Abstract
A method operates a hearing device which performs active noise suppression for suppressing noise signals having one or more frequency components. An audiogram is provided which specifies a hearing threshold of a user of the hearing device as a function of frequency, wherein by using the audiogram it is determined which frequency components of the noise are audible to the user and which are not audible. The noise suppression is operated selectively by suppressing audible frequency components of the noise and by not suppressing inaudible frequency components of the noise. A corresponding hearing device is operated according to the method.
Claims
1. A method for operating a hearing device having active noise cancelling for suppression of noise signals having at least one frequency component, which comprises the step of: providing an audiogram specifying a hearing threshold of a user of the hearing device in dependence on frequency, the audiogram is used to determine which frequency components of noise are audible to the user and which are not audible, the audiogram having at least one dead region within which the hearing threshold is above a minimum level; and operating noise suppression selectively dependent on the audiogram by suppressing audible frequency components of the noise and by not suppressing inaudible frequency components of the noise, an operation of the noise suppression is frequency-selective, by not suppressing the frequency components which are disposed within the at least one dead region of the audiogram, so that only the frequency components which are not within the at least one dead region of the audiogram are actively suppressed.
2. The method according to claim 1, wherein an operation of the noise suppression is amplitude-selective, by not suppressing the frequency components which have a level below the hearing threshold, so that only the frequency components in which the level is above the hearing threshold are actively suppressed.
3. The method according to claim 2, which further comprises: defining a maximum level which specifies a power limit of the hearing device; and not suppressing the frequency components having the level which is above the maximum level.
4. The method according to claim 1, wherein a local maximum of the hearing threshold is disposed within said at least one dead region.
5. The method according to claim 1, wherein the noise suppression suppresses intrusive ambient noise by recording the intrusive ambient noise with an external microphone of the hearing device and outputting it in inverted form via a receiver of the hearing device.
6. The method according to claim 1, wherein the noise suppression has an active occlusion reduction which suppresses intrusive noise arising from an occlusion in a user's auditory canal, by recording the intrusive noise with an internal microphone of the hearing device in the user's auditory canal and outputting it in inverted form via a receiver of the hearing device.
7. The method according to claim 1, wherein the audiogram specifies the hearing threshold in a frequency range from at least 10 Hz to at most 20 kHz.
8. The method according to claim 1, wherein a frequency range for speech is not suppressed by the active noise cancelling.
9. A hearing device, comprising: a controller configured to perform a method of operating the hearing device having active noise cancelling for suppression of noise signals having at least one frequency component, the method comprises the step of: providing an audiogram specifying a hearing threshold of a user of the hearing device in dependence on frequency, the audiogram is used to determine which frequency components of noise are audible to the user and which are not audible, the audiogram having at least one dead region within which the hearing threshold is above a minimum level; and operating noise suppression selectively dependent on the audiogram by suppressing audible frequency components of the noise and by not suppressing inaudible frequency components of the noise, an operation of the noise suppression is frequency-selective, by not suppressing the frequency components which are disposed within the at least one dead region of the audiogram, so that only the frequency components which are not within the at least one dead region of the audiogram are actively suppressed.
10. The hearing device according to claim 9, wherein said controller has a signal processor for modifying input signals to compensate for a hearing impairment of the user.
Description
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING
(1) FIG. 1 is an illustration of a hearing device;
(2) FIG. 2 is a graph showing an audiogram and an amplitude-selective suppression of sounds;
(3) FIG. 3 is a graph showing the audiogram of FIG. 2 and a frequency-selective suppression of sounds; and
(4) FIG. 4 is a graph showing the audiogram of FIG. 2 and an amplitude- and frequency-selective suppression of sounds.
DETAILED DESCRIPTION OF THE INVENTION
(5) Referring now to the figures of the drawings in detail and first, particularly to FIG. 1 thereof, there is shown an exemplary embodiment of a hearing device 2. In FIGS. 2 to 4, an example of an audiogram 4 of a user is shown, on the basis of which an active noise cancellation 6 of the hearing device 2 is selectively operated in different ways as part of a method for operating the hearing device 2. The active noise cancellation 6 is generally used to suppress noise, where the term “noise” is also intended to mean individual sounds, without limitation of generality. The noise cancellation 6 suppresses noise in such a way that a quieter listening situation is created for the user. This is carried out by generating counter-noise in order to partially or even completely eliminate the noise signals. This requires energy, which in this case is extracted from an energy storage device 8 of the hearing device 2.
