AUDIO PLAYBACK OF AN ELECTROMAGNETIC METAL DETECTION SIGNAL

Description

[Technical field]

[0001] The present invention relates generally to the field of metal detection using portable electromagnetic detectors, and more particularly to the audio reproduction of an electromagnetic metal detection signal.

[0002] Such metal detectors are used primarily for leisure purposes, for example for searching for and discriminating buried metal objects such as coins, jewelry, treasures or gold nuggets, but also by professionals, for example for mine clearance, ballistics research, or the search for pipes in the construction industry or polluting metal particles in the food sector. These application examples are not exhaustive. Modern handheld metal detectors allow the user not only to locate but also to identify a target buried in the ground. [Technological background]

[0003] Usually, metal detectors include an electromagnetic detection head, or more simply detection head or detection disc, which constitutes the active part of the detector, and which is mounted at the end of a handling rod. This detection head takes the form of a disc because it includes one or more inductive coils which make it possible to generate an incident electromagnetic field, and to receive a resulting magnetic field modified by the immediate environment.

[0004] The detection head may also incorporate control electronics, adapted to produce an alternating electromagnetic signal at the origin of the incident electromagnetic field, and to process a received electromagnetic detection signal corresponding to the resulting magnetic field. This processing makes it possible to detect and discriminate possible metallic objects exposed to the incident electromagnetic field. Alternatively, the detection head only incorporates part of the electronics for processing the received electromagnetic detection signal, for example a preamplifier, the rest of this electronics being remote in a control box to which the detection head is connected.

[0005] In portable electromagnetic detectors operating in continuous wave mode, also called continuous wave detectors (or CW, from the English "Continuous Wave" ) ,an alternating magnetic field is emitted "continuously" and the detection is based on the amplitude and phase variations between the frequency components of the emitted signal and those of the received signal. More particularly, CW detectors use the measurement of amplitude and phase of the different components of the alternating electromagnetic signal emitted by at least one transmitting coil powered by an electric voltage, and the components of the electromagnetic signal received by one or more receiving coils arranged close to the transmitting coil. The frequency of the signal emitted by the transmitting coil is generally less than 100 kHz. In certain embodiments, the transmitting coil and the receiving coil(s) are only one and the same coil.

[0006] The implementation of the principle of induction balance (or IB, from the English "induction balance") allows very small changes in the properties of the medium to be measured by induction. This is achieved by arranging the transmitting and receiving coils in such a way that the mutual inductance of the receiving coil(s) and the transmitting coil(s) is as low as possible. This minimizes the electromagnetic field measured in the absence of a target. The coupling of the two coils is then such that the signal received by the receiving coil is substantially zero when no metallic element is placed in the field of the transmitting coil. This optimizes the dynamics of the measurement signal by measuring only the variations in the resulting electromagnetic field.

[0007] The most common configuration is the "double D" head, but other configurations are possible, for example with concentric coils, or by placing the coils at 90° to each other, or with auxiliary coils used to generate a compensation field, or with other geometries.

[0008] On the other hand, there are other types of metal detectors that, instead of working in the "frequency" domain, work in the "time" domain. In this case, it is not the relative amplitudes and phases between the frequency components of the emitted signal and the received signal that are analyzed, but rather the shape of the received signal. This may seem similar because there is a correspondence (through Fourier transforms) between the time domain and the frequency domain, but generally this also covers a different way of generating the emitted magnetic field. In fact, while in the case of a CW detector the signal is emitted continuously, in the case of a pulsed induction (or "PI" mode) detector "Pulsed Impulse") a pulse signal is generated for a few tens of microseconds (µs) in the detection head, then the signal received after this magnetic excitation of the ambient medium is received and analyzed. The same coil can then be used alternately to generate a pulse and to analyze the received field. But two coils can also be used, namely one for transmission and the other for reception in various configurations (for example the aforementioned "double D" configuration, or others). In all cases, the received signal can be seen as the sum of different exponential decays that are representative of the environment in which one operates. Signals associated with soil or salt water have generally rapid exponential decays, with time constants of less than a few µs, while the targets of interest have longer time constants.Obviously, the excitation signal must be renewed periodically (generally with a recurrence of around 1 kHz) and the received signal is "averaged" to be used, but the processing is done in the time domain and not in the frequency domain.

[0009] We can also have detectors that work simultaneously in both modes, namely CW mode and PI mode.

[0010] During a detection session, the user surveys a specific area to be explored by scanning the ground with the portable detector, using the handling rod, which avoids him having to bend down or lean over. More specifically, he stands and moves the detection disc parallel to the ground using the rod, sweeping the detection head from left to right, then from right to left, and so on. The transmission coil produces a magnetic field which, at the level of a possible fixed metal target, varies over time due to the fact that the detection head is thus set in motion. This magnetic field generates eddy currents in the metal target that is sought to be detected.

[0011] These currents in turn generate magnetic fields that are picked up by the receiving coil (which is sometimes the same as the transmitting coil). The received electromagnetic signal, also called the target detection signal or more simply the detection signal or just signal in the following, has an amplitude that varies when the detection head moves above the target, proportionally to the magnetic fields produced by the circulation of eddy currents generated in the target, and received by the detection head. The received electromagnetic signal is then amplified and processed by the control electronics, so as to isolate the signal associated with the target being sought and separate it from parasitic signals (industrial or meteorological interference, variation associated with the ground, the Earth's magnetic field, parasitic targets, etc.).

[0012] The effective detection range of the metal detector depends on the type of target and the technology used by the detector manufacturer. It is limited by the signal-to-noise ratio ( "Signal-to-Noise Ratio" or SNR, in English) which depends on the physical characteristics of the metal detector's control electronics, the signal processing implemented and the various parasitic signals to which the detection coil is subjected. The majority of detectors therefore have a threshold setting below which the detection signal is not taken into account, either because it is in the noise, or because one deliberately wishes to ignore targets that are too small.

[0013] Other information is deduced from the signals analyzed by the detection head. This is so-called "discrimination" information, which enriches the main information relating to the intensity level of the target detection signal, with data concerning the characteristics of the target (for example, its ferrous or non-ferrous nature) and / or its environment (for example, soil characteristics such as conductivity). The relevance of this discrimination information is highly dependent on the signal-to-noise ratio of the detector.

[0014] Signal amplitude and discrimination information can be returned to the user through various sensory channels, including auditory or visual. Audio reproduction is preferred and most commonly used because it allows the user to remain focused on scanning the ground with the detection head.

[0015] The document titled "DEUS V5 USER'S MANUAL" is a user manual that describes several functions and settings available for a metal detector. The document "Outlaw Operator Instruction Manual" is also a user manual for another metal detector.

[0016] Document WO0015001 A2 describes a hearing correction system for people with hearing impairment, in which it is intended to segment an audio signal into three regions to which different gains are applied. [Earlier technique]

[0017] The approaches described in this subsection could be applied, but are not necessarily approaches that have been previously devised or applied. Therefore, unless otherwise indicated herein, the approaches described in this section do not constitute prior art prior to the present application and are not recognized as such by their inclusion in this section.

[0018] The detection signal has an amplitude that varies when the detection head moves over the target in proportion to the magnetic fields produced by the circulation of eddy currents generated in the target and received by the detection head. The analysis of this signal makes it possible to detect targets, by providing the user with indications to assist in the detection and identification of metallic targets. These indications are extracted from the electromagnetic detection signal by signal processing techniques known per se. They are generally presented, i.e. restored to the user in sound form using an electroacoustic transducer such as, for example, an audio headset. For this purpose, an audio signal is generated from the electromagnetic detection signal and is delivered to the user's audio headset, for example, a wireless headset.

