Eight Centimetres Across the Grain: Listening for Insects Inside Wood

1. Introduction: a method governed by physics

Most detection problems in this journal are biological: the animal hides, moves, or is present at a density nobody can estimate. This one is acoustics, and the constraints are stated as coefficients.

1.1 Why that is the whole subject

Because background noise falls with frequency and so does range. The instrument you can use is decided by which of those two problems the substrate imposes on you, and §§5 to 10 are that.

1.2 And one thing this method does that no other in this journal does

It can report that no measurement was possible, which is §§16 to 18.

2. What makes the sound

The signal source.

Sensors and algorithms leverage the sounds and vibrations generated by insects as they move, communicate, masticate, and procreate within a material.⁴

2.1 Four behaviours, not one

Which matters because they are not equally loud or equally frequent. A feeding larva produces repeated impulses; a resting one produces nothing.⁴

2.2 And feeding is the useful one

Because it is repetitive, mechanical and produces a characteristic train of impulses rather than a single event. A pattern-matching system has something to match against.²

Which is also why the method suits wood-boring larvae particularly: an animal whose entire existence consists of chewing through the substrate the sensor is attached to is an unusually cooperative target. That observation is ours.

2.3 And the term is imprecise

One paper notes that although many studies refer to these techniques as vibro-acoustic methods, this paper adopts the term acoustic to denote elastic waves within the tested materials.⁴

Which is worth holding onto: what is being detected is vibration travelling through a solid, not sound travelling through air, and every distance figure below is a distance through material.

3. The quiet at high frequency

The first half of the trade-off.

Acoustic emission sensors are successful because they are nondestructive and operate at high frequencies (>40 kHz) where there is negligible background noise to interfere with detection and interpretation of insect sounds.⁵

Similarly, ultrasonic sensors are particularly effective for detecting wood-boring pests because background noise is negligible at >20 kHz frequencies, and ultrasonic signals attenuate much less rapidly in wood than in air, grain, or soil.²

3.1 Nothing else in a building is making that noise

Traffic, plumbing, footsteps, ventilation and voices are all low-frequency. Above twenty kilohertz a building is close to silent, which means almost any signal detected there is biological.²

3.2 Which is a rare gift

Specificity supplied by the medium rather than by interpretation. There is no classifier to train, no threshold to set and no judgement to make, because at that frequency there is essentially nothing else to confuse the signal with.⁵

3.3 Which is the ideal condition for detection

High specificity for free, supplied by physics rather than by processing. Our detection articles have generally found specificity the hard part, and here it comes built in.

4. And the price of it

The second half.

The thousandfold increase in attenuation between 500 Hz and 120 kHz in air, with greater rates in soil and grain.¹

4.1 What attenuation is

The rate at which a vibration loses energy as it travels through a material, dissipated as heat through internal friction. It rises with frequency because a higher-frequency wave flexes the material more times over the same distance.

Which is why the relationship is a physical necessity rather than a property of any particular substance, and §5.1 is the consequence.

4.2 Which means the signal dies with distance

So the quieter band is also the band in which the insect must be nearly against the sensor.

4.3 And the alternative

The range of acoustic detection is much greater at frequencies <10 kHz, where low-frequency accelerometers have been used to detect larvae over one to two metres in grain.⁵

Which is the band in which everything else in the building is also audible.

5. The trade-off

Stated as one thing.

Go high and you get silence and no range. Go low and you get range and no silence. There is no frequency at which both problems are solved.¹²⁵

5.1 Which is not a technological limitation

It is a property of how elastic waves propagate. Better sensors, better processing and better software do not change the attenuation coefficient.

That observation is ours and it distinguishes this from most of the technologies this journal examines, where the limitation is usually that the work has not been done.

6. Which the substrate resolves for you

Because the coefficient depends on the material.

Soil and grain are described as highly effective insulators against high-frequency sound.¹ Wood has a low attenuation coefficient.¹

6.1 So the choice is made for you

In wood you can use the high-frequency band and get its specificity. In grain or soil you cannot, so you work low and accept the noise.