(6) The hearing device 2 in FIG. 1 has a control unit 10, which is configured to carry out the method. The control unit 10 is arranged inside a housing 12 of the hearing device 2. The audiogram 4 in this case is stored in a memory 14. The memory 14 and the noise cancellation 6 are each part of the control unit 10. This is not mandatory, however.
(7) The hearing device 2 shown is configured as a hearing aid device to compensate for a hearing deficit of the user, and for this purpose has a signal processing unit 15, which is also part of the control unit 10. The signal processing unit 15 is used to modify input signals to compensate for the user's hearing deficit. The input signals are detected by means of a microphone 16 of the hearing aid 2, in FIG. 1 two external microphones 16 are shown. The modification of the input signals takes place in the signal processing 15 according to the audiogram 4, i.e. user-specifically. The modified input signals are then output signals of the signal processing 15 and are forwarded to a receiver 18 of the hearing device 2 for output to the user. The receiver 18, in the embodiment shown, is a part of an earpiece 20, which is inserted into the user's ear canal. Alternatively, the receiver 18 is arranged in the housing 12 and the sound signals generated by the receiver 18 are routed into the ear canal via a sound tube. Alternatively or additionally, the input signals are electrical audio signals, which are transferred to the hearing device 2 from a suitable playback device or are stored in the hearing device 2.
(8) The input signals are divided into several frequency bands by means of a filter bank, not shown in detail, as part of the signal processing 15 of the hearing device 2, i.e. in the present case within the control unit 10. For example, the filter bank has 48 channels and generates 48 frequency bands accordingly. A given frequency component f1-f8 is then suppressed by suppressing the particular frequency band in which the frequency component f1-f8 to be suppressed is located.
(9) FIG. 1 shows the hearing device with only one single device. In a variant not shown, the hearing device 2 has a binaural design and contains two individual devices, e.g. as in FIG. 1, one for each of the two ears of the user.
(10) The audiogram 4 generally indicates a frequency-dependent hearing threshold 22 of the user and is determined, for example, in a corresponding test or calibration procedure. The audiogram 4 typically differs from user to user. The audiogram 4 shown in FIGS. 2 to 4 is therefore only one example from a plurality of possible audiograms 4. The audiogram 4 shown indicates, for each frequency f of a frequency spectrum from 10 Hz to 20 kHz, a hearing threshold 22 above which the respective frequency f is audible to the user, i.e. the user-specific hearing threshold 22 is specified as a function of frequency. The hearing threshold 22 is a level p, i.e. an amplitude. In FIGS. 2 to 4, a number of vertical arrows also show various frequency components f1-f8, each of which has a specific level p. The frequency components f1-f8 shown as examples are here individual frequencies, but are alternatively frequency ranges with a plurality of frequencies. The hearing thresholds 22 of the various frequencies f together form a hearing curve H. As will be evident from the graphical representations of FIGS. 2 to 4, the hearing curve H divides the space defined by the two dimensions of level p and frequency f into two regions, namely an actually inaudible region nB underneath the hearing curve H and an actually audible region hB above the hearing curve H.
(11) The audiogram 4 is thus formed in such a way that it can be used to determine which sounds are audible to the user and which are not audible. A particular noise consists of one or more frequency components f1-f8, which are audible or inaudible, or a combination of these. A frequency component f1-f8 is audible to the user if and only if this frequency component f1-f8 has a level p which exceeds the hearing threshold 22 of the user for this frequency range. Thus, in FIG. 2 the frequency components f1-f5 are actually audible by the user, while the frequency components f6-f8 are not. In FIG. 3 the frequency components f1, f2, f5 are actually audible by the user, while the frequency components f4, f6 are not. In FIG. 4, the frequency components f1-f3 are actually audible by the user, whereas the frequency components f4-f6 are not.
(12) The noise cancellation 6 is further operated selectively by suppressing audible frequency components f1-f8 of the noise and by not suppressing inaudible frequency components f1-f8 of the noise. The noise suppression 6 is therefore selectively used only for such frequency components f1-f8 for which their suppression also has sufficient benefit for the user. Such frequency components f1-f8, which the user cannot hear at all, do not need to be actively suppressed and are therefore ignored during the suppression. It is not only those frequency components f1-f8 which are already outside the acoustic spectrum, i.e. in FIGS. 2 and 4 below 20 Hz and above 20 kHz, that are not suppressed, but the individual audiogram 4 of the user is used to perform the suppression on an individual basis. Thus in the exemplary embodiment shown, the user is hearing-impaired and has a hearing deficit under which the hearing threshold 22 in the range from 1 kHz to 2 kHz is at least approximately 100 dB. Sounds at these frequencies and below this hearing threshold 22 are then not perceptible to the user, i.e. are not audible, and are therefore not actively suppressed.
