Field of the invention

[0001] The present invention is generally related to the field of techniques for calibrating headphones.

Background of the invention

[0002] Acoustic calibration of headphones (i.e. a device comprising a pair of earphones joined by a band placed over the head) has been traditionally performed by specialized institutes because of the hardware requirements and required level of expertise. Acoustic calibration requires very specialized and expensive hardware, as well as the right "environmental" conditions, such as a sound-proof room (where it is very silent and quiet), etc. All this makes it a task not so easy to perform properly. To carry out such calibration correctly, a considerable investment of money, time and effort is required.

[0003] Acoustic calibration of headphones in professional audiological applications is even more complicated and specialized. Standard devices to measure the output of headphones are based on a standardized acoustic coupler between the sound device (headphone, headset, in-the-ear set etc.) and the microphone. Typically, Bruel & Kjaer systems are used for that purpose. However, even the sound pressure levels (SPLs) measured with such a coupler, do not provide the final information about the intensity levels at the eardrum, as the acoustic coupler and the ear have about the same volume (6 cc) but still different frequency-dependent characteristics. Therefore, each headphone type has to be calibrated subjectively with a large group of listeners. All this is not done for consumer headphones.

[0004] Standards for calibrating headphones and insert phones (headsets) for application in audiometrical equipment have been published in two ISO standards:

- ISO 389-1 (1998). Acoustics - Reference zero for the calibration of audiometric equipment. Part 1 - Reference equivalent threshold sound pressure levels for pure tones and supra-aural earphones

- ISO 389-2 (1994). Acoustics - Reference zero for the calibration of audiometric equipment. Part 2 - Reference equivalent threshold sound pressure levels for pure tones and insert earphones

[0005] Basically, the calibration procedures include two major elements : measuring the acoustic output in a reproducible way (objective measure: dB SPL) and relating these acoustically measured values to the mean hearing threshold of a group of human listeners (subjective measure: converting to dB HL, whereby HL stands for hearing loss). [0006] However, even when all requirements and conditions are met, differences between supra-aural headphones, earphones and other types of transducers still remain, due to the differences in ear canal shapes and how those transducers fit in them, compared to an artificial ear but also to other human ears.

Reference is made to Richard Seewald on Audiology Online (http://www.audiologyonline.com/ask- the-experts/advantages-and-disadvantages-insert-earphones-48 8) and Mitzi Walkup at al. ("Clinical Observations on Insert Earphones vs. Headphones", VA Tennessee Valley Healthcare System), see http://www.myavaa.org/documents/conferences/2008/Walkup_lnse rts_v_Headphone.pdf.

[0007] In US2013/003981 a calibration in real time of headphones to improve the accuracy of recorded audio content is presented. Microphones used to record the audio content (including orientation and recording characteristics) are characterized to indicate a first audio coloration. Audio playback devices used to process the audio content are characterized to indicate a second audio coloration. Headphones corresponding to the user headphones are characterized to indicate a third audio coloration. An equalization signal is computed based on the audio colorations, and is applied to calibrate the user headphones during playback of the audio content. A database of the characterizations is maintained so that calibration of different models of headphones using different playback devices can be accomplished for audio content recorded using different models of microphones.

[0008] US2011/002471 is concerned with the calibration and tuning of audio transducers (headphones, speakers, microphones etc). To compensate for inaccurate reproduction of an applied signal the digital signal processor has to transform the input signal taking into account the transducer characteristics. Because a transducer has its own characteristics that need to be compensated for separately, a profile is supplied to the DSP either by a database lookup based on an identification made by the user or transducer itself or by profile information stored on the transducer. Once the characteristics of a transducer are known, many additional DSP algorithms can be applied in order to improve the audio performance.

[0009] Application US2013/345594 relates to a digital headset system for use in audiometric testing. The headset system includes a stored calibration reference relating the exact frequency and volume response of each speaker to analog input signals. A microprocessor accesses the calibration reference to determine the required analog signal needed to produce the desired sound. An on-board digital to analog converter generates the required analog signal and transmits it to the speaker. The headset is used with software based audiometric test methods that allow the generation of an electronic user hearing profile. [0010] Hence, there is a need for a cheap and simple method for calibrating a test headphone set.