6.2 Which is an unusual kind of constraint

The operator does not choose a strategy and then find the right instrument. The material chooses the frequency band, the band determines whether background noise is a problem, and the problem determines what the result can mean.

Three consequences from one property of the substrate, and none of them is negotiable.

6.3 Which means the same insect needs two instruments

A beetle larva in a joist and the same species in a sack of flour are different detection problems with different physics, and our stored product article describes the species moving between those settings.

7. Wood is the good case

The figure.

Because of that low coefficient, ultrasound (i.e., frequencies >20 kHz) could be detected from termites over active spaces of up to 2.2 m in wood.¹

7.1 Why wood is different

Because it is a stiff, low-density, fibrous solid, which is a good combination for transmitting elastic waves. Grain is a heap of loose particles with air between them, and every particle boundary is an interface where energy is lost.

That explanation is ours and it accounts for why the same frequency behaves so differently in the two substrates.¹

7.2 Two metres is a useful distance

It is the span of a joist bay, and it means a sensor does not have to be placed on the infested member itself.

Which would make this a genuinely practical structural inspection tool, if §8 did not exist.

8. Eight centimetres

The number that changes the picture.

Signals were detected at metre distances from the sensor location along the wood grain, but only ca. 8 cm away across the wood grain.¹

8.1 Same wood, same frequency, two directions

Timber is anisotropic. It conducts vibration readily along the fibres and poorly across them, and the difference here is more than an order of magnitude.¹

8.2 So the two-metre figure is one-dimensional

It describes a line, not a sphere. A sensor on a joist listens along that joist and hears almost nothing in the adjacent one.

That reading is ours and it is the most practically important thing in this article.

9. Which makes placement a geometry problem

And the geometry is set by the framing.

Every structural member is a separate acoustic channel. The sensor covers the member it is on and does not cover the one beside it, so coverage is a function of how many members there are rather than of area.

9.1 Which reverses the usual scaling

Most survey methods get cheaper per unit area as the area grows, because a sweep covers ground. Here a larger building with more framing needs proportionally more readings, and a heavily framed wall needs more than a lightly framed one of the same size.

So the cost of a thorough acoustic survey scales with construction complexity rather than with floor area, which is the opposite of how inspections are priced.

9.2 Which is why sensor spacing is the open question

A 2026 paper identifies as a research priority the systematic experiments that would enable evidence-based sensor spacing recommendations for structural applications.³

Meaning they do not exist yet.

9.3 And joints make it worse

The same paper notes that a measurement taken on a piece of wood joined to an infested piece, rather than on a single piece of solid wood, constitutes an additional acoustic interface and poses suboptimal signal transmission conditions for the sensor.³

A building is made of joined pieces. Every joint is an interface.

10. Grain is the other case

Where the physics forces the low band and the results are still useful.

A system originally designed for termite detection was used to track the movement of one to twenty adult insects of four named species in corn meal and flour mix, and was also able to quantify the level of infestation.⁴

10.1 One to twenty individuals

Which is a remarkably low detection floor for a concealed population, and quantification rather than presence or absence is more than most methods in this journal manage.⁴

10.2 And the named species are ours

A cigarette beetle, a sawtoothed grain beetle, a carpet beetle and a flour beetle. Our packaging article covered the first, second and fourth, and our fabric pest article the third.⁴

10.3 And it works because grain is uniform

A bin is a homogeneous medium with no joints, no members and no anisotropy. Every direction attenuates equally, so coverage is a sphere rather than a line and §8 does not apply.

Which is why a substrate that is bad for the physics can be good for the application, and that observation is ours.

10.4 The caveat attached

That the method can only detect the movement of the insects and is highly susceptible to vibrational interference.⁴

Which is §13 and §15, stated by the source in one sentence.

11. The seven determinants

The review's list.

Efficacy depends on many factors, including the sensor type and frequency range, the substrate structure, the interface between sensor and substrate, the assessment duration, the size and behavior of the insect, and the distance between the insects and the sensors.²

11.1 Only two are under the operator's control

Sensor choice and assessment duration. The substrate, the interface geometry, the insect and the distance are given by the situation.²

That division is ours and it is why §12 matters.