[0019] Methods of audio reproduction that can be used involve generating a sound whose amplitude and / or frequency are modulated by the amplitude of the detected signal. For example, the audio signal with the loudest amplitude may indicate the position of the target closest to the detection coil. Such amplitude modulation of the signal may also be used, the frequency of which is further determined by the discrimination information. Also, the amplitude and frequency of the audio signal may be modulated simultaneously according to the detection signal. Generally, the loudest and highest-pitched audio signal indicates the position of the target closest to the detection coil. Amplitude and / or frequency modulation may also be achieved by reserving certain frequency ranges for certain types of targets, and other frequency ranges for other types of targets, respectively.For example, ferrous targets can be signaled by much lower frequencies than non-ferrous targets.

[0020] It is also possible to modulate the amplitude of the audio signal by varying the duty cycle of a rectangular signal driving the electroacoustic transducer. For example, the duty cycle of the latter goes from 0% for a zero detection signal to 50% (i.e., to a square signal) for the strongest detection signal. In addition to the simplicity of the implementation at the hardware level ( "hardware" in English), this has the advantage of coupling the amplitude variation of the audio signal to a spectral variation, that is to say a variation in the frequency content of this audio signal, which allows the user a better sensory analysis than a pure amplitude variation, while reserving the possibility of frequency modulation to indicate discrimination information.

[0021] Various settings can be used to adjust the detection threshold at which a sound is generated (no sound is produced for a detection signal below this threshold) as well as the sound level of the audio signal reproduced, and possibly to modify the different frequencies used for sound generation. In the absence of frequency modulation, the sound can be cut ( "mute" in English) when the target level is below the user-selected threshold, or the detector is detecting particular targets (e.g., ferrite). Where appropriate, the use of frequency modulation to indicate the amplitude of the target detection signal may include an adjustment to adjust the level of the continuous sound that is produced in the absence of a target ( "baseline" in English), or to opt for the absence of audio signal ("mute"), as well as an adjustment of the basic frequency, that is to say the frequency of this continuous sound.

[0022] A problem remains, however, with regard to taking into account the dynamics of the detection signal in the reproduced audio signal. It will be noted that, in the context of the invention and within the scope of the present description, the term "dynamic" applied to the detection signal means the difference between the lowest amplitude and the highest amplitude of said detection signal. Furthermore, when used in reference to the audio signal reproducing the detection signal in sound form to the user, the term "dynamic" means the difference between the lowest sound level and the loudest sound level of said audio signal. In either case, the difference is generally expressed in decibels (dB), which is a relative unit of measurement, on a logarithmic scale, between a given level and a determined reference level.

[0023] Indeed, we observe that the dynamic range of the signal associated with metallic targets is very significant. It can commonly exceed 100 dB depending on whether the target is small and distant, on one side, or large and close, on the other side. However, if the amplitude modulation of a sound signal is used to translate, ie , restore the amplitude variation of the target signal, this produces variations in sound level which exceed the possibilities of the human ear.

[0024] The dynamic range of the human hearing system is theoretically around 120 dB at 1 kHz, but the useful dynamic range is actually much smaller. In the maximum values, the limit is given by the damage that can be caused by excessively high levels. European Directive 2003 / 10 / EC, for example, recommends in this regard not to expose company employees to sound levels of more than 80 dB SPL(A), where the acronym SPL stands for sound pressure level ( "Sound Pressure Lever", in English), for eight hours. Also, the limit allowed by European legislation for portable audio players is 85 dB SPL (A). For the minimum values of the dynamic range, the reproduction of the audio signal is limited by the background noise in which one is immersed. Inside a quiet office space, the background noise is around 40 dB SPL (A), and it can be around 30 dB SPL (A) in a quiet outdoor environment, for example in a forest. But it can be much higher in other outdoor environments, typically around 43 to 48 dB SPL (A) in the jungle, for example. However, to be clearly perceptible to the human ear, a sound signal must have a sound level of around ten decibels above that of the background noise.This gives a useful range of sound levels typically between 30 and 48 dB(A) plus 10 dB(A) on one side for quiet sounds, and 85 dB(A) on the other side for loud sounds, i.e. a dynamic range of 27 dB to 45 dB.

[0025] The excess dynamics between the dynamics of the detection signal (i.e. more than 100 dB) and the desired dynamics of the audio signal for sound reproduction to the user (from 27 dB to 45 dB) is therefore of the order of 50 to 80 dB. With the reproduction of the detection signal by amplitude modulation of a sound signal, the user then has the choice of either lowering the overall sound level, which reduces the amplitude of the audio signal uniformly throughout its variation range, even if it means no longer hearing relatively small and / or distant targets, or adjusting the sound level associated with small targets so as to hear them comfortably, but at the risk of being bothered by too high a sound level in the event of detection of relatively large and / or nearby targets. In extreme cases, excessively high sound levels may cause hearing loss in the user.

[0026] In the scientific article "Auditory issues in Hand-Held Landmine Detectors," by Nancy L. Vause et. al., SPIE Conference "Detection and Remediation Technologies for Mines and Minelike Targets IV," Orlando, Florida, SPIE Vol. 3710 (April 1999), the authors note that many commercially available detectors use acoustic signals that are reproduced at a level considered annoying or that may require the user to temporarily change the detection threshold.

[0027] In the article "Auditory signals for enhanced operator performance with hand-held mine detector," by Gene Ferguson et al., in "Detection and Remediation Technologies for Mines and Minelike Targets V," Proceedings of SPIE Vol. 4038 (2000), the authors (partly in common with those of the above article) note that this issue requires urgent attention because the current audio interface of the HSTAMIDS detection system (put for " Hand-held Stand-off Mine Detection System » in English) is capable of generating acoustic signals at a height of 120 dB (MD) and 114 dB SPL (GPR), for GPR technologies (put for " Modem Ground Penetrating Radar" in English) and MD (put for " Metal Detection» in English), respectively, which exceed the sound levels of non-hazardous presentations for military and civilian applications. In addition, they note that the current MD subsystem of the HSTAMIDS system generates a continuous base signal that can reach 83 dB SPL with the volume control fully open.

[0028] US Patents 4,594,559 and 4,644,290 (1987) issued to David S. BERNZWEIG, which relate to an audio amplifier for a metal detector, disclose an accessory device for use with a metal detector and having means for amplifying the low audio output of the metal detector, which is barely audible, so that it can be easily heard with additional means which are adapted to dampen the louder audio output signals by making these signals weaker. This device allows for limiting the dynamics of the signal in a very frustrating way and introduces signal distortions, which does not make it usable with modern devices.

[0029] Some metal detector manufacturers (e.g. MINELAB ELECTRONICS PTY LTD, on its models GPX4000 ™< , GPX4500 ™< , GPZ7000 ™< , ...) are aware of the problem and add a volume limiting setting ( "Volume Limit" in English) which allows the volume of the sound emitted by the detector to be clipped to a high limit value. The manufacturer warns the user in the user manual of one of these models, in these very explicit terms (free translation): "If the volume limit is set to maximum, all target signals will be heard, and the sound will be proportional to the size and depth of the target. The maximum limit allows you to hear the difference between a small and a large target, but may be uncomfortable for your hearing if a large target itself found near the reel » ; or again: “If the volume limit is set to minimum, most target signals will potentially be very limited. A low volume will be more comfortable for your hearing but will increase the risk of missing targets.” The same manufacturer also warns users of its models with the following warning: “Warning: The detector is capable of producing an extremely high volume if a large or shallow target is located. Protect your ears!” ".

[0030] In other words, this means that the volume limiting setting and the methods based on limiting the amplitude of the audio signal effectively make it possible to reduce the maximum level of the audio signal which is presented to the user, but that the dynamic variations of the detection signal corresponding to the area of the audio dynamics in which this limitation occurs are then completely annihilated.