Summary of the invention

[0011] It is an object of embodiments of the present invention to provide for a method of directly calibrating headphones, which is both inexpensive and fast, while providing high accuracy.

[0012] The above objective is accomplished by the solution according to the present invention.

[0013] In a first aspect the invention relates to a method for calibrating a set of headphones comprising

- performing for each person of a test group of at least 10 persons the following steps :

performing a first measurement over a given frequency range of the intensity of a reference acoustic signal provided by a sound generating device to the person via a set of reference headphones with known characteristics,

performing a second measurement over the given frequency range of the amplitude of a digital signal produced in the sound generating device, said digital signal being converted to a corresponding acoustic signal via a set of headphones to be calibrated, said acoustic signal having an intensity equal to the intensity of the reference acoustic signal,

taking the difference between first measurement values from the first measurement and second measurement values from the second measurement as calibration values for the set of headphones to be calibrated,

- performing an averaging of the calibration values to obtain resulting calibration values over the given frequency range for the set of headphones to be calibrated. [0014] The proposed solution indeed allows calibrating the set of headphones under test.

First, the intensity of a set of acoustic signals covering the frequency range and serving as reference is measured while the test person wears the reference headphones with known behaviour. Next, measurements are performed using the set of headphones to be calibrated. When the test person observes the same acoustic intensity on the test headphone set as the reference acoustic signal, the corresponding amplitude of the digital signal provided by the sound generating device is measured. From the difference between the measured intensity values related to the reference acoustic signal resulting from the first measurement (obtained with the reference set of headphones) and the 'loudness' values produced by the sound generating device from the second measurement (obtained with the set of headphones under test) a set of calibration values is derived for the person taking the test. By averaging these values over the test group a resulting set of calibration values is obtained. Obviously, the larger the test group, the more reliable calibration values result. In the proposed procedure the ears of the test persons are used as a measurement device. More precisely, the ears have to indicate that a same intensity level is experienced in the two measurements.

[0015] In a preferred embodiment the first and the second measurement is an audiogram measurement. As well known in the art, in an audiogram the acoustic intensity is expressed as a minimum hearing threshold. Note however that instead also the most comfortable level could be considered for comparison or any other quantity that can be measured on both headphones in a similar way.

[0016] In a preferred embodiment the first measurement is performed by presenting to the test person a series of stimuli, said stimuli being either a single pure tone, a composition of a plurality of short tones or a silence.

[0017] Preferably, the sound generating device is a smartphone or a music player.

[0018] In one embodiment the set of headphones has a microphone attached, so forming a headset. In another embodiment the sound generating device comprises a microphone.

[0019] Advantageously the microphone is used to monitor the environmental noise and to determine an indication of the noise level. This information may later be exploited in an algorithm to reduce the noise level.

[0020] In one embodiment the method comprises a step of determining at least one output characteristic of the sound generating device. In case the entire system, i.e. sound producing device and headphone set is measured, this additional information can then be used to characterize the effect of the sound generating device and so to derive characteristics of the headphone set alone.

[0021] In a preferred embodiment the sound generating device is provided with a digital connection for headphones.

[0022] In another aspect the invention relates to a program, executable on a programmable device containing instructions, which when executed, perform the method as previously described.

[0023] For purposes of summarizing the invention and the advantages achieved over the prior art, certain objects and advantages of the invention have been described herein above. Of course, it is to be understood that not necessarily all such objects or advantages may be achieved in accordance with any particular embodiment of the invention. Thus, for example, those skilled in the art will recognize that the invention may be embodied or carried out in a manner that achieves or optimizes one advantage or group of advantages as taught herein without necessarily achieving other objects or advantages as may be taught or suggested herein. [0024] The above and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter.

Brief description of the drawings

[0025] The invention will now be described further, by way of example, with reference to the accompanying drawings, wherein like reference numerals refer to like elements in the various figures.

[0026] Fig.l illustrates a typical sound production chain.

[0027] Fig.2 illustrates a comparison of the reference set-up and the test set-up.

[0028] Fig.3 illustrates the average distribution of the hearing loss of a population of humans.

Detailed description of illustrative embodiments

[0029] The present invention will be described with respect to particular embodiments and with reference to certain drawings but the invention is not limited thereto but only by the claims.

