CN114257913A

CN114257913A - In-ear earphone

Info

Publication number: CN114257913A
Application number: CN202111122714.6A
Authority: CN
Inventors: 黄顾翠玉; 阿图尔·佐佐
Original assignee: Devialet SA
Current assignee: Devialet SA
Priority date: 2020-09-25
Filing date: 2021-09-24
Publication date: 2022-03-29
Also published as: FR3114722B1; EP3975583B1; FR3114722A1; EP3975583A1

Abstract

The invention relates to an in-ear headphone (10) comprising an electroacoustic transducer (18) adapted to propagate sound, located in an anatomical housing (14), a control unit (20) adapted to receive a signal to be reproduced and to control the electroacoustic transducer (18), the control unit comprising an equalization filter applied to the signal to be reproduced, the control unit (20) being connected to an internal microphone (30) of the headphone. The control unit (20) further comprises means for estimating the infrasound path, and means for defining and successively applying equalization filters based on the estimation of the infrasound path. The headset includes an external microphone (32). The infrasound path estimating means is adapted to estimate the infrasound path also from the acoustic signal measured by the external microphone (32).

Description

In-ear earphone

Cross Reference to Related Applications

Priority and benefit of the application filed on french patent office patent application No. 2009767, on 25/9/2020, is hereby incorporated by reference in its entirety.

Technical Field

The present invention relates to an in-ear headphone, generally comprising an electroacoustic transducer located in an anatomical housing adapted to propagate sound in an ear canal, an electroacoustic transducer control unit adapted to receive a signal to be reproduced, the control unit comprising an equalization filter applied to the signal to be reproduced, the control unit being connected to an internal microphone of the headphone, the internal microphone being configured for measuring an audio signal in the ear canal, the control unit further comprising continuous estimation means defined as an infrasound path between the transducer and the internal microphone, and means for defining and continuously applying the equalization filter in dependence on the estimation of the infrasound path.

Background

The in-ear headphone has an anatomical shape and is insertable into the ear canal. Which defines a small volume of air substantially enclosed between the earphone and the tympanic membrane, so that relatively low frequency sounds, down to 5Hz, can be produced. However, in view of the large difference in ear size, a completely closed space is sometimes not obtained between the earphone and the tympanic membrane, and leakage occurs between the ear canal and the earphone.

These leaks greatly change the frequency response of the headset to low frequencies.

This phenomenon is known and signal processing schemes have been proposed whereby these leaks can be compensated for by correcting the signal to be reproduced by the earphone electro-acoustic transducer.

This solution is described, for example, in US 9,264,823B 2. In this document, it is proposed to process the signal coming from an equalization filter constituted by a shelf filter referred to by the english "folding filter" in the document. The gain of such a filter varies continuously with the signal of the microphone measuring the sound waves in the ear canal. The gain of the filter is adjusted based on the ratio between the signal energy produced by the electromagnetic transducer and the audio signal energy measured by the internal microphone over different frequency ranges.

Disclosure of Invention

This solution only partially compensates for the leakage problem of the in-ear headphone in the low frequency range.

It is an object of the invention to propose a solution that allows a better correction of the audio signal to take into account leakage between the in-ear headphone and the ear canal.

To this end, the present invention has for its object to provide an in-ear headphone of the above-mentioned type, characterized in that the in-ear headphone comprises an external microphone configured for measuring acoustic signals outside the ear canal, and in that the infrasound path estimating means are adapted for estimating the infrasound path also from acoustic signals outside the ear canal measured by the external microphone.