[0031] The use, for sound reproduction, of a rectangular signal whose duty cycle varies with the level of the target detection signal, provides an advantage. Indeed, if the effective value of a pulsed signal varies linearly with its duty cycle, the same is not true of the perceived sound level ( "loudness" Indeed, if we observe the spectral content of different rectangular signals with different duty cycles, we see that the spectral distribution of energy varies considerably. The lower the duty cycle, the richer (relative to the fundamental) the harmonic content becomes, resulting in a "flat" spectrum in the audible frequencies for duty cycles of less than 1% (for frequencies of 500 Hz or more). However, the sensation of sound level as perceived by a human being varies in a complex way with the spectral composition and the sound level.

[0032] For example, ISO 226:2003 shows the isophonicity curves (i.e., equal perceived level) as a function of frequency and sound pressure for pure tones (sinusoidal signal), which clearly illustrate this variation in perception.

[0033] For complex signals such as those corresponding to a rectangular signal with a variable duty cycle, the isophonicity curve is more difficult to develop, because it depends in a complex way on the energy present in different frequency bands analyzed by the ear. These, called critical bands, have a width of approximately 0.2 x F for frequencies F above 500 Hz, which implies that, for a signal whose fundamental is at 500 Hz (which is a fundamental frequency of audio signal typically used), the first ten harmonics are in separate critical bands and each contribute to the overall sensation of loudness. These first ten harmonics are also in the zone of maximum sensitivity of the ear, and this is all the more so as the acoustic level is low.Thus, as the duty cycle decreases, the relative increase in the amplitude of the first harmonics in the various critical bands analyzed by the ear largely compensates for the decrease in the effective value of the signal as a whole (and this is even more so as the signal decreases), which contributes to a sensation of constant level even as the effective value of the rectangular signal decreases. This results in a kind of natural "dynamic compression" if this type of rectangular signal with modulated duty cycle is used.This "dynamic compression" depends a lot on the sound reproduction devices, in particular the quality of the headphones and / or the loudspeaker, and their ability to reproduce said harmonics, as well as the sensitivity of the user's ear: hearing losses often produce attenuation in frequencies around 4000 Hz in the area in which the variations in the level of the harmonics of the signal are particularly sensitive, which affects the effect of this dynamic compression for users affected by such hearing losses.

[0034] This dynamic compression can reach around thirty dB in the best case for a 500 Hz signal. To further improve this compression, the variation curve of the duty cycle value could be modified so as to favor the dynamic variation felt in a particular area of the dynamics of the detection signal. This dynamic adaptation has limits, however, because it is part of a rectangular signal with a variable duty cycle whose sound, for low levels (and therefore for low duty cycles), can be unpleasant due to its particularly rich harmonic content. [Statement of the invention]

[0035] The object of the invention is to improve the dynamics of the audio signals generated by a metal detector, so that the weakest signals are audible, the strongest signals do not cause hearing discomfort, and the assessment of the sound volume can be carried out gradually for intermediate level signals.

[0036] This goal is achieved, in accordance with embodiments, by modifying the dynamics of the electromagnetic target detection signal, so as to make this dynamic compatible with the actually usable dynamics of the human auditory system.

[0037] To this end, a first aspect of the invention provides a method according to claim 1.

[0038] Segmenting the dynamics of the detection signal into multiple portions and applying respective dynamics change rates to each of these portions means that the overall change in the dynamics of the detection signal can be more complex, resulting in an audio signal with a potentially better feel for the user.

[0039] According to a particular implementation mode, portions of the dynamics of the detection signal are defined by linear segments. The implementation is thus simpler.

[0040] According to a particular implementation mode, at least one local dynamic modification rate is greater than or equal to unity. Thus, and notwithstanding the global compression of the dynamics, local portions of the dynamics can be magnified by extending the dynamics locally.

[0041] The method may further comprise, before the step of reducing the dynamic range of the detection signal, a logarithmic conversion of said detection signal according to a logarithmic scale, and the application of the global dynamic range compression rate to the clipped detection signal or the application of a local dynamic range modification rate to a determined portion of the clipped detection signal, may then comprise the multiplication of said clipped detection signal or of said portion of the clipped detection signal, respectively, by said global dynamic range compression rate or by said local dynamic range modification rate, respectively. The calculations are therefore simpler to carry out by a processor, since it involves a simple multiplication of the values of the amplitude of the detection signal on the logarithmic scale.

[0042] The method may further comprise, after reducing the dynamic range of the detection signal and before generating the audio signal, inversely converting the logarithmic conversion of the received detection signal to bring the dynamic range of the compressed and shifted detection signal back onto a linear scale.

[0043] According to a particular implementation, the offset of the compressed detection signal may then comprise the addition of a value corresponding to a desired minimum level of the audio signal.

[0044] Depending on the particular implementation mode, the detection threshold and / or the value corresponding to the desired minimum level of the audio signal are programmable. These parameters can thus be adjusted by the user according to his expectations and / or his experience in metal detection.

[0045] In a particular embodiment, generating the audio signal may comprise generating an acoustic signal having a fundamental component of determined pitch, as well as modulating said acoustic signal by the compressed and shifted detection signal.

[0046] According to a particular embodiment, the generation of the audio signal further comprises a change in timbre of the acoustic signal, between two respective portions of the dynamics of the detection signal. The audio signal is thus nuanced by spectral contents slightly modified from one zone to another of the dynamics of the detection signal, in addition to the variation in the sound level.

[0047] For better user experience, the change in timbre of the acoustic signal between two contiguous portions of the detection signal dynamics can be gradual from one to the other.

[0048] According to a particular embodiment, the reduction of the dynamic range of the received detection signal further comprises, after clipping the detection signal but before applying the overall dynamic compression rate to the clipped detection signal, a half-wave rectification of said clipped detection signal. In other words, any negative values of the dynamic range of the detection signal after subtraction of the audibility threshold value are eliminated. This ensures that there is a minimum value that is equal to zero at the input of the downstream part of the processing chain.

[0049] In a second aspect, the invention also provides a computer program product comprising instructions which, when the program is executed by a computer, cause the latter to implement all steps of the method according to the first aspect.

[0050] A third aspect relates to a data processing device comprising a processor configured to implement all the steps of the method according to the first aspect.

[0051] Finally, a fourth and final aspect of the invention also relates to a portable metal detection apparatus comprising a device according to the third aspect above as well as at least one electroacoustic transducer for restoring the audio signal to a user of the apparatus. [Brief description of the drawings]

[0052] Other characteristics and advantages of the invention will become apparent from reading the description which follows. This description is purely illustrative and must be read in conjunction with the appended drawings in which: [ Fig. 1 ] there figure 1is a schematic representation of the main functional elements of an example of a metal detector in use conditions, during a detection session conducted by a user; [ Fig. 2 ] there figure 2 is a functional diagram representing, in the form of a block diagram, the architecture of an electromagnetic metal detector capable of implementing the audio restitution method according to embodiments; [ Fig. 3 ] there figure 3 is a graph of the evolution, as a function of the distance "d" between the detection head and the target, of the attenuation (in decibels, or dB) of the amplitude of the electromagnetic detection signal coming from the detection head of a metal detector; [ Fig. 4 ] there figure 4 is a functional diagram (block diagram) of a device according to embodiments of the invention, suitable for implementing the method according to embodiments of the invention; [ Fig. 5 ] there Figure 5is a graph showing the evolution, as a function of time, of an example of an electromagnetic detection signal as it comes from the detection head during a single scanning movement of the detection head of a metal detector over a target; [ Fig. 6 ] there figure 6 gives graphs of an example audio signal for the sound reproduction of the signal from the figure 4 according to a constant dynamic method, for respective gain values; [ Fig. 7 ] there figure 7 gives a graph of an example audio signal for the sound reproduction of the signal from the figure 4 by implementing a method of limiting the dynamics of the audio signal by high and low clipping; [ Fig. 8 ] there figure 8 gives a graph of an example audio signal for the sound reproduction of the signal from the figure 4 by implementing the dynamic compression method in accordance with embodiments of the invention; [ Fig. 9 ] there figure 9gives the graphs of the figure 7 and of the figure 8 superimposed on each other, to illustrate the comparison between the shape of the audio signal for the sound reproduction of the signal of the figure 4 according to the method of limiting the dynamics of the audio signal by high and low clipping and the shape of this audio signal for sound reproduction according to the method of compressing the dynamics in accordance with embodiments of the invention, respectively; and, [ FIG. 10 ] there figure 10 is a graph illustrating an implementation mode in which the target detection signal is segmented so as to define several zones in which different respective compression levels can be had. [Description of embodiment(s)]

