rotors in response to disturbances, while background sound and propagation affect what is observable. Wind can influence both the acoustic environment and the aircraft's operating state. Added payload can also change required thrust and rotor operation. These influences need not produce the same change on every aircraft, but they rule out assuming that wind, manoeuvre or payload leaves BPF unaffected.
Observability
The sound produced is not identical to the sound received
A rotor can continue producing periodic sound while the available recording reveals less of it. Source behaviour and the observability of that behaviour are different questions.
ISO 9613-1:1993 describes atmospheric absorption as dependent on sound frequency and air temperature, humidity and pressure. Absorption can therefore change the relative prominence of acoustic components rather than simply making every part equally quieter. Outdoor sound is also affected by reflections and refraction, which the standard distinguishes from absorption. Background activity can mask components, and source orientation can change the sound reaching a listener. A recognizable structure in one recording need not remain equally recognizable in another.
This distinction prevents a physical explanation from becoming an unsupported performance claim. Reduced harmonic detail may leave evidence consistent with rotating machinery while weakening a more specific interpretation. It does not establish a universal boundary at which detection remains possible but classification becomes impossible. Both depend on the source, the remaining information and the assessed system. Likewise, absence of recognizable BPF in an observation is not proof that no aircraft was present; the relevant evidence may simply have been obscured or unavailable.
Competing sources
Harmonics are informative, but not exclusive to drones
Periodicity is a consequence of repeated physical motion. Aircraft rotors share that property with non-aircraft machinery, so a harmonic pattern is a clue rather than a unique signature.
Fans and other rotating equipment can produce blade-passage tones, while pumps and compressors can contribute recurring mechanical or aerodynamic sound. Intaratep et al. (2016) explicitly relate aspects of the measured quadcopter spectrum to the acoustic behaviour of axial-flow machines. An industrial background is therefore more than undifferentiated noise: some of it can contain structure resembling the evidence of interest. Calling a tone rotor-like does not by itself establish that its source is airborne, a drone or a particular aircraft model.
Harmonic relationships can still add useful context beyond a single tone. Their relative prominence and variation can help distinguish examples when those differences are represented in the evaluation. But the same aircraft can sound different as its state changes, and unrelated sources can share some features. Even a familiar-looking harmonic pattern needs an interpretation supported by relevant comparisons. Real machinery examples matter because a model may separate drones from generic background clips while remaining uncertain about mechanically similar sounds.
Meaning of a label
Aircraft class, individual identity and payload are separate claims
A useful acoustic label names the proposition it supports. Evidence of aircraft-like activity is not equivalent to a model name, a unique airframe identity or a statement about what it carries.
Classification assigns an observation to a defined category. That category might be drone versus non-drone, a broad aircraft class or one of the models included in a study. These are different tasks. Propeller-driven fixed-wing aircraft and multirotors can both produce blade-passage sound, but their configurations differ; a result on one does not automatically establish a result on the other. Nor is there a general rule that one configuration always has a cleaner or more identifiable acoustic pattern.
Individual identity means distinguishing one particular aircraft from other examples, including others of the same model. BPF alone does not establish that uniqueness. Payload inference is different again: a changed sound could reflect loading, manoeuvre, propeller condition or the observation environment. An association measured under controlled conditions would not by itself identify the payload of an unfamiliar aircraft, much less its contents. A report should not quietly expand a tested class label into an identity or payload conclusion.
Learned-model evidence
Research supports learning from sound, not assuming what every model learns
Machine learning can use recurring acoustic differences, but a successful output does not prove that a model relies only on BPF. Its evidence may include other sound characteristics or recording-specific cues.
Al-Emadi et al., in their Sensors study on audio-based drone detection and identification (2021), evaluated deep learning with recorded and artificially generated drone audio. They found useful results for familiar drones and examined previously unseen drone types separately. The classifier was less effective on entirely new types, while the hybrid recorded/generated training material improved the unseen case relative to its comparison. This is concrete evidence that training-data diversity can matter, not proof of unrestricted recognition of any new aircraft.
The original drone recordings were collected in a quiet indoor environment; additional material included background-augmented examples. Those conditions help define the meaning of the reported result. The unseen-type experiment concerns recognizing drone activity from unfamiliar types, not establishing the exact identity of an unknown model. Neither this study nor rotor physics proves that all learned models merely extract BPF. A model's robustness and the information it actually uses require evidence beyond an appealing explanation of the sound source.
Generalization
A stronger result on related data is not the same as transfer
Generalization means remaining useful beyond the examples and conditions that shaped a model. Shared rotor physics gives a reason to investigate transfer, not evidence that transfer has already succeeded.
Kümmritz's The Sound of Surveillance, published in Drones (2024), reports classification improvements from acoustic data augmentation and examines defined drone classes. The paper also identifies a major weakness: training and validation material came from the same underlying measurements. The author explains that this relationship can encourage overfitting and overestimate performance on genuinely new data. The improvement remains meaningful within that evaluation, but its scope is narrower than a claim of dependable performance across unfamiliar environments.