11.2 Which is why results are hard to compare

Two studies reporting different detection ranges may differ in any of the seven, and a figure quoted without its conditions is close to meaningless.²

Which is why §26 lists the provenance of each number in this article.

11.3 And the interface is easy to get wrong

One review notes that microphones are useful sensors for airborne signals, but vibration sensors interface better with signals produced in solid substrates, such as soil, grain, or fibrous plant structures.²

12. Duration is the one operators control

And it is not arbitrary.

A 2026 paper identifies the need for experiments to establish species-specific measurement duration recommendations and detection thresholds.³

12.1 Why duration is a parameter at all

Because the insect has to do something audible while you are listening. Longer listening raises the chance of catching an active period, which means a short measurement and a long one are different tests rather than the same test done twice.

12.2 Which connects to temperature

Insect activity is temperature-dependent, so a measurement taken on a cold wall in January samples a less active population than the same measurement in July.

We found no source addressing that, and in this climate it is the obvious confounder. A negative acoustic reading in an unheated crawlspace in winter tells you considerably less than the same reading in summer, which is our inference and one we would want tested.

12.3 And the recommendations do not exist

So an operator listening for thirty seconds and an operator listening for ten minutes have no published basis for choosing either.³

13. It detects activity, not presence

The central epistemic point, and the sources make it.

The method can only detect the movement.⁴ And the 2026 validation paper's title concerns the detection of active wood-boring larvae specifically.⁷

13.1 Which is unusual among detection methods

A visual inspection detects the animal or its traces whether or not it is doing anything. A genetic method detects material it left behind, which our environmental DNA article found persists too long rather than too briefly.

Acoustics detects only the present tense, which is both its weakness and, for a question like whether a treatment worked, its advantage.

13.2 A quiescent insect is silent

Which covers eggs, pupae, moulting individuals, dormant larvae and anything cold enough to be inactive.

13.3 So a negative has a specific meaning

Nothing was moving audibly within range during the measurement window, which is a narrower claim than absence in three separate ways: audibly, within range, and during that window.

That parsing is ours.

14. Which is a different claim from a negative

And better than most.

Our article on what a zero means argued that a negative inspection result is uninterpretable without a stated detection probability, and that the trade routinely reports absence when it has only failed to detect.

14.1 Here the limitation is in the physics

An acoustic negative can be qualified precisely: the range was this, the duration was that, the substrate attenuates at this rate. Each qualification is a number somebody could in principle supply.

14.2 Which is what our detection article asked for

A detection probability attached to a negative result. Acoustics cannot supply the number yet, per §19, but it is the kind of method for which the number is obtainable in principle.

14.3 Which makes it a better negative than a visual one

Not because it detects more, but because its failure modes are enumerable. An inspector who found nothing cannot tell you the radius within which they would have found something.

15. The noise problem

The practical obstacle in the low band.

Problems in distinguishing sounds produced by target species from other sounds have hindered usage of acoustic devices, but new devices and signal processing methods have greatly increased the reliability of detection.²

15.1 Why filtering alone fails

A filter removes a frequency band, which works when the noise and the signal occupy different bands. Below ten kilohertz they do not: the building noise and the insect noise are in the same range.

So the separation has to be made on something other than frequency, which is what §15.2 describes.

15.2 The approach that works

One new method considers spectral and temporal pattern features that prominently appear in insect sounds but not in background noise, and vice versa.²

Which is pattern recognition rather than filtering: insect signals have a characteristic temporal structure that a pump or a passing vehicle does not.

15.3 And the commercial version

A patent describes an apparatus capable of comparing the detected sounds to a library of previously recorded known termite sounds, using pattern matching or recognition to find matches.⁶

16. And the honest response to it

The sentence that made us want to write this article.