[0030] Furthermore, the terms first, second and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a sequence, either temporally, spatially, in ranking or in any other manner. It is to be understood that the terms so used are interchangeable under appropriate circumstances and that the embodiments of the invention described herein are capable of operation in other sequences than described or illustrated herein.

[0031] It is to be noticed that the term "comprising", used in the claims, should not be interpreted as being restricted to the means listed thereafter; it does not exclude other elements or steps. It is thus to be interpreted as specifying the presence of the stated features, integers, steps or components as referred to, but does not preclude the presence or addition of one or more other features, integers, steps or components, or groups thereof. Thus, the scope of the expression "a device comprising means A and B" should not be limited to devices consisting only of components A and B. It means that with respect to the present invention, the only relevant components of the device are A and B.

[0032] Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment, but may. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner, as would be apparent to one of ordinary skill in the art from this disclosure, in one or more embodiments. [0033] Similarly it should be appreciated that in the description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. This method of disclosure, however, is not to be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.

[0034] Furthermore, while some embodiments described herein include some but not other features included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention, and form different embodiments, as would be understood by those in the art. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0035] It should be noted that the use of particular terminology when describing certain features or aspects of the invention should not be taken to imply that the terminology is being redefined herein to be restricted to include any specific characteristics of the features or aspects of the invention with which that terminology is associated.

[0036] In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the invention may be practiced without these specific details. In other instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

[0037] The invention aims to present a method of directly calibrating a set of headphones in a same unit as a reference, pre-calibrated set of headphones. The reference set has known characteristics, so that an output signal with a known acoustic level can be produced. In a preferred embodiment an audiological set of headphones calibrated in dB HL is used. The main idea is to use human ears as measurement devices and compare the set of headphones to be calibrated to the reference headphones, without requiring the use of a coupler or other advanced hardware. Although human ears are very imprecise as measurement devices, they are readily available. Given enough measurements, it is possible to cancel out the measurement errors and deduce the "true" value for the calibration of headphones.

[0038] The proposed method can also be applied to a headset, i.e. a set of headphones with a microphone attached. [0039] Calibration values are frequency-dependent curves describing the relation between the amplitude of the digital signal (as measured in dBFS, dB full scale, relative to the maximum digital value that the device can output) and the actual loudness values as measured at the user's ear level. Hence, the calibration values obtained in the scheme of Fig.l indicate the relationship between the digital signal amplitude as produced by the device (e.g. a mobile phone or a music player) and the loudness of the audio signal as perceived at the user's ear level.

[0040] The method of the invention is primarily intended to calibrate a set of headphones in dB HL (dB hearing loss, used to measure a person's hearing impairment), but can also be used to derive absolute calibration levels in dB SPL or any other unit in which the reference headphone set has been characterized.

[0041] Having calibration values for a system comprising a device producing digital audio and a headphone set coupled to that device allows the device to produce a digital signal which precisely results in a desired output level at the ear. This is required e.g. in the case of hearing loss compensation, as a certain amount of amplification must be given, not more and not less.

[0042] Another advantage of the proposed method is that it can be run multiple times with different sets of people repeating the procedure in different places in the world. By continuously accumulating this data it is possible to better absorb statistical anomalies in order to further fine-tune the resulting measurements. This can create a positive feedback loop, whereby a system as described in patent US9,055,377 is able to provide better hearing loss compensation due to better knowledge of the exact sound output it is producing, which in turn should attract more users that will also be contributing to the calibration of the headphones they use.

[0043] First the procedure according to the invention is explained for a controlled environment. In the proposed approach use is made of a headphone set with known characteristics, preferably an audiological set with very precise calibration. This set of headphones is hereafter referred to as the reference headset. The set of headphones to be calibrated is hereafter called the test set. Also needed is a group of a certain number of people, e.g. at least 10, to ensure sufficient accuracy in the measured values. Obviously, more test persons will add more accuracy.

[0044] The main idea is to compare both headphone output responses using humans as reference points between the two sets of headphones. This is done by performing a task which is measurable by humans on the two headphone sets, and deducing calibration values for the test set by comparing the latter with the reference set. In a preferred embodiment an audiogram measurement of the subjects is used (typically yielding a minimal hearing threshold as a function of frequency), but in principle any task that can be measurable and that measures loudness is adequate. For example, one can measure the most comfortable level instead of the minimum hearing threshold. [0045] Using human subjects to perform these tasks, incurs measurement errors much higher than what can be expected of a standard calibration. However, the use of multiple subjects allows averaging the calibration error over all the subjects, which should yield the true value being measured as measurement errors have a null arithmetic mean, and allows estimating the standard deviation of the measurement values and, hence, also inferring the measurement precision.