According to a particular embodiment, the in-ear headphone comprises one or more of the following features:

the control unit comprises means for correcting the acoustic signal measured by the external microphone outside the ear canal according to an attenuation model adapted to the earphone;

-the in-ear headphone comprises means for disregarding acoustic signals outside the ear canal measured by the external microphone when the level of leakage between the in-ear headphone and the ear canal is above a predetermined threshold;

-the defining and applying means are adapted to define the gain, cut-off frequency and quality factor of the filter from the estimate of the infrasound path and to apply it continuously to the equalization filter;

-the infrasound path continuation estimating means comprises an infrasound path transfer function model;

the estimation means comprise a regulation loop to run an optimization algorithm applied to the model;

the infrasound path continuation estimating means includes calculating means for calculating an energy factor composed of a quotient of energy of the frequency signals in the low frequency range based on the energy of the frequency signals in the intermediate frequency range;

the filter definition means comprise calculation means for calculating the gain, the cut-off frequency and the quality factor of the filter by applying a function predefined according to the energy factor;

the in-ear headphone comprises confidence score establishing means, the confidence score being defined in terms of the equalized audio signal to be reproduced applied to the transducer and the error between the audio signal to be reproduced and the signal measured by the internal microphone to which the infrasound path estimate provided by the model is applied; the in-ear headphone further comprises stopping and varying means for filter gain, cut-off frequency and quality factor when the confidence score is below a predetermined threshold.

Drawings

The invention will be better understood from the following description, given by way of example only, taken in conjunction with the accompanying drawings, in which:

FIG. 1 is a schematic view of an in-ear headphone worn by one ear;

FIG. 2 is a functional schematic diagram of an in-ear headphone;

FIG. 3 is a flow chart of the acoustic processing of an in-ear headphone;

fig. 4, 5, 6 are graphs showing the gain, cut-off frequency and quality factor of the filter as a function of the energy factor, respectively.

Detailed Description

A part of the face 8 is shown in dashed lines in fig. 1. The in-ear earphone 10 is inserted into the ear canal (reference numeral 12) of the ear. In fact, the user has a pair of earphones symmetrical to each other and wears them in both ears. Only one is described herein.

As is well known, the in-ear earphone 10 comprises an anatomical housing 14, which is slightly frustoconical in shape, adapted to be inserted into the ear canal 12 as closely as possible.

The end of the circular truncated cone portion forms an opening 16 as a sound outlet, facing which opening an electroacoustic transducer 18 is arranged, which is connected to a control unit 20 for excitation.

The headset comprises receiving means 22 for audio signals to be reproduced from a broadcasting means, such as a mobile phone 24. Communication between the mobile phone and the receiving means 22 of the in-ear headphone is provided by e.g. bluetooth.

The control unit 20 further comprises a calculator 26 adapted to provide information processing described below and amplifying means 28 for amplifying the excitation signal of the transducer 18. The output of the amplifying means 28 is connected to the transducer 18.

The calculator 26 further comprises a first internal microphone 30 arranged at an outer surface of the housing placed in the ear canal 12. The microphone is adapted to measure sound waves in the ear canal, more precisely sound waves acting on the eardrum.

Likewise, the housing 14 at the ear canal 12 comprises an external microphone 32 on its outer surface, which is also connected to the calculator 26 and is adapted to measure acoustic signals outside the ear cavity corresponding to ambient noise.

In fig. 2, the electro-acoustic transducer 18 and the internal microphone 30 are shown. The infrasound path 50 is defined as being between them, and is commonly referred to as a "secondary path" in english. The infrasonic path refers to the sound flow emitted by the transducer 18 and received by the tympanic membrane and is actually measured by the microphone 30, which is set to represent a measurement on the tympanic membrane. The primary acoustic path refers to the reception of the eardrum into the acoustic stream outside the earphone, corresponding to the ambient noise.

In this figure, the output of the receiving means 22 is connected to an equalization filter 52 with variable parameters obtained in the calculator 26. For example, the filter is a Shelving filter or what is known in french as a Shelving filter. As a modification, the equalization filter 52 is an Infinite Impulse Response (IIR) or Finite Impulse Response (FIR) filter, a cascade filter, or any other filter.

The output of the filter 52 for such equalized signal to be reproduced is connected to the amplifying means 28 via a digital/analog converter 54.

The earpiece further comprises continuous estimation means 60 implemented in the calculator 26 for continuously estimating the infrasound path 50, which estimation means are associated with definition and continuous application means 62 for defining and continuously applying the parameters of the equalization filter 52.