If, based on this experience, noise cannot be distinguished from larval activity, a measurement is not possible.³

16.1 Not a negative result

Not a reading of zero, not an inconclusive finding, not a low-confidence absence. No measurement.³

16.2 And it is the third option almost nobody offers

A client asks whether there are insects in the wall. The available answers are usually yes and no. This method supplies a third, which is that the conditions here did not permit an answer.³

16.3 Which is a design decision

The instrument could have been built to return something in that situation, and returning something would be commercially easier. Declining to is a choice about what the method is willing to assert.

17. A third kind of result

Which this journal has not encountered before.

Our detection articles have described two failure modes. The uninterpretable zero, where absence cannot be distinguished from non-detection. And the uninterpretable positive, where a signal is detected but its meaning in time and space is unclear, which our environmental DNA article described.

17.1 This is neither

It is a declared non-measurement: the instrument reporting that the conditions for producing a result were not met.

17.2 Which is what a good instrument does

A laboratory assay with a failed control is reported as a failed assay rather than as a negative. Almost nothing in structural pest inspection works that way, and this is the exception we have found.

That framing is ours and we intend to use it as a standard elsewhere.

18. Why that matters to this journal

Briefly, because it connects to everything.

Our article on the brown recluse found researchers building an exclusion instrument because confirmation was impossible, and we concluded that a well-built exclusion is the appropriate response to an unanswerable question rather than a consolation prize.

18.1 This is the same principle, one step further

There the tool said what a thing is not. Here the tool says that on this occasion it cannot say anything at all.³

18.2 And it changes what a report should say

Our inspection reporting articles have argued that a finding of no evidence should state the method's limits alongside it. The stronger version available here is that some conditions do not permit a finding at all, and that saying so is a result rather than an evasion.

18.3 Both are refusals

And both are more useful than an answer that does not mean anything.

19. What is still missing

From the 2026 paper's own limitations, which are extensive.

Signal attenuation characteristics, effective detection range, and the relative impact of background noise may all differ substantially in field conditions.³

A systematic experimental investigation of range and attenuation in structural elements of varying dimensions remains a priority for future research.³

19.1 Field conditions may all differ substantially

Which is the phrase to notice. The validation work is laboratory work, and the authors say plainly that the three parameters governing performance may behave differently in a real structure.³

That is a limitation most of the technologies this journal examines have and few of them state.

19.2 And the species list

Extended experiments are needed on named beetle species and termites, whose expanding European range under climate change increases their relevance.³

19.3 So the physics is settled and the protocol is not

Attenuation coefficients and frequency bands are known. What sensor spacing, what listening duration and what detection threshold to use in an actual building are not.³

That is our summary of the gap.

20. The two-stage design

A commercial proposal that gets the architecture right.

A patent describes performing a thermal scan of the structure to identify potential infestation sites, then positioning acoustic sensors at the potential infestation sites to detect vibration signals in a stated band, comparing them with control signals.⁶

20.1 Thermal and acoustic detect different things

A thermal scan sees a temperature anomaly, which may be a void, damp, a pipe, a draught or an insect gallery. It has coverage and no specificity, which is exactly the property a screen needs.

20.2 The stated rationale

That the combination couples a quicker but low-specificity screening technique for speed with a high-specificity, slower technique for accuracy.⁶

21. Which is the right architecture

And worth naming, because it is a general principle.

A fast test with poor specificity narrows the search. A slow test with good specificity resolves what the first one flagged. Neither is adequate alone and the combination is better than either at the same total cost.

21.1 And it is the structure of good diagnosis generally

A screening test tolerates false positives because the follow-up catches them, and a confirmatory test tolerates slowness because it runs on few cases. Reversing the order, or using either alone, wastes whichever property the other supplied.

Our brown recluse article described an exclusion instrument built because confirmation was unavailable. Here confirmation is available and expensive, so it is rationed by a cheap screen, which is the other half of the same design problem.

21.2 It solves the coverage problem in §9

An acoustic sensor covering one member at a time cannot survey a building. A thermal scan can survey a building and cannot confirm anything. Together the coverage comes from the first and the specificity from the second.⁶

21.3 And the patent is a patent

Which is an argument for its own invention and not evidence that the combination performs as claimed. We found no independent evaluation and flag it accordingly.⁶

22. The commercial instrument

What is actually sold, as described in the patent literature.