[0053] In reference to the figure 1 ,A modern 10 metal detector has a three-element architecture, these three elements communicating with each other via a digital wireless link, for example a radio link. These three elements are: firstly, a detection disc or detection head which incorporates the transmission and / or reception coil(s); secondly, a control module 12, also called a remote control, which allows all of the various functions of the detector to be configured using a suitable human-machine interface (or HMI); and thirdly, an audio reproduction device 13, which may be a wired audio headset or preferably a wireless audio headset, one or more loudspeakers or a bone conduction device, comprising electroacoustic transducers (for example, one for each ear) capable of reproducing the detection signal in a form audible to the user.

[0054] It should be noted that all or part of the electronics for managing the transmission, reception, and processing of detection signals to achieve target detection and discrimination can be integrated into the detection head, or into the remote control, or be distributed between the detection head and the remote control.

[0055] In addition, the detector is usually supplied with, as an accessory, a rod having a handle and an armrest. The end of the rod opposite the armrest and the handle is adapted to support the search coil, which can be removably coupled to it. The remote control is adapted to be fixed on the rod, just in front of the handle as shown in figure 1 .

[0056] Thus, when user 1 extends the cane thus equipped with the detection disc and the remote control in front of him, and he holds said cane in one hand via the handle with the corresponding arm firmly wedged in the armrest as shown in figure 1 , he can make the detection disc perform lateral scanning movements parallel to the ground, while having the possibility of intervening with his other hand on the remote control in order to modify the detector settings if necessary.

[0057] Those skilled in the art will appreciate that the embodiments of the invention are not limited by the technology of the metal detector. By way of example only, and with reference to the technologies set forth in the introduction to the description, it will be considered here for the purposes of this description that this is an induction balance (IB) detector operating in CW mode.

[0058] It is possible to decompose the architecture of a metal detector in the manner that will be described below with reference to the functional diagram of the figure 2 . This functional breakdown makes it possible to distinguish the main functions provided by the detector, but is not limiting of the way in which these functions are implemented at the hardware level, certain functions being able to be grouped or conversely broken down within determined hardware elements, which can also implement additional functions which are not considered here because they are not essential for the presentation of embodiments of the invention.

[0059] The detector comprises an electromagnetic transmission / reception assembly 21, comprising the coil(s), namely at least one transmission coil and at least one reception coil, knowing that these two coils can also be one. This assembly 21 is installed in the detection head 10 of the figure 1 .

[0060] The detector also includes a 22 analog front-end module (or eVLF FE, where "FE" stands for "front end " in English, and where the term "eVLF" is derived from the term "VLF" (from the English "Very Low Frequency") which designates the frequency band from 3 kHz to 30 kHz), as well as a digital front-end module 23 (or DFE, put for "Digital front end " in English). The set of these two front modules ensures the emission (Tx) of the incident magnetic field.

[0061] In the reception direction (Rx), the front-end modules 22 and 23 ensure the transposition of the analog signals adapted to the coils, into very low frequency signals (generally less than 100 Hz) representative of the passage of the detection head over the target, by applying a frequency transposition, which can be achieved by demodulation and associated filtering. Most often, the analog front-end module 22 integrates the power amplifications of the transmission chain (Tx) and the low noise amplifiers of the reception chains (Rx). The VLF band is generally used but lower frequencies and higher frequencies, i.e., frequencies in a wider frequency range than the VLF band, for example between 1 kHz and 100 kHz, can also be used for metal detection during specific searches.This is why we speak here of "extended" VLF band (noted eVLF), this expression (and this acronym) not necessarily belonging to the everyday language of the person skilled in the art in the technical field considered. The analog front-end module 22 can also integrate the demodulators, but the demodulators are most often produced in the digital front-end module 23 for the narrow-band detectors.

[0062] Those skilled in the art will further appreciate that the analog / digital separation depends on the design of the equipment, and is not limited, in practice, by the examples considered herein.

[0063] The device 10 further comprises a detection processor 24, which is adapted to apply the detection processes in order to determine the presence and possibly the type of metal targets.

[0064] In practice, the detection processor 24 can be implemented in the form of a miniaturized electronic circuit, which allows it, in certain metal detectors, to be integrated into the detection disc 11 ( figure 1 ). Such a circuit is adapted to digitize and analyze the detection signals and to produce the detection data and the discrimination data, this data then being sent in real time to the user interface 25 (see below), for example by digital radio link as shown in figure 1 or by wired connection, for sound reproduction via the audio headset 13 and possibly, in addition, for display on the remote control 12.

[0065] The performance of the detector can be affected by the different levels of mineralization of the ground that can be encountered depending on the use cases. For example, this can be mineralization of natural origin of magnetic type: iron oxide, ferrites, magnetites. It can also be point mineralization linked to ancient places of human occupation (also magnetic), for example hearths, terracotta, ferrites, slag, etc. It can also be seaside mineralization which can range from magnetic grade (black sands) to electrically conductive grade (salt water), depending on the beaches and regions. figure 2 , electromagnetic disturbances are symbolized by cloud 26.

[0066] Detection is also affected by electromagnetic interference (or EMI, from the English “Electromagnetic interference”) which are numerous, particularly in urban areas (high voltage lines, electrical transformers, electric fences, power lines, radio relays, mobile phones, computers, televisions, other metal detectors operating nearby, etc.), as well as by metal pollution. figure 2 , electromagnetic disturbances are symbolized by cloud 27.

[0067] Finally, the detector includes a human-machine interface also called user interface 25 (or HMI, put for "Human-Machine Interface",in English). The user interface 25 makes it possible to configure the equipment, on the one hand, and to restore information indicating the possible presence of metal targets and, possibly, also information relating to their identification, as well as possibly detection assistance information to assist the user in his search, on the other hand. As already indicated in the introduction, audio reproduction is the interface most commonly used for this, and it is this which is the subject of the present invention. The user interface 25 thus comprises the audio headset 13 of the figure 1 and / or any other electroacoustic transducer, for example one or more loudspeakers.

[0068] However, all or part of the aforementioned information may also be restored in another form, for example in a visible form by display on a screen. The user interface 25 comprises for this purpose the remote control 12 ( figure 1 ). The remote control also allows the user to adjust the main detection settings such as sensitivity, discrimination, ground effects, tones, the frequency(ies) used for the signals emitted and processed by the detection head, the volume, etc., but also to select factory programs or those previously created by the user himself with the remote control. These settings modify the detection, which is symbolized at figure 2 by arrow 20 in dotted lines.

