Patent Application
20070121976
The application relates to a hearing aid, especially a hearing aid with multiple modes of operation and efficient switching between these modes of operation.
It is well known in the art of hearing aids to have multiple modes of operation, such as different directional characteristics and different methods of noise suppression. The purpose of the different modes of operation is to suit the wearer's needs as good as possible in different environments and situations. As an example, directionality makes it possible for a hearing aid to “focus” on a sound source, such as a speaking person, located e.g. in front of the wearer of the hearing aid. Thereby, possibly disturbing sounds coming from the sides and back of the wearer will have a limited influence on the wearer's perception of sound from a source in front of him/her.
Since it is quite inconvenient for a hearing impaired person to manually switch between different modes of operation in dependence of his/her needs, several attempts have been made to develop efficient methods for automatically selecting “the best” mode of operation as the listening environment or conditions changes. See e.g. published international patent application WO 01/76321 for an example of this.
Another example of such an automatic selection algorithm was presented by Starkey Laboratories, Inc. at the Union der Hörgeräte-Akustiker (UHA) fair in Nürnberg (Germany) in October, 2003. It was stated that “the omni-directional mode is activated in environments with good signal to noise ratio and the directional mode is activated in environments with poor signal to noise ratio”.
Recently, research has shown that hearing aid users prefer a directional mode of operation over an omni-directional mode primarily when they find themselves in a noisy environment and there is a speech source in front of them at a fairly close range, e.g. 2-3 meters. These research results was presented on 1 Feb. 2003 by Brian E. Walden, Ph.D., Director of Research, Army Audiology & Speech Center, Walter Reed Army Medical Center at The First Annual Southern California Conference on Speech, Language, And Hearing Sciences under the title: “Predicting Hearing Aid Microphone Preference in Everyday Listening”.
It is an object to provide an efficient and user friendly system and method for switching between modes of operation in a hearing aid.
Thus, in a first aspect, a hearing aid is provided that comprises means for performing a first mode of operation, means for performing a second mode of operation, first and second quality measuring means for providing first and second quality measures corresponding to the first and second modes of operation, respectively, and selecting means for automatically selecting one of said first and second modes of operation by applying a predetermined selection criterion to the first and second quality measures.
Preferably, the first mode of operation is an omni-directional mode of operation and the second mode of operation is a directional mode of operation. This means that in the first mode of operation, a wearer of the hearing aid will hear sound originating from all directions equally well, more or less, while in the directional mode, the wearer will hear sound coming from a certain direction, such as straight in front of him/her, much better than sound coming from other directions. Thus, the means for performing the first mode of operation may comprise an omni-directional microphone and the means for performing the second mode of operation may comprise a directional microphone. The latter may be a separate directional microphone unit or be composed of the omni-directional microphone in combination with a second omni-directional microphone.
Preferably, the first and second quality measuring means may comprise means for measuring first and second signal-to-noise ratios, respectively. Thereby, a signal-to-noise ratio is provided for each of the two modes of operation
Alternatively, the first and second quality measuring means may comprise means for measuring first and second modulation depths, respectively. The modulation depth could be measured by continuously tracking maximum and minimum in the signal amplitudes in each of the two modes of operation. The modulation depth is to be understood as the difference between the maximum and minimum amplitude when measuring over suitable time intervals, such as intervals with a length of 1 ms, such as 5 ms, such as 20 ms, such as 50 ms, such as 0.1 s, such as 0.25 s, such as 0.5 s, such as 1 s. In common situations where the noise level is substantially constant over the time interval, the modulation depth will thus provide an indication of the signal strength of faster varying signals, such as speech, relative to the static or quasi-static noise level.
Yet alternatively, the first and second quality measuring means may comprise first and second speech detectors, respectively, for providing respective first and second probabilities of speech being present in the respective modes of operation. Such speech detectors are well known in the art. The probability of speech thus provides a measure of how “speech-like” a signal is according to the speech detector. Thus, if the first mode of operation is an omni-directional mode and the second mode of operation is a directional mode, and the wearer of the hearing aid is facing a speaking person while a lot of noise is coming from a location e.g. behind the wearer, the probability of speech provided by the first speech detector will be significantly lower than the probability provided by the second speech detector.