Apparatus for detecting insect sounds over the full range of frequencies from about 100 Hz to about 15 kHz, with a sensor specially optimized for detecting termite sounds in the 100 Hz to 15 kHz frequency range while excluding the effect of ambient noise, generally including a highly sensitive electronic microphone coupled with a mechanical sound amplification means, such as a stethoscope.⁶

22.1 The stethoscope is not a joke

A mechanical amplifier coupling a small contact area to a diaphragm is exactly the right instrument for picking vibration out of a solid surface, and it has been the right instrument for two centuries.⁶

22.2 Note the band

Entirely below the quiet region. The device operates where the range is good and the background noise is at its worst, and compensates with pattern matching rather than with physics.⁶²

22.3 Which is a defensible engineering choice

Given §8. A device restricted to eight centimetres across the grain would be almost unusable in a building, so accepting noise in exchange for range is the trade an inspection tool has to make.

That reading is ours.

23. What a contractor should expect

Four things.

It hears along members, not through them. Section 8, and it determines how many readings a structure needs.

A negative means nothing was active in range during the measurement. Which is worth saying to a client in those words.

Noise defeats it. A building with running plumbing, traffic or occupants is a harder measurement, and the honest response is to say so rather than to report a zero.³

And it is a confirmation tool rather than a survey tool. Section 21.

23.1 Which sets what it is worth paying for

Not a whole-house survey, which §9.1 shows scales badly, but a specific question about a specific member that visual inspection has raised and cannot settle.

That is a narrow application and a genuine one, and it is the only use we would defend on the evidence available.

24. Our own position

The disclosure, corrected.

We own acoustic detection equipment, so this is an article about a tool we use, and §25.2 limits the use we would defend to a narrow one, which constrains what our own surveys can be relied on to show. An earlier version of this section stated that we did not own such equipment, which was wrong. We wrote it because the failure mode in §16 is the standard we think inspection reporting should meet, and because we wanted to understand why a technology this old has not displaced visual inspection.

24.1 The answer to that second question

Appears to be §9. A method that surveys one structural member at a time cannot replace a method that surveys a room, whatever its accuracy on the member it is on.

25. The Manitoba position

Two notes.

The wood-destroying insect case is limited here. Our carpenter ant articles describe the principal structural wood pest in this province, and we found no acoustic detection literature specific to that genus.

The grain case is not limited at all. Section 10 describes detection and quantification of stored product insects at low numbers, in a province that stores a great deal of grain, and our grain storage article describes hot spots that are localised and hard to find.⁴

25.1 And the bin has a second advantage

It is a controlled acoustic environment, unoccupied for long periods, with no plumbing and no footfall. The ambient noise problem that defeats the low band in an occupied building is largely absent in a grain bin at night.

25.2 Which is the application we would want to see tested

A hot spot is a concentration of metabolically active insects in a defined location, which is the best possible target for a method that detects activity. We found no study pairing the two and it is an obvious pairing.

26. Limitations and open questions

Several figures reach us through a review rather than the original studies. The 2.2 metre figure, the eight centimetre figure and the attenuation coefficient all appear in a review citing earlier work.¹

Some are old. The underlying measurements date from the 1990s and early 2000s, and sensor technology has moved.

Two sources are patents. Arguing for their own inventions, flagged accordingly.⁶

Two distance figures were truncated in our sources. The stored product detection distances are reported with their units cut off in the text available to us, and we have not reproduced them.⁴

We found no field validation in occupied buildings. No sensitivity or specificity figures against a known ground truth, which is the measurement that would settle whether the method is useful.

And no cost information at all. Neither instrument prices nor the time a survey takes, which determine whether any of this is practical.

Sections 1.1, 2.1, 3.1, 5.1, 6.2, 8.2, 9, 11.1, 12.1, 13.2, 14, 17, 18, 19.2, 21, 22.2, 23 and 25.1 are our reasoning. The framing of the trade-off as unresolvable, the reading of the anisotropy as a geometry constraint, the parsing of what an acoustic negative asserts, the identification of the declared non-measurement as a third result category, and the proposed grain hot spot application are ours rather than sourced positions.