[0046] More in detail, the steps to be performed in a preferred embodiment of the invented method are :

- measuring the audiogram of a subject on the reference set of headphones and record these values as dB HL,

- measuring the audiogram of the same subject using the test set. To the extent possible, the conditions in which the procedure is run, are identical to those for running the previous step (e.g.: same sound booth to ensure same noise conditions, etc.). The values as produced by the sound generating device (e.g. smartphone, etc.) are hereby recorded in dB FS,

- recording the difference between the values measured in the two previous measurements (i.e. the audiogram in dB HL of the reference headphones and the measurement of the amplitude levels in dB

FS) as a single measurement of the calibration of the test set of headphones,

- repeating the preceding steps over all the subjects and averaging the found calibration values.

[0047] Fig.2 illustrates a comparison of the calibrated reference set-up and the test set-up. In Fig.2 S _d represents the digital signal (measured in dB FS), S _e the electrical signal (in Volts) and S _a the audio signal (in dB HL). The "ref" or "test" qualifier denotes whether the value is measured on the reference headphone set or on the test headphone set. If the same loudness is measured on the reference set-up and on the test set-up, this means S _{a re} f = S _{a test} . The procedure relies on the fact that indeed the same quantity (e.g.: the hearing threshold of a human for a given ear) is observed on the test person using the reference headphone and using the test headphone. Next, a difference in the intensity level of the acoustic signal produced by the reference headphone and the amplitude level of the digital signal applied to the test headphone is calculated.

[0048] The quantity S _{d ;test} represents the digital level at which the sound is presented and is known by measurement. S _3:test = S _a,ref is known too, as it is the value measured on the reference headset, which yields dB HL values. Thus, the relation

^J a,test ( ) = Sd.test tf) + Calibration(f)

can be rewritten as

Sa.ref if) = ¾,test ( ) + Calibration(f)

In other words: Calibration(f) = S _{a ref} (f) - S _ditest (f)

All these values are frequency-dependent and should thus be measured for each frequency independently. The frequency resolution of the final calibration is the same as the one used during the fitting process, so a pertinent choice of fitting process can result in better measurements.

[0049] In a preferred embodiment the repeated DuoTone procedure (as described in patent EP2572640 Bl) can be used, as it is fast and flexible in the choice of frequency points and can either be repeated for a lot of different frequencies, which yields a higher frequency resolution, or can be repeated multiple times for the same frequency points in order to achieve a higher accuracy of the final resulting values. With the DuoTone procedure a user's audiogram can validly be measured.

[0050] DuoTone is a fully-automated procedure allowing a user to perform pure-tone audiometry on himself/herself without supervision by a professional. The procedure is implemented as a series of stimuli/answers, where in each round the device presents a sound stimulus to the user and waits for him to answer via a touch screen whether he has heard the stimulus, and if yes, which one (out of a predetermined list). A typical DuoTone procedure uses two types of stimulus, namely a lower-frequency stimulus consisting of a single pure tone and a higher-frequency stimulus composed of 3 short tones. At each stimulus presentation, the user hears either one of the stimuli or nothing (silence), and has to answer using the touch screen interface by tapping on a button corresponding to the stimulus that was heard. Once the user has prompted a correct answer to a test stimulus with a set frequency, the next test stimulus of that frequency is presented with a lower intensity, usually with a step size of 10 dB. Once the user has presented a wrong answer, the next stimulus presented to the user with that frequency is higher again, usually twice the normal step size. When the user has presented a wrong answer, this is considered as a lower hearing threshold, as the previous stimulus for that frequency was heard, but the current one not anymore. This procedure is usually repeated until three lower hearing thresholds are determined for each frequency utilized. In that way, the lowest intensity just to be heard by the user can be reliably determined. Once the system has found the lowest intensity for the same frequency for the third time, the method is ended and the result (the threshold) is calculated from the three measured intensity levels, usually calculating the arithmetic mean plus half the step size. In this way a user's audiogram can validly be measured.