[0069] In summary, the user interface can be implemented in the remote control 12, and it also includes the audio headset 13 for the audio reproduction of the detection signal to the user, the latter being able to be regarded as a peripheral dedicated to the reproduction in sound form of the detection signal. In certain embodiments, the audio headset can also include buttons or the like for carrying out certain adjustments of the detector, in particular but not only adjustments relating to the reproduction in sound form of the detection signal. Furthermore, it will be noted that the audio signal which is generated for the audio reproduction of the electromagnetic detection signal is derived from said detection signal received in "raw" form by the detection head by demodulation operations (in the broad sense) and derivation(s) / filtering(s) to arrive at a very low frequency signal (with a frequency generally less than a hundred Hertz) characteristic.

[0070] The electromagnetic detection signal (more simply called here the "signal") has an amplitude which varies when the detection head moves above the target in proportion to the fields produced in the target by the circulation of eddy currents and received by the detection head. For small targets, the amplitude of the magnetic field at the target is proportional to 1 / (a 2< +d 2< ) 3 / 2< , where "a" denotes the diameter of the detection head (and therefore substantially the diameter of the transmitting and / or receiving antennas), and where "d" denotes the distance from the head to the target. The amplitude of the signal received by the receiving head when it scans a target at a given distance corresponds to a round trip of the signal. It therefore varies approximately with a relationship of the form K x 1 / (a 2< +d 2< ) 3< .

[0071] In a simplified case where the distance "d" from the head to the target is large relative to the diameter "a" of the head, curves 31 and 32 of the figure 3 illustrate, respectively, the amplitude of the magnetic field at the target which varies in 1 / d 3< i.e. in the form K x 1 / d 3< , and the signal received at the receiving head which varies in 1 / d 6< i.e. in the form K x 1 / d 6< , where K is a proportionality factor depending on the conditions of use. Curves 31 and 32 correspond to an example of a use case for a small target, in which the detection head has a diameter of 28 cm (a=28 cm) and in which the reference distance d 0 (which gives an attenuation of 0 dB) is substantially equal to 1 cm.

[0072] The characteristics which can be used to present to the user information relating to the detection of a metallic target, and discrimination information enabling the user, to a certain extent, to identify this target, will now be explained.

[0073] Reactivity is a parameter that determines the detector's behavior in terms of analysis speed and selectivity. Selectivity refers to the separation power between targets. If a soil is polluted with ferrous materials, ferrites or other mineralized debris (hereinafter "pollutants"), soil penetration is generally reduced and so is the detector's ability to detect targets close to ferrous materials. Under these conditions, it is possible to choose a high reactivity level, which speeds up signal analysis. On the other hand, if the ground is clean, i.e. relatively free of pollutants, it may be appropriate to slow down the reactivity and scanning, in order to be more sensitive to deep masses and penetrate further into the soil. Adjusting the detector's "reactivity" parameter therefore makes it possible to make the detector more or less fast and selective.

[0074] Depending on the reactivity, the duration of the audio signal varies as a target passes under the moving detection head: with a relatively lower reactivity, a target gives a relatively longer sound and, vice versa, with a relatively higher reactivity, the same target gives a relatively shorter sound. It will be noted that the duration of the sound of false signals (corresponding to cracking of ferrous metals for example), also varies in proportion, i.e., varies in the same way as the duration of a target signal depending on the reactivity.

[0075] There are basically three characteristics to consider in a sound: the pitch of the sound, which determines its degree of elevation between what is perceived (by the human being) as low and what is perceived as high; the intensity of the sound, which differentiates a loud sound from a soft sound; and, the timbre of the sound, which makes it possible to distinguish an "aggressive" sound from a "soft" sound.

[0076] Furthermore, a sound is said to be "pure" when the sound wave is perfectly sinusoidal.

[0077] The pitch of a pure tone corresponds to its vibration frequency, which is measured in hertz (number of periodic vibrations per second). The faster the vibration, the higher or sharper the sound. Conversely, the slower the vibration, the lower or bassier the sound. Frequencies audible to humans range from 20 Hz to 20,000 Hz (20 kHz). A sound is generally considered high-pitched when its fundamental frequency is above (approximately) 2,000 Hz (2 kHz). It is considered low-pitched when its fundamental frequency is below 200 Hz. In nature, however, a sound is almost never pure, that is, it is made up of several sounds of different frequencies. Furthermore, differences in the timbre of a sound result mainly from the combination of the different harmonics of a sound, with different intensities. The timbre therefore depends mainly on the frequency spectrum of the sound, as explained below.

[0078] The intensity of a sound is a quantity that gives an indication of the "strength" of the sound: the higher the sound intensity, the louder the sound perceived by the human ear. Sound intensity is noted with the letter "I" and is expressed in watts per square meter (Wm -2< ). This unit indicates that it corresponds to a "flow" of energy per unit of surface area. The human ear can on average perceive sounds with a sound intensity greater than a value of the order of 10 -12< Wm -2< . This minimum sound intensity is called the audibility threshold "I 0 ". A sound with a very high sound intensity can cause pain in humans, as well as partial or total hearing loss. It is generally estimated that the pain threshold in a normal individual corresponds to a value of approximately 10 Wm -2< which corresponds to 130 dB SPL, knowing that the value on a decibel scale of a sound intensity I is given by 10 x log(I / I 0 ).Furthermore, repeated exposure to sound waves of high intensity but below the maximum intensities, without appropriate protection, can also cause damage to the human ear.

[0079] The timbre of a sound refers to the set of particular characteristics that allow us to distinguish two sounds having the same fundamental frequency and the same intensity but which do not give the same physiological sensations. For example, the same musical note, such as "C", produces a specific sensation depending on whether it is played by different musical instruments, even though they then produce sounds of the same fundamental frequency and possibly the same intensity. The timbre of a sound is therefore independent of its pitch and intensity. In nature, it is specific to the instrument or voice that emits it. Physically, it is linked to the number and relative intensities of the harmonics that make up the sound.

[0080] It should be noted that the feeling also depends on the shape of the sound, that is to say the waveform of the different vibrations that compose it, which is linked to the attack transients, the extinction transients, and the body of the sound, that is to say its properties between the attack and extinction transients. The attack of a sound can be relatively slow or relatively fast, and relatively gradual or relatively abrupt. The body of a sound can be relatively regular or with peaks, substantially sinusoidal or more complex, and with more or less amplitude. We thus classically distinguish a soft sound from a harsh (or aggressive) sound even when they have substantially the same pitch and the same intensity.

[0081] In order to overcome the drawbacks of the technique which were presented in the introduction, embodiments of the invention propose a method and a device making it possible to modify the dynamic range of the audio signal associated with the target which is returned to the user, so as to make this dynamic compatible with the dynamic actually usable by the human auditory system, while allowing a dissociation between the variation in level felt and the spectral content of the audio signal.

[0082] According to embodiments, this modification of the dynamic range is carried out, for the simplest implementation, by a reduction thereof obtained by the method and the device illustrated by the functional diagram of la figure 4 .This diagram is purely schematic, and the elements of this block diagram symbolize both the stages of the sound reproduction process according to the methods of implementing the process and the functional means of a device for this implementation. In other words, the diagram of the figure 4 illustrates the detection signal processing chain that leads to the production of an audio signal Audio_Lev whose pitch, intensity and timbre are generated in such a way as to reproduce in a sound manner, comfortably audible to the user, the entire amplitude dynamics of the target detection signal as well as, where appropriate, associated discrimination information when such information is also available.

[0083] In reference to the figure 4, the target detection signal 20, resulting from processing (not shown) intended to isolate it from noise, is routed to a linearization module 21 which is adapted to carry out a conversion of the detection signal into decibels (dB). This makes it possible to obtain a signal Signal_Amp whose amplitude varies approximately linearly as a function of the distance d from the target to the detection head. In one embodiment, the linearization device 21 may be a simple logarithmic amplifier. It goes without saying that other linearization methods may be envisaged.