27. Conclusion

Acoustic detection is constrained by a trade-off that cannot be engineered away. Background noise is negligible above twenty kilohertz and acoustic emission sensors above forty exploit that, but the attenuation coefficient rises a thousandfold between five hundred hertz and a hundred and twenty, and rises faster still in soil and grain, which makes both highly effective insulators against exactly the frequencies that would be easiest to interpret.¹⁵² Wood is the exception, passing ultrasound from termites over two point two metres, and then not: the same wood carries the same signal about eight centimetres across the grain.¹

That anisotropy is the reason a technology described as having a century of history has not replaced looking at things. A sensor hears along the member it is on and hears almost nothing in the one beside it, so coverage scales with the number of structural members rather than with area, and the sensor spacing recommendations that would make it a protocol are identified in a 2026 paper as not yet existing.³ The commercial answer is a two-stage design, a fast low-specificity thermal screen followed by a slow high-specificity acoustic confirmation, which is the correct architecture and which we found no independent evaluation of.⁶

What we will carry forward from this subject is a sentence. Where noise cannot be distinguished from larval activity, a measurement is not possible.³ Not a zero, not an inconclusive finding, not a low-confidence absence: no measurement. This journal has spent a great deal of time on the uninterpretable zero and the uninterpretable positive, and here is a third category neither of us had a name for, which is an instrument declining to produce a number it cannot stand behind. Almost nothing in structural pest inspection works that way, and everything in it should.