[0051] The procedure as described above works well in a controlled environment such as an audiological clinic. However, the proposed method is usable in all types of environments where conditions are not ideal, and, as such, has to deal with some issues described hereafter.

[0052] Measuring someone's audiogram using calibrated material is quite a common and recognized audiological procedure, which is also described in international standards, as are the precision requirements of the audiological headphone sets and audiometers. This ensures a quite precise and reproducible audiogram measurement, regardless of the hardware on which it is executed.

[0053] In this case the step of measuring the acoustic signal intensity via a known reference set of headphones in the aforementioned procedure can be performed previously to and independently of the procedure. Thus, the total time of the procedure can be reduced if all persons participating come to the procedure with their audiogram already measured. In order to still achieve as high a precision as possible, the time between the audiogram measurement on an audiological headset and on the test headset should be as small as possible to ensure similar conditions. It is known for instance that being ill (flu, ear infections, etc.) can influence your hearing and thus would invalidate the result of the experiment. In fact, based on this observation, it is possible to use such a repeated procedure (e.g., weekly) as screening for illnesses such as ear infections in children.

[0054] The main issue one has to deal with when running the procedure outside a controlled environment such as an audiological clinic, is the level of ambient noise in the environment. As the hearing threshold is measured, it is of paramount importance than the noise level is well below the volume of the tones used to measure the hearing threshold. In case the noise level is constant and cannot be reduced (e.g.: air conditioning system, city noise such as proximity to a busy street ...), the measured values are higher than what they would be in a controlled environment, and cannot be used reliably. In case the noise level is sporadic (e.g. typical office noise, such as feet shuffling, birds chirping outside, people talking at a distance, ...), the device used for performing the measurement (e.g. a smartphone) is also typically equipped with a microphone, which can be used to measure the noise level during the tone presentation. If the noise level is too high at this moment, it can be decided to discard the measured value and repeat the tone to get a more accurate reading. It has been shown that smartphones can actually provide a very good noise level estimate, nearly comparable to professional calibrated audiometers.

[0055] When running the procedure as described above, the measured values actually are calibration values for the entire system comprising the sound producing device (e.g. a smartphone) and the test set. It is not generally possible to deduce the calibration values for the headphone set only, as the same set plugged into different devices might have a different output, given that the output impedance of such devices might not be identical on all devices. To mitigate this several options are available. If the selection of devices is relatively small, there might not be such a big increase in the total number of possible configurations when the entire system is always calibrated. This still should be a tractable problem. A better solution, however, is to compare the impedance and power output of different devices with the same set of headphones plugged in. It may be that different devices have the same or similar output characteristics (output impedance and output power) which can either be known from published specifications from the manufacturer or can be measured independently. For example, it can be expected that the iPhone 4, iPhone 5 and iPod Touch 5 ^th generation all use the same or similar components and thus have the same output characteristics, in which case the measurements done on one device are also valid on the other ones and thus the calibration values would only depend on the headphone set. Another alternative is to use a set of headphones with a digital connection to the smartphone. With the advent of new 'intelligent' connectors on mobile phones, a new type of headphone set appears that connects digitally to the mobile device, which implies that the DAC is embedded directly in the headphone set. Thus, in this case, the system calibration is in effect a headphone set calibration. The system calibration does not depend on the device anymore at all, as all the components needed to convert the original digital signal to a sound wave are contained inside the headphone set. Another advantage of those types of headphones is that they can convey more information on this digital link, so they could for instance record the calibration values themselves or carry them as technical specs publically available from the connected mobile device.

[0056] This method, if successfully implemented and applied over the Internet, allows easily gathering a lot of data and producing calibration values for a large number of combinations of device + headphone set. This can be used advantageously to perform a characterization of both the headphone set and the device and possibly infer calibration of a new device + headphone set combination that has not been measured previously. For instance, if there are two devices Dl and D2 and two headphone sets HI and H2 and if Dl+Hl, D1+H2 and D2+H1 are known and one wants to deduce D2+H2, while knowing that Dl+Hl = D2+H1 (i.e., the headphone set HI performs the same on both devices), one can reasonably assume that D2+H2 = D1+H2, so bypassing the need to actually measure this combination. Even when the condition Dl+Hl = D2+H1 does not hold, but there is a constant and reproducible difference over more headsets, i.e. Dl + Hi = D2 + Hi + constant, for various headsets Hi, this constant difference can be considered as the calibration values for the sound generating device itself and can be used to infer calibration values for the entire system (with headphone set included), even when said headphone set is only characterized on a different sound generating device.