[0084] A logarithmic amplifier can be made of analog electronic components. As is known, an analog logarithmic amplifier is a nonlinear amplifier that produces an output representing the logarithm of the input signal.

[0085] Alternatively, the logarithmic amplifier 21 can also be produced by digital processing, by linear or polynomial interpolation, for example.

[0086] The detection threshold of the metal detector is then subtracted from the linearized signal Signal_Amp. This subtraction operation is symbolized here by a subtractor 22, whose positive input "+" receives the linearized signal Signal_Amp and whose negative input "-" receives a value representative of the detection threshold. In certain embodiments, as shown in figure 4 , this value is generated by a threshold adaptation module which allows the detection threshold of the metal detector to be varied. This variation can be controlled "on demand" by the user using the detector's remote control, depending on their feelings and / or their metal search strategy. It can also result from the implementation of a particular pre-recorded program.

[0087] In one embodiment, the signal resulting from the aforementioned subtraction operation can then undergo a single-wave rectification, symbolized in figure 4 by the symbol of a diode 23. Those skilled in the art will appreciate that the negative values of the signal at the output of the subtractor 22 correspond to signal levels Signal_Amp which are lower than the detection threshold of the detector. In other words, any negative values of the dynamics of the detection signal after subtraction of the value of the audibility threshold are eliminated by the filtering carried out by the rectifier 23. This ensures that there is a minimum value equal to zero at the input of the downstream part of the processing chain.

[0088] The value of the signal at the output of the rectifier 23 therefore varies from 0 dB if the target signal is equal to said threshold, to the maximum value of the dynamic range between the threshold and the maximum amplitude (in decibels) of the detection signal, i.e. potentially more than 100 dB as explained in the introduction to this description.

[0089] In accordance with the teaching of the embodiments of the invention, this deviation from the threshold is in turn multiplied by the desired dynamic compression rate. This is equal to unity (i.e. 1) for no compression. To obtain compression, it is strictly between 0 and 1.

[0090] The dynamic range compression ratio is typically set to a value between 0.2 and 0.5 to reduce a dynamic range of 0 to 100 dB to a dynamic range of 0 to 30 dB, for example. figure 4, this compression operation is represented by a multiplier, or multiplication operator 24, receiving the signal at the output of the rectifier 23 on a first input, and receiving a value representative of the compression ratio on a second input. In certain embodiments, as shown in figure 4 , this value can be provided by a compression adaptation module 24a which allows the dynamic compression rate to be varied.

[0091] A value representative of the audibility threshold "I 0 " is then added, which corresponds to the minimum desired intensity of the audio signal, i.e. the weakest sound returned to the user ( "baseline" ) . To the figure 4, this addition is symbolized by an adder, or addition operator 25, which receives the signal at the output of the multiplier 24 on a first input, and which receives the value representative of the audibility threshold "I 0 " on a second input. In certain embodiments, as shown in figure 4 , this value is provided by a module 25a for adapting the audibility threshold. This threshold can in fact be variable, in order to be able to be adapted to the acoustic perception specific to each user.

[0092] In one example, the output signal of the adder 25 can vary with a desirable dynamic range, i.e. within an ideal range, for example between 20 dB and 50 dB.

[0093] In some embodiments, the inverse operation of the linearization performed by the linearization module 21 can then be carried out. In other words, the signal is converted from decibel values to linear values. This operation can, for example, be carried out by an exponential function, with an adequate gain.

[0094] The output of the conversion module 26 amplitude modulates an audio signal produced by a sound generator 27, which is used to generate the sounds associated with the detection of a determined target. More particularly, the sound generator 27 is adapted to produce sounds with a pitch (frequency) which is a function, for example, of the distance between the detection head and the detected target. As will be understood, the device according to embodiments of the invention which is shown in figure 4has the sole function of adapting the dynamics of the intensity (volume) of the sound delivered at the output, to the dynamics of the detection signal received at the input.

[0095] The aforementioned function of modulating the amplitude of the audio signal generated by the generator 26 by the signal at the output of the conversion module 26 is symbolized, at figure 4 , by a multiplier or multiplication operator 28. This receives on a first input the signal delivered by the sound generator 27 (modulated signal), and receives on a second input the signal at the output of the conversion module 26 (modulating signal). The multiplier delivers the output signal of the device, namely an audio signal Audio_Lev with variations corresponding to those of the amplitude of the input signal 20, but within the desired sound intensity interval, i.e. the desired dynamics of the audio signal.

[0096] In some embodiments, the audio signal generator 27 may further receive information from a timbre variation module 27, making it possible to vary the timbre of the output audio signal based on the discrimination information generated by the metal detector, as presented in the introduction.

[0097] Those skilled in the art will appreciate that the final conversion from a logarithmic scale (decibel scale) to a linear scale performed by the conversion module 26 is not necessarily necessary. Indeed, certain devices, such as analog-to-digital converters (DACs) or audio amplifiers, can receive the decibel information directly as input.

[0098] Furthermore, in certain embodiments, a sound cutting (or "mute") circuit 99 may be arranged at the output of the detection threshold subtraction module 22. This module makes it possible to detect whether the signal is below the threshold level and to control, if necessary (for example depending on the audio mode chosen by the user), a "mute" of the audio signal for signals below the detection threshold. This "mute" function is here symbolized by a controlled switch 29 operating in all or nothing, arranged between the output of the modulator 26 and the audio headset 13, and controlled by the circuit 99. It will be noted that a "mute" may also be generated, in the same way, in the case of the detection of targets of a certain particular type or of several particular types, in order not to unnecessarily alert the user if the detected target does not correspond to the type of target sought.

[0099] Finally, those skilled in the art will appreciate that an adjustment of the overall audio gain of the processing chain of the figure 4 , not shown here, can be integrated into the processing chain. Alternatively, it can be implemented in the form of a volume control associated with the audio headset 13. The latter can of course be replaced by a loudspeaker without this changing the basis of the invention. The gain control is not part, in itself, of the present disclosure insofar as the invention relates to the dynamics of the output audio signal to satisfactorily restore the entire dynamics of the input detection signal.

[0100] In summary, the invention makes it possible to compress the dynamics of the target detection signal in order to correspond to the desired dynamics of the audio signal which is generated to reproduce this detection signal in an audible manner, the dynamics being understood here as being the difference between two extreme values. With regard to the detection signal, the dynamics is understood more particularly as the range of amplitude values of the signal which is between a lower end and an upper end. For the audio signal reproduced to the user, the dynamics represents the ratio between the loudest sound (having the highest sound intensity) and the quietest sound (having the lowest sound intensity).The invention makes it possible, in the context of metal detection, to avoid the loss of detection signal at the lower end and the limitation or clipping at the upper end of the amplitude variation range of the detection signal, in the corresponding audio signal which is generated in order to restore this detection signal to the user in an audible manner.

[0101] This technical effect will now be explained with reference to the graphs of the figures 5 to 9 .

[0102] There Figure 5 ,first of all, shows the evolution of the amplitude, as a function of time, of an example of an electromagnetic signal for detecting a target. In this example, a metal target is detected within a time window (time window) of a width equal to 40 ms, between the instants t=30 ms and t=70 ms, plotted on the abscissa. The portion of the signal represented in the figure comes from the detection head of a metal detector during a single scanning movement of the detection head over a metal target. More specifically, on the Figure 5 the detection signal is represented as an ordinate on a logarithmic scale. This is for example the version Signal_Amp of the detection signal taken at the output of the linearization module 21 of the figure 4On this scale, curve 51 representing the signal has a bell shape: the rising part corresponds to the approach of the detection head towards the target, the peak of the Gaussian corresponding to the moment (at time t=50 ms in the example) when the head is closest to the target and the falling part of the Gaussian corresponding to the distance of the detection head from the target.