More precisely, the estimation means 60 comprise a transfer function model 70 of the infrasound path 50. The input of the model 70 receives the equalized signal to be reproduced from the equalization filter 52.

The output signal of the model 70 is subtracted in a subtractor 72 from the composite signal, in particular from the internal microphone 30.

The analog signal at the output of the internal microphone 30 is routed to an analog-to-digital converter 86, the output of which is connected to the input of the subtractor 72 through an intervening subtractor 88. The negative input of the subtractor 88 is connected to the external microphone 32, in turn, via an analog/digital converter 89 and a passive attenuation modeling module 90 that models the structure of the earphone. A switch 92 controlled by the control unit 20 is interposed between the block 90 and the subtractor 88.

The module 90 is adapted to apply the digitized signal corresponding to the external sound stream to a filter that is attenuated by the corresponding earphone structure such that the output signal of the module 90 represents the sound stream generated by the unique external sound stream within the ear canal when the earphone is present.

The attenuation function of the filter is measured experimentally on the headset and stored in memory.

An output of subtractor 72 is connected to an input of an estimation unit 100 that receives at another input the equalized signal from the output of filter 52.

The estimation unit 100 is adapted to run a least squares algorithm and to connect the model 70 to ensure that it is adjusted according to a suitable optimization algorithm. The evaluation unit 100 thus forms a regulation loop of the model 70.

The filter parameter definition means 62 comprise continuous calculation means 120 for continuously calculating the parameters of the filter 52, in particular the gain with reference G, the cut-off frequency with reference Fc and the quality factor with reference Q.

The algorithm run by the headset calculator 26 is shown in fig. 3.

In addition, the acoustic signal source 22, the filter 52, the internal microphone 30, and the external microphone 32 are shown at the top for ease of understanding.

The infrasound path estimation means 60 generate an impulse response of the infrasound path 50 on the basis of the information received from the equalized signal to be reproduced and from the two

microphones

30, 32.

In step 212, a fast fourier transform is applied to the impulse response obtained by the model 70.

In step 214, a frequency signal energy calculation in the low frequency range is performed. The low frequency range is for example between 50 and 200 hz.

Also, in the same step 214, a calculation of the frequency signal energy in the mid frequency range selected outside the sound frequency range affected by the leakage is performed. The mid-frequency range is, for example, between 300 hz and 500 hz.

In step 216, the calculated energy is smoothed over time, making it more stable against external noise, since it is time-varying.

In step 218, an energy factor EF is continuously defined from the smoothed energy values. Based on the energy of the frequency signal in the intermediate frequency range, which is equal to the energy quotient of the frequency signal in the low frequency range.

In addition, the means 60 generate an error signal, independent of the calculation of the energy factor EF, intended to generate a confidence score SNE as described below.

The error signal corresponds to the signal obtained at the output of the subtractor 72 and is equal to the difference between the equalized signal to be reproduced, to which the infrasound path transfer function modelled by the module 70 is applied on the one hand, and the digitized signal coming from the internal microphone 30, corrected by the digitized signal measured by the external microphone 32 and filtered to take account of the attenuation of the earphone structure on the other hand.

In step 312 the error signal and the equalized signal to be reproduced from the filter 52 are filtered in the intermediate frequency range, for example 50Hz to 750 Hz. And then smoothed in step 314.

In step 316, a confidence score SNE is calculated. Which is equal to the amplitude quotient of the equalized signal to be reproduced, which has been filtered and smoothed by the filtered and smoothed error signal amplitude.

Step 400 includes defining and applying filter parameters based on the energy factor EF and the confidence score SNE.

In case the confidence score SNE is above a predetermined threshold, each of these parameters is calculated from the energy factor EF obtained in step 218 using a predefined polynomial function. If the confidence score SNE is above a predetermined threshold, the new parameters of the filter are applied to the filter 52 and the algorithm is run again. Otherwise, the algorithm is run again without changing the parameters.