(13) Alternatively, the user is not hearing-impaired in the sense of a pathological condition. The noise cancellation 6 is generally a personalized noise cancellation 6. In this respect, the method is not only suitable for hearing devices 2 which are designed as hearing aids, for example as in FIG. 1, but also for headphones, headsets and the like, which in themselves mainly output useful sounds to the user, but these useful sounds are superimposed by other sounds. These other sounds are then suppressed in a user-specific manner by means of the noise cancellation 6.
(14) Which frequency components f1-f8 of the sounds are audible to the user and which are not audible is determined by the audiogram 4. More specifically, it is determined on the basis of the audiogram 4 which frequency components f1-f8 are audible or cannot be perceived. The sounds are thus divided into audible and non-audible frequency components f1-f8 based on the known audiogram 4. Whether a frequency component f1-f8 of the audiogram 4 is determined as audible or inaudible can therefore differ in principle from whether it is actually audible or not, depending on the manner of operation of the selective noise cancellation 6. In general, however, the aim is to operate the noise cancellation in a selective manner such that the frequency component f1-f8 will be correctly identified as audible or inaudible with an overwhelming probability by applying a definition based on the audiogram 4.
(15) In this case, two variants are particularly suitable for distinguishing audible and inaudible frequency components f1-f8 and thus for implementing a selective noise cancellation 6. On the basis of FIG. 2 an embodiment of the first variant is described, on the basis of FIG. 3 an embodiment of the second variant is described, and in the embodiment according to FIG. 4 both variants are combined with each other. In FIG. 2-4, a number of frequency components f1-f8 which form one or more sounds are also shown as examples. The frequency components f1-f8 shown represent the actually existing sounds, i.e. not the sounds that are output to the user via the receiver 18. These actual sounds normally enter the user's ear canal directly, but are sometimes further attenuated due to the earpiece 18. The actual sounds in this case also reach the microphone 16, are picked up by it, processed in the control unit 10 as appropriate, and output to the user via the receiver 18.
(16) In FIG. 2 in accordance with the first variant the noise cancellation 6 is operated amplitude-selectively, by not suppressing those frequency components f6-f8 which have a level p below the corresponding hearing threshold 22, so that only those frequency components f1-f5 in which the level p is above the corresponding hearing threshold 22 are actively suppressed. For this purpose, the respective level p of a frequency component f1-f8 is compared with the corresponding hearing threshold 22 of the audiogram 4, and those frequency components f1-f5 which have a level p above the hearing threshold 22 are considered as audible frequency components f1-f5, whereas those frequency components f6-f8 which have a level p below the hearing threshold 22 are considered as inaudible frequency components f6-f8. A distinction is thus made according to the level p, i.e. the amplitude of the frequency components f1-f8 relative to the audiogram 4, more precisely relative to the hearing curve H. As a result, during the method the noise is actively suppressed above the hearing curve H and not unnecessarily so below it.
(17) In addition, in the example of FIG. 2 a maximum level 24 is defined which specifies a power limit of the hearing device 2, and those frequency components f4, f5, the level p of which is above the maximum level 24, are not suppressed. The maximum level 24 indicates the level p above which suppression of the respective frequency component f1-f8 is no longer meaningful or no longer possible due to technical limitations of the hearing device 2. Such technical limitations are the result, for example, of a maximum power of the receiver 18 or a power amplifier stage of the hearing device 2. Since above the maximum level 24 an effective suppression cannot therefore be carried out with the hearing device 2, but instead arises automatically due to the power limit being exceeded, suppression is not used in this case and the frequency components f4, f5 are excluded from the noise cancellation 6, although in this case they are audible. However, at the output these frequency components f4, f5 are automatically reduced to the maximum level 24 due to the power limit. As is evident from FIG. 2, the maximum level 24 is normally above the respective hearing threshold 22. This is not essential, however. In this case, the maximum level 24 is constant for all frequencies f, whereas in a variant not shown, the maximum level 24 is frequency-dependent. The use of a maximum level 24 as described is independent of the amplitude-selective noise cancellation 6 described and can also be omitted.