[0103] In fact, curve 51 corresponds to the useful detection signal S_in mixed with detection noise which comes from the different disturbers symbolically illustrated in figure 2by clouds 26 and 27. The maximum noise level N is represented by the horizontal line 52 in dotted line. The maximum amplitude of the noisy signal (S_in + N) is represented by the horizontal line 53 in broken line. As can be seen, the dynamic range Dyn_Sin of the useful detection signal (noise-free signal S_in), here represented by the vertical arrow 50 in broken line, corresponds to a range of amplitude values which extends, on the logarithmic scale represented, between the minimum and maximum values of the noisy signal S_in+N which are equal to 20 dB and 120 dB, respectively. This dynamic range 50 is therefore equal to 100 dB.

[0104] There figure 6 gives graphs of an example audio signal for the sound reproduction of the signal from the figure 4according to a constant dynamic method, in which one would simply vary the gain of the processing chain of the detection signal to be restored generating the corresponding audio signal. More particularly, graphs 61 and 62 correspond to maximum gain values G+ and minimum G-, respectively, which would be chosen in the following manner: the gain would be set to the maximum value G+ so that the entire audio signal is approximately 10 dB(A) above the ambient acoustic noise level (which can be estimated at a minimum of 30 dB(A) as explained in the introduction, but is attenuated due to the use of headphones by a value that can be estimated to be equal to 10 dB) in order to be able to be distinguished from this noise; whereas; the gain would be set to the minimum value G- so that the entire audio signal is below the limit level imposed by the regulations (set at 80 dB(A) as explained in the introduction).

[0105] It will be appreciated that the restitution of the dynamic 50 of the amplitude of the detection signal of the Figure 5gives, in the case of the gain G+ corresponding to curve 61, values of the audio signal which can reach 130 dB SPL (A), thus largely higher than the maximum threshold of hearing comfort Lev_max of 80 dB SPL (A) represented by the horizontal line 64 in broken line, which is not tolerable and therefore requires clipping the audio signal at this threshold in order to protect the user's hearing aid. Conversely, in the case of the gain G- corresponding to curve 62, the audio signal restoring the full dynamic range of the detection signal (which is 100 dB) can drop to values of the audio signal largely below the threshold of audibility Lev_min represented by the horizontal line 63 in dotted line, that is to say below 30 dB SPL (A) taking into account the noise. In this case, the corresponding variations in the detection signal are lost in the audio signal that is played back to the user.

[0106] The combined effect of limiting the Dyn_Sout dynamic range resulting from top-end clipping and bottom-end loss of the detection signal amplitude variations is graphically illustrated by curve 70 of the figure 7 . This curve represents the evolution, as a function of time and for a given value of the gain of the restitution chain, of the audio signal S_Ltd which is limited by the two phenomena presented above. As can be seen, the dynamic Dyn_Sout of the audio signal which restores the detection signal to the user is advantageously included in the range of comfortable values, between 20 dB SPL (A) and 80 dB SPL (A) corresponding to a signal which is both audible taking into account the ambient noise and respectful of the integrity of the user's hearing aid.

[0107] Those skilled in the art will appreciate that any value of the gain G of the processing chain between the minimum value G- and the maximum value G+ presented above with reference to the graphs of the figure 6, presents more or less, to respective degrees weighted relatively to each other, the disadvantages mentioned above with regard to curve 62 and curve 61, respectively. The higher the gain G between the values G- and G+, the more the sound reproduction of the variation of the low amplitude values of the detection signal is favored, to the detriment of the reproduction of the variation of the high amplitude values which are suppressed by clipping the audio signal. Conversely, the lower the gain G between the values G- and G+, the more the sound reproduction of the variation of the high amplitude values of the detection signal is favored, to the detriment of the reproduction of the variation of the low amplitude values which are lost in the audio signal to the extent that they are not audible to the user.

[0108] In other words, the user can certainly vary the gain linearly for the entire amplitude range of the detection signal in order to favor the sound reproduction of the low amplitudes or that of the high amplitudes of the detection signal, but in all cases the useful dynamic range Dyn_Sout of the audio signal represented in figure 6 by the vertical arrow 60 in broken line only very imperfectly restores the dynamic Dyn_Sin of the target detection signal (see arrow 50 on the Figure 5). In other words, again, the variation of the gain of the processing chain carried out linearly in the entire range of variation of the detection signal allows neither more nor less than to find a compromise between the loss of the variations of the detection signal at the lower end of the amplitude range on the one hand, and the limitation by clipping of the audio signal at the upper end of the amplitude range, on the other hand, in the corresponding audio signal which is generated in order to restore this detection signal to the user in a sound manner.

[0109] Conversely, the audio signal generated in accordance with embodiments using the technique of compressing the dynamics Dyn_Sin of the detection signal supplied as input, with a compression rate adapted to correspond to the useful dynamics Dyn_Sout of the output signal, which is represented by graph 81 of the figure 8 ,respects the variations in the amplitude of the detection signal. This can be seen from the fact that curve 81 of the figure 8 corresponds to curve 51 of the Figure 5 , by being "picked up", that is to say compressed in the interval of values (on the decibel scale) between 30 dB SPL (A) and 80 dB SPL (A).

[0110] To facilitate comparison, curve 71 of the figure 7 and curve 81 of the figure 8 are superimposed on the figure 9 .

[0111] In an implementation mode illustrated by the graph of the figure 10 ,the target detection signal is segmented to define several portions or zones in which there can be associated compression rates, called local dynamic compression rates, which can be different from one zone to another. The graph shows, on the ordinate, the level of the audio signal in decibels (dB), noted "Audio_Lev" as a function of the amplitude (in dB) of the target detection signal noted "Signal_Amp" and represented on the abscissa. In the figure 10 , the reference level giving 0 dB corresponds to the maximum level which is comfortable for the user, and which the user can adjust with the volume control. Furthermore, in the example shown the Dyn_Sout dynamic of the audio signal is here 40 dB, between - 40 dB and 0 dB.

[0112] In the implementation mode shown in the figure 10, the detection signal is segmented into three portions or zones, respectively Z1, Z2 and Z3, corresponding to respective intervals of its dynamics, for example contiguous two-by-two zones. This is however not limiting. In particular, those skilled in the art will appreciate that the dynamics of the target detection signal can be subdivided into only two portions, or conversely into four or more portions.

[0113] In the example considered here, the first portion corresponds to the so-called "false signal" zone which corresponds (for example) to a dynamic range of 10 to 20 dB above the detection threshold. A second portion Z2 corresponds to the so-called "average signal" signal zone, from the end of zone Z1 up to, for example, 60 dB above the detection threshold, followed by the portion Z3 corresponding to the strong signal zone, after 60 dB above said threshold. The portions of the dynamics corresponding to the first and last zones, Z1 and Z3 respectively, can thus be modified by lower local dynamic compression rates than that applied to the portion of the dynamics corresponding to zone Z2, with the aim, for example, of magnifying the target signals in zone Z2 of the average signals.

[0114] It should be noted that there are in fact two additional implicit zones: the first is the Z0 zone of the dynamics which is below the detection threshold which can correspond to a "mute" of the reproduced audio signal; and the second is the Zsat zone of electrical (or digital) saturation which corresponds to the highest levels of the dynamics of the detection signal. For reasons relating to simplicity of implementation, the different zones can be defined by linear segments. Smoothing can also be used to erase the slope breaks at the "boundaries" between two contiguous zones, for example with a Spline type smoothing.