New aircraft, new recording contexts and new background activity are distinct changes. Success on one does not resolve the others. Evaluation evidence should make clear which differences were genuinely represented and which examples remained closely related to training material. Findings should also distinguish class-specific errors from an overall score. A model might recognize broad drone-like activity without resolving type, or classify familiar recordings well while confusing new machinery sounds. Generalization is demonstrated by those comparisons, not inherited automatically from BPF's physical origin.
Evaluation responsibility
Keep detection, classification and judgment accountable to different evidence
The commercial value of a rotor signature is clearest when each conclusion has an explicit scope. Detecting an event, assigning a class and evaluating a product should not be collapsed into one confident-looking label.
Detection concerns whether the evidence supports activity of interest, with false reports and missed events as separate outcomes. Classification concerns whether a supported observation belongs to a defined category; ambiguous and unfamiliar cases remain important. Evaluation concerns whether either capability was demonstrated for the proposed product and stated conditions against reliable reference information. These are responsibilities to distinguish, not a prescribed processing sequence. A successful detection does not excuse a wrong type label, and a correct label on selected examples does not establish dependable detection elsewhere.
For technical and business review, a defensible claim connects the named outcome to the evaluated system version, data context and limitations. It also explains unsuccessful or unresolved cases rather than hiding them inside a headline score. Published research supplies useful questions and physical context, but it does not validate a New Ark product. BPF earns its place as an interpretable reference when it helps explain the available evidence. Its value is weakened, not strengthened, by presenting it as an immutable fingerprint or a substitute for product-specific evaluation.
SOURCE ATTRIBUTION
The evidence behind this discussion
Sources are named here without outbound links. Research findings and reported events do not establish New Ark product performance or site suitability.
Intaratep et al., AIAA: Experimental Study of Quadcopter Acoustics and Performance at Static Thrust Conditions · 2016
Reports BPF and shaft-rate tones and harmonics alongside broadband and motor noise in static-thrust quadcopter tests; discusses rotor-speed differences, propeller-dependent sound and rotor interactions. These observations do not establish field detection performance, unique identity or payload inference.
ISO, ISO 9613-1: Acoustics—Attenuation of Sound During Propagation Outdoors, Part 1: Calculation of the Absorption of Sound by the Atmosphere · 1993-06
Describes atmospheric absorption as dependent on sound frequency and air temperature, humidity and pressure; distinguishes absorption from other propagation effects. It is not a drone detection standard and supplies no product range or classification guarantee.
Al-Emadi, Al-Ali and Al-Ali, Sensors: Audio-Based Drone Detection and Identification Using Deep Learning Techniques with Dataset Enhancement through Generative Adversarial Networks · 2021-07-21
Evaluates familiar and unseen drone types using recorded and generated audio; reports reduced effectiveness on new types and benefits from hybrid training material in the unseen case. Quiet indoor source recordings and augmented data limit direct inference about unfamiliar environments; unseen-type detection is not unique-aircraft identification.
Kümmritz, Drones: The Sound of Surveillance—Enhancing Machine Learning-Driven Drone Detection with Advanced Acoustic Augmentation · 2024-03-19
Reports augmentation-related classification improvements and discusses scene-specific acoustic effects. Explicitly identifies the close relationship between training and validation measurements as a major limitation that may overestimate generalization. Defined class labels and validation gains do not prove identity, payload or New Ark product performance.
PUBLIC REFERENCE BOUNDARY
A clearer project discussion, not an operating instruction.
01The information roles described here depend on lawful purpose, environmental conditions and independent project evidence. No detection, coverage, identification, interoperability or safety outcome is guaranteed.
02Passive reception is not a hazardous-area certification. Qualified professionals must review equipment approval, site permissions, protected services and the written scope before any engineering decision.
03This article supplies no sensor layout, signal-processing method, threshold, target-selection method, RF-control procedure or operational response instruction. Data rights and decision authority remain with the responsible parties.
ARTICLE FAQ / 536
Questions on Blade-Pass Frequency: A Useful Rotor Signature, Not an Immutable Fingerprint
Answers address evidence, limitations and written project scope, not operational instructions.
RECORDACOUSTIC TECHNICAL LIBRARY
PROJECT DISCUSSION
Take the evidence questions into a written project scope.
Explore the PCBA delivery references or discuss your procurement, compliance and information-rights questions with New Ark. Available documents are project references, not a performance or certification commitment.
Review the full ARC-PERIMETER perimeter solution.
A consolidated reference from microphone-array networking to layered perimeters and project validation.
Microphone Array Drone Detection: Technology, Applications and C-UAS Integration
Microphone-array drone detection combines a physical sensing context with an integration and governance question. This guide brings together technology, application, and existing C-UAS-system integration without treating a public description as a design or operating manual.
Passive Acoustic Drone Detection for Low-Altitude Security
Low-altitude security is a site-specific information challenge. This guide explains how passive acoustic inputs may be considered within perimeter review, safe-offset planning, and multi-sensor corroboration without publishing deployment instructions or field-performance claims.