Patent Application
20140169570
This invention pertains to devices and methods for treating hearing disorders and, in particular, to electronic hearing aids.
Hearing aids are electronic instruments worn in or around the ear that compensate for hearing losses by amplifying and processing sound so as to help people with hearing loss hear better in both quiet and noisy situations. Hearing aid wearers often complain of a diminished ability to perceive and appreciate the richness of live music. Their diminished experience is due (at least in part) to the inability to perceive the binaural cues that convey the spatial aspects of the live music experience to listeners with normal hearing. It has also long been recognized that listeners prefer music that appears to emanate from a broad spatial extent over that emanating from a narrow point source. Stereo and surround sound consumer audio formats recognize this preference, and correspondingly generate spacious audio experiences for listeners. Concert hall architects also recognize this preference and design halls to enhance the spaciousness of a musical performance. Listeners with hearing loss, especially those whose impairment is moderate-severe to severe, have deficits in the perception of the binaural cues that convey spaciousness. Indeed, even listeners with milder hearing losses can have such deficits.
Designers of concert halls achieve a sense of spaciousness or envelopment by ensuring that there are significant reflections of the direct sound coming from the lateral walls. These lateral reflections cause a sense of spaciousness by de-correlating the signals at the two ears. Intuitively, the sense of spaciousness comes from the de-correlated signals giving an impression that the same sound is arriving simultaneously from multiple locations. Indeed, inter-aural de-correlation is manifested in the brain as random fluctuations of the binaural disparity cues that underlie the perceived lateral angle of a sound source. The perceptual effect is that of an auditory image that has a broad spatial extant.
Described herein are techniques for artificially enhancing spaciousness in a hearing aid to improve the music listening experience. Such spatial enhancement is produced by doing signal processing in the hearing aid that mimics the acoustic effects of well-designed concert halls. Although the primary objective is to improve the experience of a live music performance, the same techniques can also be applied to improving the experience of listening to recorded music reproduced and amplified over a speaker system, or to music streamed to the direct-audio input of a hearing aid. Spatial enhancement may also be applied to speech listening, wherein the spaciousness is enhanced subsequent to signal processing such as directional filtering that degrades binaural cues for spaciousness. For such speech listening applications, it may be desirable to restrict it to situations in which speech reception is good so that the spatial enhancement processing, which has the potential to degrade speech reception, has minimal impact on intelligibility.
It may be desirable to apply the spatial enhancement processing only in some environments, specifically those in which natural cues for spaciousness are absent. Examples of such environments are music listening outdoors or in very large indoor venues, music listening when directional processing is activated in the hearing aids (e.g., in a noisy nightclub where it might be desirable to activate directionality in order to suppress background noise), listening to music streamed directly to the hearing aid, and speech listening with directional processing. In each of these examples, spaciousness processing should enhance sound quality.
The electronic circuitry of a hearing aid is contained within a housing that is commonly either placed in the external ear canal or behind the ear. In an example embodiment, a hearing assistance system comprises first and second hearing aids for providing audio outputs to both ears such as shown in
An example of the basic components of either hearing aid 10A or 10B are as shown in
In the embodiment illustrated in
The signal processing modules 120, 130, and 135 may represent specific code executed by the processor or may represent additional hardware components. The processing done by these modules may be performed in the time-domain or the frequency domain. In the latter case, the input signal is discrete Fourier transformed (DFT) prior to processing and then inverse Fourier transformed afterwards to produce the output signal for converting into sound. Any or all of the processing functions may also be performed for a plurality of frequency-specific channels, each of which corresponds to a frequency component or band of the audio input signal. Because hearing loss in most patients occurs non-uniformly over the audio frequency range, most commonly in the high frequency range, the patient's hearing deficit is compensated by selectively amplifying those frequencies at which the patient has a below-normal hearing threshold. The filtering and amplifying module 120 may therefore amplify the input signal in a frequency specific manner. The gain control module 130 dynamically adjusts the amplification in accordance with the amplitude of the input signal to either expand or compress the dynamic range and is sometimes referred to as a compressor. Compression decreases the gain of the filtering and amplifying circuit at high input signal levels so as to avoid amplifying louder sounds to uncomfortable levels. The gain control module may also apply such compression in a frequency-specific manner. The noise reduction module 135 performs functions such as suppression of ambient background noise and feedback cancellation.
As noted above, hearing aids typically perform signal processing in a frequency-specific manner, usually referred to as multichannel or multiband processing. In the time domain technique, a filter bank is used to separate the input signal into a multiplicity of frequency bands. The lowest frequencies are output by a low-pass filter, the highest frequencies by a highpass filter, and the remaining intermediate frequencies by band-pass filters. The input signal is convolved with the filters one sample at a time, and the output signal is formed by summing the filter outputs. The alternative frequency domain technique divides the input signal into short segments, transforms each segment into the frequency domain, processes the computed input spectrum, and then inverse transforms the segments to return to the time domain. Hearing aids may perform some functions in the time domain and others in the frequency domain. The spatial enhancement techniques described below may be performed in either the time domain or frequency domain upon discrete segments of the input signal that are then joined together to form the final output signal.
In one embodiment spaciousness is enhanced by randomly modifying phase in each channel of multiband signal processing in the hearing aids independently at the left and right ear. Such jittering is easily done, and requires little computational overhead, in hearing aids that already do multiband frequency domain signal processing for other purposes. Computational savings can be gained by doing the processing in a band-limited manner, for instance below 1500 Hz which is the frequency range in which humans are particularly sensitive to inter-aural de-correlation.