[0115] It will be appreciated that the overall dynamic compression that is sought does not prohibit at least one rate of local dynamic modification from being greater than or equal to unity. In other words, one or more zones of the dynamics of the detection signal can be locally extended (we then speak of dynamic extension or expansion, instead of dynamic compression), although the dynamics are compressed globally, that is to say if we consider globally the entirety of this dynamics.

[0116] In a further improved embodiment of the invention, at least certain portions of the dynamics may be associated with respective timbres of the audio signal. For example, each zone among zones Z1, Z2, Z3 and Zsat may be associated with a different timbre. Furthermore, the change in timbre may be done gradually when moving from one portion of the dynamics to another. In other words, the change in timbre of the acoustic signal between two contiguous portions of the dynamics of the detection signal is progressive from one to the other.

[0117] Those skilled in the art will appreciate that the changes in timbre of the acoustic signal generated by the generator 27 which have been presented above are found in the audio signal which is returned to the user. Thus, these different modes of implementation provide a selective spectral enrichment (i.e. which is a function of the level of the signal) which improves the sound perception, by the user, of the sounds generated to restore the detection signal of the metal target.

[0118] The segmentation described above also makes it possible to apply the spectral enrichments associated with the target discrimination methods in a differentiated manner, depending on the level of the received signal and therefore depending on the dynamic range in which it is located, it being observed that the accuracy of target discrimination is generally degraded for weak signals and for saturated signals. Several implementations can be considered, which are specified in the following.

[0119] In the first implementation, the oscillator that generates the final sound proceeds by reading a wavetable. Each zone boundary is associated with a previously calculated waveform, for example based on a desired spectral content. The oscillator that generates the sound interpolates between the two waveforms defined at the boundary of the activated zone. For example, a waveform with a relatively "soft" spectrum can be reserved for the Z1 zone of "false signals" (the zone of the weakest signals, lost in the noise), and a waveform with a more "aggressive" spectrum can be reserved for the signals in the Z3 zone of the dynamic range, whose amplitude is the highest. Similarly, waveforms reminiscent of the spectrum of saturated signals can be used for the loudest signals.The advantage of the variable spectral modification as a function of the level is to allow the user's ear to improve the perception of level, without resorting to a frequency variation as a function of the level, which makes it possible to reserve the possibility of using a frequency variation in connection with discrimination information. Waveform reading makes it possible to generate varied sounds, and even, with a single oscillator, to simulate third or fifth chords by generating waveform tables whose fundamental components are calibrated on the 4th, 5th, 6th harmonics of the signal.

[0120] In a second implementation, the spectral change within an area can be obtained by subtracting two waveforms whose phase depends on the level of the detection signal at the input of the device. This synthesis mode allows the generation of pulse-type signals of variable widths (pulse width modulation or PWM) by subtracting two "sawtooth" (or SAW) type signals. "Sawtooth" ) .It is thus possible to generate signals whose spectral content varies radically depending on the phase shift value, from a practically "flat" spectrum for low duty cycles, to a "square" signal (spectrum with a 1 / f decay, without odd harmonics) for a duty cycle of 50% (see the spectra associated with PWM signals above). An advantage of this type of synthesis is that it allows the generation of signals similar to PWM signals (with recognizable spectral shifts) while having control over the overall richness of the spectrum by acting on the basic signals (initially "sawtooth" signals) used for the subtraction.

[0121] In a third implementation, spectral change within an area can be achieved by frequency modulations of the oscillator used for sound generation, using methods developed by John Chowning in 1967.

[0122] In a fourth implementation, the spectral change within a zone can be obtained by subtractive synthesis, for example by inserting into the audio signal path a filter whose characteristics are dynamically varied.

[0123] In a fifth implementation, spectral changes can be achieved by additive synthesis, for example by implementing multiple audio oscillators whose mixing can be determined based on the desired frequency spectrum.

[0124] All or part of the implementations given above as examples can also be combined with each other, so as to obtain a spectral enrichment which is more complex in order to better respond to particular use cases.

Claims

1. Method for generating an audio signal for the acoustic rendering, to the user of a portable metal detection apparatus (10), using at least one electroacoustic transducer, of an electromagnetic target detection signal the amplitude of which may vary with a determined dynamic range, the method comprising the following steps: - reducing the dynamic range of the detection signal so that it corresponds to a desired dynamic range of the audio signal; and - generating (27, 28) the audio signal on the basis of the detection signal with the reduced dynamic range, wherein: - reducing the dynamic range of the detection signal comprises: - - clipping (22) the detection signal on the basis of a determined detection threshold; clipping (22) the detection signal comprising clipping said signal below the detection threshold; - - applying (24), to the clipped detection signal, an overall determined dynamic range compression rate, which depends on the ratio between the dynamic range of the detection signal and the desired dynamic range of the audio signal; and - - shifting (25) the compressed detection signal so as to set a desired minimum level of the audio signal; and characterized in that : the overall determined dynamic range compression rate is strictly between 0 and 1; and - applying (24) the overall dynamic range compression rate to the clipped detection signal comprises segmenting the dynamic range of the detection signal into a plurality of separate portions (Z1, Z2, Z3), and also applying respective local dynamic range modification rates to different portions of the clipped detection signal.

2. Method according to Claim 1, wherein some portions of the dynamic range of the detection signal are defined by linear segments.

3. Method according to either one of Claims 1 and 2, wherein at least one local dynamic range modification rate is greater than or equal to unity.

4. Method according to any one of Claims 1 to 3, furthermore comprising, before the step of reducing the dynamic range of the detection signal, logarithmic conversion (21) of said detection signal on a logarithmic scale, and wherein applying (24) the overall dynamic range compression rate to the clipped detection signal or applying a local dynamic range modification rate to a determined portion of the clipped detection signal comprises multiplying said clipped detection signal or said clipped detection signal portion, respectively, by said overall dynamic range compression rate or by said local dynamic range modification rate, respectively.

5. Method according to Claim 4, furthermore comprising, after reducing the dynamic range of the detection signal and before generating the audio signal, inverse conversion (26) of the logarithmic conversion of the received detection signal in order to return the dynamic range of the compressed and shifted detection signal to a linear scale.

6. Method according to either one of Claims 4 and 5, wherein: - shifting (25) the compressed detection signal comprises adding a value corresponding to a desired minimum level of the audio signal.

7. Method according to any one of Claims 1 to 6, wherein the detection threshold and / or the value corresponding to the desired minimum level of the audio signal are programmable (22a, 25a).

8. Method according to any one of Claims 1 to 7, wherein generating the audio signal comprises generating (27) an acoustic signal having a fundamental component of a determined level, and modulating (28) said acoustic signal with the compressed and shifted detection signal.

9. Method according to any one of Claims 2 to 8, wherein generating the audio signal furthermore comprises changing the timbre (27a) of the acoustic signal between two respective portions of the dynamic range of the detection signal.

10. Method according to Claim 9, wherein the change in timbre of the acoustic signal between two contiguous portions of the dynamic range of the detection signal is gradual from one to the other.

11. Method according to any one of Claims 1 to 10, wherein reducing the dynamic range of the received detection signal furthermore comprises, after clipping (22) the detection signal but before applying (24) the overall dynamic range compression rate to the clipped detection signal, half-wave rectification (23) of said clipped detection signal.

12. Computer program product comprising instructions that, when the program is executed by a computer, prompt said computer to implement all of the steps of the method according to any one of Claims 1 to 11.

13. Data processing device comprising a processor (24) configured to implement all of the steps of the method according to any one of Claims 1 to 11.

14. Portable metal detection apparatus (10) comprising a device according to Claim 13 and at least one electroacoustic transducer (13) for rendering the audio signal to a user of the apparatus.