[0057] The described procedure can be run in a variety of different settings, all of which can lead to a successful calibration of the headphone set, given that care is taken not to introduce systematic errors (as previously described). A non-exhaustive list of possible variations contains : - the location where the procedure is run : controlled quiet environment (audiological clinic), uncontrolled quiet environment (apartment room at night), uncontrolled noisy environment (in the street, office with air conditioning or open windows, ...)

- the type of sound used for the audiometry : pure tones, band-limited noise signals (white noise, pink noise, etc.), animal noises (more entertaining for kids)

- only using user audiometry or possibility of adding a microphone probe in the ear canal of the user performing the audiometry to have an additional measure of loudness, which can later be correlated with the user's response

- use of different target groups for performing the audiometry : normal hearing subjects vs. hearing impaired subjects, headset calibration using specific age groups (children, young adults, etc.)

These variations may influence the number of individuals required to obtain sufficiently precise results, but should not influence the final results if proper care is taken not to introduce systematic errors. For instance, it is expected that the use of 1/3 octave noise band instead of pure tones should give more precise measurements or at least measurements with a smaller standard deviation, as they are much less influenced by frequency dependent resonance effects caused by individually different resonance effects in the external ear. These effects are expected to be larger when using in-the-ear phones (smaller volumes) as compared to headphones (larger volumes).

[0058] The method as described above uses a previously characterized headphone set as reference point for the calibration of the new headphone set. This can be considered as the most preferred embodiment as it allows quickly obtaining the characterization for the new headphone set using few measurements. The method, however, is able to work with any reference point that can be measured using human ears. In one particular embodiment that can prove very useful, the average minimum hearing threshold is used for a given population of normal hearing persons (http://en.wikipedia.org/wiki/Absolute_threshold_of_hearing, ATH), which is a measured value, reliable with a quite low standard deviation (typically 6-7 dB) and can serve as absolute reference point. This allows many more people to participate in the gathering of data, as this is done subconsciously if the method is adapted in the following way:

- people are asked to take an audiometry test (initiated by them because they are using an audiometry app, this is not a clinically controlled environment)

- those people also introduce their headphone set model in the audiometry app (or the phone detects it automatically for new digital headsets which carry their own identification)

- the data is uploaded to a server.

At this point, the users have only performed a normal audiometry test. However, if at the server side a histogram of measured values for a given set of headphones is plotted, it can be expected to contain two Gaussian curves, one quite narrow for normal hearing persons and another, wider curve corresponding to hearing-impaired people (see Fig.3). As these measured curves are in dB FS relative to the device/headphone set combination, and the (known) average levels are in dB SPL (or 0 dB HL, by definition), the calibration of the system used can be deduced by comparing the two values. The fact that the ATH is a less precise measurement than a calibrated headphone set's reference levels is compensated by the fact that it is much easier to gather a lot more data in order to get a precise average, where statistical anomalies are easier to identify and isolate.

Note that in Fig.3 the two separate Gaussian curves are a simplification of the reality that has been made for illustrative purposes for the case of this description. In reality, there is no gap between normal hearing and hearing-impaired people. However, it doesn't take away the fact that the left part of the distribution has a steep slope which still allows accurately estimating the ATH.

[0059] As more and more data comes over time, the calibration values for a given set of headphones can also be continuously refined. This makes it easy to calibrate a new headphone set as soon as users start using them for doing audiometric tests on their devices. In effect, this is the preferred method if the user base of the audiometry app is big enough, as it requires nearly no setup (compared to running a clinical experiment) and should yield the same results.

[0060] While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive. The foregoing description details certain embodiments of the invention. It will be appreciated, however, that no matter how detailed the foregoing appears in text, the invention may be practiced in many ways. The invention is not limited to the disclosed embodiments.

[0061] Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single processor or other unit may fulfil the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems. Any reference signs in the claims should not be construed as limiting the scope.