Preferably, the selecting means is programmed to select the mode of operation for which the estimated quality measure is largest when the difference between the first and second quality measures exceeds a predetermined threshold value. The threshold ensures that the selecting means does not switch between the modes of operation until a certain improvement in the quality measure is achieved by doing so. The threshold value may be a difference in signal-to-noise ratio or modulation depth of 1 dB, such as 1.5 dB, such as 2 dB, such as 2.5 dB, such as 3 dB, such as 3.5 dB, such as 4 dB, such as 4.5 dB, such as 5 dB, such as 6 dB, such as 8 dB, such as 10 dB, such as 12 dB. Alternatively, the threshold value may be an absolute difference in speech probability of 0.1, such as 0.15, such as 0.2, such as 0.25, or a relative difference of 5%, such as 10%, such as 15%. Furthermore, the threshold value may be dependent of the absolute level of the quality measures in the sense that the threshold may be larger at high quality measure levels than at low quality measure levels. Thus, if the quality measure of the first mode of operation is high, such as a signal-to-noise ratio of 9 dB, the threshold value may be 6 dB, while in the case of a low quality measure of the first mode of operation, such as 2 dB, the threshold value may be only 1.5 dB. This way, the second mode of operation will only be selected when it provides a substantial improvement in the quality measure.
Preferably, the selecting means is programmed to stay in a selected mode of operation for at least a predetermined minimum period of time. This ensures that the selecting means will not switch back and forth when the difference between the two quality measures is just around the threshold value. The minimum period of time may be as long as 1 ms, such as 5 ms, such as 20 ms, such as 50 ms, such as 0.1 s, such as 0.25 s, such as 0.5 s, such as 1 s, such as 2 s, such as 5 s, such as 10 s.
Preferably, the first quality measuring means comprises a noise detector for providing a noise level, the second quality measuring means comprises a speech detector for estimating a speech probability, and the selecting means is programmed to select the second mode of operation when the noise level exceeds a predetermined noise level threshold and said speech probability exceeds a predetermined probability threshold. Otherwise, the first mode of operation will be selected.
The noise level and probability thresholds may be set to standard values based e.g. on experience of audiologists, but may also be set individually in accordance with feedback from the wearer of the hearing aid.
Furthermore, the first and second quality measuring means may be arranged to operate simultaneously and independent of one another. This assures that the hearing aid may always operate in the currently most user-preferable mode.
In a second aspect, a method for controlling the operation of a hearing aid having at least a first and a second mode of operation is provided. The method comprises estimating a first quality measure relating to the first mode of operation, estimating a second quality measure relating to the second mode of operation, and selecting a mode of operation based on at least said first and second quality measures.
Preferably, the first mode of operation is an omni-directional mode of operation and the second mode of operation is a directional mode of operation.
The first and second quality measures may comprise first and second signal-to-noise ratios, respectively. Alternatively, the first and second quality measures may comprise first and second modulation depths, respectively. As a further alternative, the first and second quality measures may comprise first and second speech probabilities, respectively, provided by respective first and second speech detectors.
Preferably, the step of selecting is further based on a difference between the first and second quality measures. The step of selecting preferably results in selection of the mode of operation for which the estimated quality measure is largest when the difference between the first and second quality measures exceeds a predetermined threshold value.
Furthermore, the step of selecting may be constrained to stay in a selected mode for at least a predetermined minimum period of time.
Preferably, the first quality measure is a noise level provided by a noise detector, the second quality measure is a speech probability provided by a speech detector, and the step of selecting results in the second mode of operation when said noise level exceeds a predetermined noise level threshold and said speech probability exceeds a predetermined speech probability threshold.
In a third aspect, a program is provided for performing the method of the second aspect in a digital signal processing unit of a hearing aid.