In a particular embodiment, the processing circuitries of the first and second hearing aids are configured to pseudo-randomly jitter the phases of their respective output signals in the spatial enhancement mode. The jittering may be performed as the input signal is processed in the frequency domain or the time domain, the latter being equivalent to time delay jittering, and may be applied in a frequency-specific manner. For example, the jittering may be applied with different parameters to different frequency bands of the input signal and/or the pseudo-random jittering may be performed only for frequency components of the input signal below a specified frequency (e.g., 1500 Hz).
The processing for doing the jittering may also be divided between the two hearing aids for computational efficiency. For example, one hearing aid may perform the jittering for one half of the frequency bands of the input signal, while the other hearing aid jitters the second half. In one embodiment, the processing circuitry of the first hearing aid is configured to perform pseudo-random jittering for at least one frequency component of the first input signal for which the corresponding frequency component of the second input signal is not pseudo-randomly jittered by the processing circuitry of the second hearing aid. In another embodiment, the processing circuitries of the first and second hearing aids are configured to perform pseudo-random jittering for different frequency components of their respective first and second input signals. The different frequency components jittered by each hearing aid may be in contiguous or non-contiguous frequency bands.
In an embodiment in which the first and second hearing aids each further comprise a radio-frequency (RF) transceiver connected to their processing circuitries, the processing circuitries may be configured to exchange parameters for pseudo-random jittering via an RF link between the two hearing aids upon initiation of the spatial enhancement mode.
In another embodiment, the hearing aids enhance spaciousness by applying generic head-related room impulse responses to the hearing aids at the left and right ears. The impulse responses used can be measured at the left and right ears of a dummy head in rooms and source locations that give good auditory spaciousness. One might even allow a patient to select from a library of rooms that are stored on the hearing aid, or selected and load from an external device such as a smart phone. The impulse responses at the two ears will differ from each other, particularly the parts due to early lateral reflections from the side walls of the room; it is these differences that give rise to the sense of spaciousness. Because it is the early reflections that contribute most to the sense of spaciousness, computational savings can be gained by truncating the impulse responses such that only early reflections are preserved and late reflections are eliminated.
In a particular embodiment, the processing circuitries of the first and second hearing aids are configured to employ a stored head-related room impulse response for each ear to produce an output signal in the spatial enhancement mode. In this embodiment, the processing circuitry of each hearing aid convolves its input signal with the stored impulse response in the time domain or performs an equivalent operation in the frequency domain. The stored head-related room impulse response may be produced from measurements of impulse responses recorded at the left and right ears of a dummy head in a selected environment. The measurements of the impulse responses at the left and right ears of the dummy head may be truncated to preserve early reflections and eliminate late reflections. A plurality of such head-related impulse responses may be stored, where the processing circuitries of the first and second hearing aids are then configured to select from the plurality of stored head-related room impulse responses to produce their output signals in the spatial enhancement mode.
In addition to the techniques for de-correlating left and right output signals by the techniques described above, mid/side processing is another way to improve spaciousness.
Mid/side processing refers to segregating the ambient (side) part of the sound from the nearfield (mid) part. In this segregated domain, one may perform processing separately and differently on the ambient and nearfield parts of the signal before recombining them into a binaural signal presented by the two hearing aids. Mid/side processing could be combined with those de-correlation techniques or used alone.
In the mid/side processing technique, the ambient and nearfield parts of the signal are formed from a sum of the first and second input signals and a difference between the two signals. This operation may be performed by both of the first and second hearing aids, where the input signal from one hearing aid is transmitted to the other via the RF link using RF transceivers incorporated into each hearing aid. The resulting ambient and nearfield signals may then be processed non-linearly and recombined, possibly multiple times. An example sequence of operations is as follows: 1) separating each of the first and second input signals into ambient and nearfield signals by summing and subtracting the first and second input signals, 2) performing separate compressive amplification of the ambient and nearfield signals by each hearing aid, 3) generating first and second output signals by recombining the signals with a weighted combination, 4) repeating steps 1-3 a specified number of times.
In another embodiment, the spatial enhancement mode employing any of the de-correlation techniques described above may include further processing of the output signals that involves computing sums and differences between the output signals computed by each of the first and second hearing aids. In this embodiment, the first and second hearing aids each further comprise a radio-frequency (RF) transceiver connected to their processing circuitries for providing an RF link between the two hearing aids in order to communicate their respective output signals to the other hearing aid. The processing circuitry of each hearing aid is configured to produce a final output signal as a weighted sum of the de-correlated output signals produced by the processing circuitries of both of the first and second hearing aids. The processing can be inexpensively done in the time domain, but it could be done in the frequency domain as well.
The above-described embodiments have applied spatial enhancement processing to input signals produced by the hearing aids from actual sounds. Such spatial enhancement processing may also be applied to input signals transmitted directly to the hearing aids from an external device. For example, a music player (e.g., a smart phone) may wireless transmit one channel of a stereo signal to each hearing aid via the RF link or a wired connection. The received input signals are processed in the spatial enhancement mode in same manner as described above with respect to input signals derived from actual sounds.
In another embodiment, the user interface as described above may be configured to allow users to adjust the de-correlation parameters used in the above-described embodiments to suit their personal preferences for particular listening situations. For example, in the case of phase jittering, a user may adjust the amount of jittering and/or the frequency bands to which the jittering is applied. In the case of mid-side processing, the user may adjust the weightings used to combine the ambient and nearfield signals.
The subject matter has been described in conjunction with the foregoing specific embodiments. It should be appreciated that those embodiments and specific features of those embodiments may be combined in any manner considered to be advantageous. Also, many alternatives, variations, and modifications will be apparent to those of ordinary skill in the art. Other such alternatives, variations, and modifications are intended to fall within the scope of the following appended claims.