References

1. Perspective and promise: a century of insect acoustic detection and monitoring. Review article, national agricultural research service copy. Used for the statement that the attenuation coefficient increases a thousandfold between 500 Hz and 120 kHz in air, and at even greater rates in soil and grain, making both substrates highly effective insulators against high-frequency sound; for the statement that wood has a low attenuation coefficient so that ultrasound, meaning frequencies above 20 kHz, could be detected from termites over active spaces of up to 2.2 metres in wood; and for the finding, attributed to cited earlier work, that low-frequency sounds from termites and other insects can be detected at metre distances from the sensor location along the wood grain but only about 8 centimetres away across the wood grain. https://www.ars.usda.gov/ARSUserFiles/80100500/AsianLonghornedBeetleResearch/1010_2011%20Mankin%20et%20al.%20Perspective%20and%20promise%20a%20century%20of%20insect%20acoustic%20detection%20and%20monitoring.pdf
2. Research repository record for the above review, reproducing further sections of the text. Repository listing rather than the primary publication, flagged accordingly. Used for the statement that the efficacy of acoustic devices in detecting cryptic insects, estimating population density and mapping distributions depends on many factors including the sensor type and frequency range, the substrate structure, the interface between sensor and substrate, the assessment duration, the size and behaviour of the insect, and the distance between insects and sensors; for the note that considerable success has been achieved in detecting grain and wood insect pests; for the observation that microphones are useful sensors for airborne signals but vibration sensors interface better with signals produced in solid substrates such as soil, grain or fibrous plant structures; for the statement that ultrasonic sensors are particularly effective for detecting wood-boring pests because background noise is negligible above 20 kHz and ultrasonic signals attenuate much less rapidly in wood than in air, grain or soil; and for the account that problems in distinguishing sounds produced by target species from other sounds have hindered usage of acoustic devices but that new devices and signal processing methods have greatly increased reliability, with one new method considering spectral and temporal pattern features that prominently appear in insect sounds but not in background noise and vice versa. https://www.researchgate.net/publication/233566409_Perspective_and_Promise_a_Century_of_Insect_Acoustic_Detection_and_Monitoring
3. Detection of active wood-boring insect larvae using acoustic emission measurements: principles, experimental validation, and practical applications. Open-access journal article in a sensors title. Used for the limitation statement that signal attenuation characteristics, effective detection range and the relative impact of background noise may all differ substantially in field conditions, and that a more systematic experimental investigation of range and attenuation in structural elements of varying dimensions remains a priority for future research; for the disclosure that in the distance experiments the compound geometry of the test assembly, being a separate specimen clamped to a measurement piece rather than a larva embedded within the measurement piece itself, constitutes an additional acoustic interface and poses suboptimal signal transmission conditions, while noting a similar situation may arise in real infestations where a measurement is taken on a piece of wood joined to an infested piece; for the statement that if noise cannot be distinguished from larval activity, a measurement is not possible; for the identification as a research priority of systematic distance and attenuation experiments using monolithic timber of varying dimensions and species with larvae of known mass, which would enable evidence-based sensor spacing recommendations for structural applications; and for the identification of the need for extended-duration, large-sample experiments on a named beetle species and other practically relevant species including three further named beetles and termites of a named genus, whose expanding European range under climate change increases their relevance, in order to establish species-specific measurement duration recommendations and detection thresholds. https://pmc.ncbi.nlm.nih.gov/articles/PMC13469404/
4. Acoustic detection of insects in stored products in the presence of strong ambient noise. Open-access journal article in a sensors title. Used for the statement that acoustic sensors and algorithms leverage the sounds and vibrations generated by insects as they move, communicate, masticate and procreate within a material; for the terminological note that although many studies refer to these techniques as vibro-acoustic methods, the paper adopts the term acoustic to denote elastic waves within the tested materials; for the report that a system originally designed for termite detection was used to track the movement of one to twenty adults of four named stored product and fabric pest species in corn meal and flour mix, and was also able to quantify the level of infestation; for the report that a similar system successfully detected two named beetle species in wheat grains; for the caveat that this method can only detect the movement of the insects and is highly susceptible to vibrational interference; and for the note that ultrasonic emitters and receivers have been studied to detect pests and insect damage in wood by measuring changes in time of flight. https://pmc.ncbi.nlm.nih.gov/articles/PMC12987122/
5. Acoustic detection of termite infestations in urban trees. Research repository record reproducing sections of a journal article. Repository listing rather than the primary publication, flagged accordingly. Used for the statement that acoustic emission sensors are successful because they are nondestructive and operate at high frequencies above 40 kHz where there is negligible background noise to interfere with detection and interpretation of insect sounds; for the statement that the range of acoustic detection is much greater at frequencies below 10 kHz and that low-frequency accelerometers have been used to detect insect larvae over one to two metres in grain; and for the note that acoustic emission systems have been used primarily for detection of termites in wood but that there is also an obvious need to detect termites in soil around buildings. https://www.researchgate.net/publication/11061339_Acoustic_Detection_of_Termite_Infestations_in_Urban_Trees
6. Termite acoustic detection, patent specification. Commercial advocacy for the claimed invention, flagged accordingly and not independent evidence of performance. Used for the description of apparatus for detecting insect sounds over the full range of frequencies from about 100 Hz to about 15 kHz, capable of comparing detected sounds to a library of previously recorded known termite sounds using pattern matching or recognition to find matches, and of generating and transmitting detection messages to an operator; for the description of an acoustic sensor specially optimised for detecting termite sounds in that frequency range while excluding the effect of ambient noise, generally including a highly sensitive electronic microphone coupled with a mechanical sound amplification means such as a stethoscope; for the claimed method of performing a thermal scan of the structure to identify potential infestation sites, positioning acoustic sensors at those sites to detect vibration signals in the stated band, transmitting them to a computing device for comparison with control signals, and detecting infestation where the detected signals are substantially similar; and for the stated rationale that the combination of infrared and acoustic inspection couples a quicker but low-specificity screening technique for speed with a high-specificity, slower technique for accuracy. https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/7385483
7. Publisher record for the wood-boring larvae acoustic emission validation paper in a sensors title. Used for the paper's title and framing, namely the detection of active wood-boring insect larvae using acoustic emission measurements, covering principles, experimental validation and practical applications; and for the operational consideration of signal attenuation identified in the discussion. https://doi.org/10.3390/s26154672