In fig. 4, 5, 6 polynomial functions are shown which lead to a gain G (fig. 4), a cut-off frequency Fc (fig. 5) and a quality factor Q (fig. 6) in dependence on the energy factor EF, respectively.

These curves were established by successive experiments based on tests performed on the earphone.

Throughout the implementation of the algorithm, the switch 72 remains closed under the control of the control unit 20. The latter ensures that the switch is only temporarily opened when the energy factor EF is below a predetermined threshold. And therefore the level of leakage is too high, the acoustic signal corresponding to the external flow measured by the external microphone 32 is not considered.

It will be appreciated that the filter is in an improved way stably adapted to any leakage that may occur around the headset, since the acoustic signal of the external microphone 32 is taken into account.

Claims

1. An in-ear headphone (10), the in-ear headphone comprising: an electroacoustic transducer (18) located in the anatomical shell (14) and adapted to propagate sound in the ear canal (12); a control unit (20) adapted to control the electroacoustic transducer (18) for receiving a signal to be reproduced, the control unit comprising an equalization filter (52) for application to the signal to be reproduced, the control unit (20) being connected to an internal microphone (30) of the earpiece, the internal microphone being configured for measuring an audio signal in the ear canal (12), the control unit (20) further comprising continuous estimation means (60) for continuously estimating an infrasound path (50) defined between the electroacoustic transducer (18) and the internal microphone (30), and means (62) for defining and continuously applying the equalization filter (52) depending on the estimation of the infrasound path (50),

characterized in that the earphone comprises an external microphone (32) configured for measuring acoustic signals outside the ear canal (12), and the estimating means (60) of the infrasound path is adapted to estimate the infrasound path (50) also from acoustic signals outside the ear canal (12) measured by the external microphone (32).

2. An in-ear headphone (10) as claimed in claim 1, characterized in that the control unit (20) comprises means (90) for correcting acoustic signals measured by the external microphone (32) outside the ear canal (12) according to an attenuation model (90) adapted to the headphone.

3. An in-ear headphone (10) as claimed in any one of the preceding claims, characterized in that the in-ear headphone comprises means (92) for disregarding acoustic signals outside the ear canal (12) measured by the external microphone (32) when a leakage level between the in-ear headphone (10) and the ear canal (12) is above a predetermined threshold.

4. In-ear headphone (10) as claimed in one of the preceding claims, characterized in that the definition and application means (62) are adapted to define the gain, the cut-off frequency and the quality factor of the filter from the estimate of the infrasound path (50) and to apply them continuously to the equalization filter (52).

5. In-ear headphone (10) as claimed in one of the preceding claims, characterized in that the continuous estimation means (60) of the infrasound path (50) comprise a transfer function model (70) of the infrasound path (50).

6. In-ear headphone (10) as claimed in claim 5, characterized in that the estimation means (60) comprise a regulation loop (100) running an optimization algorithm applied to the model (70).

7. In-ear headphone (10) as claimed in claim 5 or 6, characterized in that the continuous estimation means (60) of the infrasound path (50) comprise calculation means (218) for calculating an energy Factor (FE) consisting of a quotient of the frequency signal energies in a low frequency range on the basis of the frequency signal energies in a mid frequency range.

8. An in-ear headphone (10) as claimed in claim 7, characterized in that the filter definition means (62) comprise calculation means (400) for calculating the gain (G), the cut-off frequency (Fc) and the quality factor (Q) of the filter by applying a function predefined according to the Energy Factor (EF).

9. An in-ear headphone (10) as claimed in any one of the claims 5 to 8, characterized in that the in-ear headphone comprises confidence score establishing means (312, 314, 316) defined in accordance with an equalized audio signal to be reproduced applied to the electroacoustic transducer (18) and an error between the audio signal to be reproduced to which an infrasound path estimate provided by the model (70) is applied and a signal measured by the internal microphone (30), the in-ear headphone further comprising stopping and varying means (400) of the gain (G), cut-off frequency (Fc) and quality factor (Q) of the filter when the confidence score is below a predetermined threshold.