(18) In FIG. 3, according to the second variant, the noise cancellation 6 is operated frequency-selectively. Here the audiogram 4 additionally has one or more dead regions 26, within which the hearing threshold 22 is above a minimum level 28 in each case. The frequency-selective operation is implemented in such a manner that those frequency components f4 which are located within a dead region 26 of the audiogram 4 are not suppressed, so that only those frequency components f1-f3, f5, f6 which are not within a dead region 26 of the audiogram 4 are actively suppressed. A respective dead region 26 thus characterizes a frequency range in which the user's hearing is particularly poor. In the dead regions 26, the noise is generally not suppressed independently of the level p, i.e. regardless of whether the level p is above or below the hearing threshold 22. Any frequency components f4 which fall within a dead region 26 are considered inaudible and hence are not suppressed. Frequency components f1-f3, f4, f5, which are located outside all dead regions 26, are considered to be audible and are actively suppressed.
(19) In general, a dead region 26 of the audiogram 4 extends from a lower frequency up to an upper frequency. Between these two frequencies, the hearing threshold 22 is consistently above the minimum level 28. FIG. 3 shows a total of three dead regions 26, wherein the two outer dead regions 26 are at the edge of the acoustic spectrum and are purely natural dead regions 26, i.e. dead regions 26 in the general sense and therefore not necessarily due to a hearing deficit. The middle dead region 26, on the other hand, is a dead region of hearing loss, i.e. due to a hearing deficit of the user, and is therefore a dead region 26 in the specific sense. While general dead regions 26 are located at the edge and the hearing curve H tapers off there, so to speak, towards high levels p, in contrast to this a specific dead region 26 can have a local maximum 30 of the hearing threshold 22 and can also frame the local maximum 30, as it were, as is the case in FIG. 3 for the middle dead region 26.
(20) Based on FIG. 4, it will now be clear that the amplitude-selective and frequency-selective operation of the noise cancellation 6 can also be combined. By overlapping these two concepts, one or more active regions 32 are then formed in the audiogram in such a way that only those frequency components f1-f3 are suppressed which are both outside the dead regions 26 and also above the respective hearing threshold 22, whereas the other frequency ranges f4-f6 are not actively suppressed, since these are not perceived by the user in any case. As can be seen from FIG. 4, the active regions 32 result as an intersection of the audible range hB and the dead regions 26.
(21) In FIGS. 2-4, the audiogram 4 indicates the hearing threshold 22 in a frequency range from 10 Hz to 20 kHz, i.e. it contains an overall frequency spectrum corresponding to the acoustic spectrum. At the edges of the audiogram, i.e. in particular below 20 Hz and above 16 kHz, as already indicated the hearing ability of most people is normally poor, regardless of whether they are hearing-impaired or not. The hearing threshold 22 here is typically above 90 dB, so that natural dead regions 26 are produced here. In addition, it makes sense to exclude such frequency ranges, in which mostly useful signals are to be expected, from the noise cancelling 6 from the outset, provided that these useful signals are not already isolated by the hearing device 2 and further processed separately. In a variant not explicitly shown, for example, a frequency range for speech similar to the dead regions 26 is not suppressed by the noise cancellation 6, regardless of whether the user has good or poor hearing there. Speech normally constitutes a useful signal, which is therefore preferably not cancelled by the noise suppression if at all possible. A suitable frequency range for speech ranges from 300 Hz to 5 kHz or over a partial range thereof.
(22) In the exemplary embodiment shown in FIG. 1, the active noise suppression 6 has active noise cancelling (ANC for short), more precisely, is implemented as such. Accordingly the noise suppression 6 suppresses intrusive ambient noise by recording the intrusive noise with one or both of the external microphones 16 of the hearing device 2 and outputting it in inverted form via the receiver 18 of the hearing device 2.
(23) In a variant not shown, the active noise cancelling 6 has an active occlusion reduction (AOR for short) or is implemented as such a system, and suppresses intrusive noise arising from an occlusion in the user's auditory canal by recording the noise interference with an internal microphone 34 of the hearing device 2 in the user's auditory canal and outputting it in inverted form via the receiver 18 of the hearing device 2. FIG. 1 shows an internal microphone 34 as part of the earpiece 18. Without AOR, the internal microphone 34 is purely optional.
LIST OF REFERENCE SIGNS
(24) 2 hearing aid 4 audiogram 6 noise suppression 8 energy store 10 control unit 12 housing 14 memory 15 signal processor 16 external microphone 18 receiver 20 earpiece 22 hearing threshold 24 maximum level 26 dead region 28 minimum level 30 local maximum 32 active region 34 internal microphone f frequency f1-f8 frequency component H hearing curve hB actual audible range NB actual inaudible range p level