In a fourth aspect, a computer readable data carrier is provided that is loaded with a program according to the third aspect.
The embodiments will now be described in further detail with reference to the accompanying drawings.
In
The purpose of the quality estimators 41 and 51 is to provide comparable measures relating to the quality of performance that a hearing impaired wearer of the hearing aid will experience in both of the two modes. The estimators can be implemented in a number of ways. A first possibility is to measure the modulation depth in the two modes. In order to do this, the quality estimators 41 and 51 should continuously track maximum and minimum in their respective signal amplitudes. The modulation depth is then provided by calculating the difference between the maximum and minimum amplitude in suitable time intervals, such as intervals with a length of 1 ms, such as 5 ms, such as 20 ms, such as 50 ms, such as 0.1 s, such as 0.25 s, such as 0.5 s, such as 1 s. In many situations, the noise level will appear to be substantially constant over the time interval, and the modulation depth will thus provide an indication of the signal strength of faster varying signals, such as speech, relative to the noise level.
Another possible implementation of the quality estimators 41 and 51 comprises a speech detector in each quality estimator 41, 51. The output from the speech detectors is probabilities of speech being present in the respective mode of operation, i.e. a numerical value between 0 and 1 indicating how “speech-like” the signal is according to the speech detector. Thus, if the wearer of the hearing aid is facing a speaking person while a lot of noise is coming from another location, e.g. behind the wearer, the probability of speech provided by the speech detector in 41 will be significantly lower than the probability provided by the speech detector in 51, i.e. Q(omni)<Q(dir). This corresponds to the observation that the wearer of a hearing aid in this situation will usually prefer the directional mode in order to better hear and understand the speaking person.
The selector 61 automatically selects the currently most preferable mode of operation in accordance with a predetermined (possibly user defined) criterion. This can be done in a number of ways.
The threshold value k ensures that the selector 61 does not switch to the directional mode until a certain improvement in the quality will be experienced by a user.
The threshold value k may be a difference in modulation depth of 1 dB, such as 1.5 dB, such as 2 dB, such as 2.5 dB, such as 3 dB, such as 3.5 dB, such as 4 dB, such as 4.5 dB, such as 5 dB, such as 6 dB, such as 8 dB, such as 10 dB, such as 12 dB. Alternatively, the threshold value k may be an absolute difference in speech probability of 0.1, such as 0.15, such as 0.2, such as 0.25, or a relative difference of 5%, such as 10%, such as 15%. Furthermore, the threshold value k may be dependent of the absolute level of the quality measures in the sense that the threshold may be larger at high quality measure levels than at low quality measure levels. Thus, if the quality measure of the omni-directional mode of operation is high, such as a modulation depth of 9 dB, the threshold value may be 6 dB, while in the case of a low quality measure of the omni-directional mode of operation, such as 2 dB, the threshold value may be only 1.5 dB. This way, the directional mode of operation will only be selected when it provides a substantial improvement in the quality measure.
In
In order to avoid rapid shifting back and forth between the omni-directional and directional modes, the selector may be programmed to stay in a selected mode of operation for at least a predetermined minimum period of time. This ensures that the selecting means will not switch back and forth when the difference between the two quality measures is just around the threshold value. The minimum period of time may be 1 ms, such as 5 ms, such as 20 ms, such as 50 ms, such as 0.1 s, such as 0.25 s, such as 0.5 s, such as 1 s, such as 2 s, such as 5 s, such as 10 s.
To best suit the individual wearer's needs, it may be beneficial to involve the user in setting the threshold values k and k′, as well as the above mentioned minimum period of time.
| Number | Date | Country | Kind |
|---|---|---|---|
| PA 2004 00349 | Mar 2004 | DK | national |
| PCT/DK05/00139 | Feb 2005 | DK | national |
This application claims priority from PCT Application No. PCT/DK2005/000139 which was filed on 28 Feb. 2005 which claims priority from Danish Patent Application No. PA 2004 00349 filed on 1 Mar. 2004, the disclosures of both of these applications are expressly incorporated by reference in their entirety herein.
