3D TruSound
Preserving the fundamentals of a natural sound experience
InterEar communication through a new, advanced wireless technology
Our new, proprietary WidexLink technology, which has been developed specifically for data exchange and audio streaming in a hearing aid system, offers new and exiting possibilities for exchange of data between hearing aids, and between hearing aids and external devices. The new technology enables the left and right hearing aid in a pair to share the information obtained by the opposite hearing aid, so that information from both ears is taken into account during signal processing. We call this data exchange InterEar communication.
InterEar communication is part of the foundation of the advanced features that comprise 3D TruSound. 3D TruSound is a new dimension of sound processing which aims at preserving the fundamentals of a natural sound experience and providing the highest possible sound quality at the same time.
The TruSound inheritance is excellent sound quality. With the introduction of 3D TruSound, a new dimension is added; namely the preservation of important localisation cues in natural hearing. In addition to the preservation of important sound localisation cues, coordinated noise reduction in difficult listening situations, and enhanced sound quality features are also central elements in 3D TruSounds.
3D TruSound includes the Digital Pinna – a feature which simulates the shadow effect of the outer ear in natural hearing. Furthermore, 3D TruSound also features TruSound Softener which can handle ultra short and extremely fast changes in sound level – for instance when somebody drops cutlery in a metal sink.
Figure 1. The3D TruSound features
The 3D TruSound features will be introduced in more detail in the following.
The Fundamentals of a Natural Sound Experience
Important sound localisation cues in normal directional hearing
The ability for hearing aid users to determine where sounds are coming from is important for a number of reasons, including safety (e.g. traffic sounds) and communication (e.g. locating a new speaker in a group).
Normal directional hearing relies on the comparison of auditory input from two ears. When an insect flies around our head we are able to determine where it is even with our eyes closed. That is only possible because we have two ears and a brain that help us coordinate the information from both sides of the head.
One of the primary psychoacoustic cues used for localising a sound source to the right or to the left is the split-second delay between the time when a sound reaches the near ear and when it reaches the far ear. This delay is referred to as the interaural time difference (ITD). The ITD will be at its maximum when the sound originates directly from the sides of the head. It is not a large difference; the maximum is around 0.65ms (Plack, 2005). When the sound is coming from the front or the back, the distance from the source to the ears is the same. Thus, there is no interaural time difference for sounds coming directly from the front or the back.
Figure 2. Sounds coming from the left or the right will reach the near ear a little sooner than the far ear. The ITD will be at its maximum when the sound comes directly from one side of the head.
Another psychoacoustic cue that is used for horizontal localisation is the interaural level difference (ILD). When a sound source is located to the right or the left of the head, the sound will have a greater intensity level when it reaches the near ear than when it reaches the far ear. This difference in sound pressure level at the near and far ear is an important cue for the localisation high-frequency sounds. The effect is predominant for high-frequency sounds because of their short wavelengths. At low frequencies, the difference in level at the two ears from sound coming from the side is low, because long waves easily flow around the head. A head is not a very big object for a low-frequent sound with a long wavelength, but it is a large obstacle for a high-frequency sound with a short wavelength.
Figure 3. Example of the interaural level difference (ILD) at different frequencies measured for one person. Notice the predominance in the high frequencies.
It is widely accepted (e.g., Middlebrooks & Green, 1991; Wightman & Kistler, 1992; Schub et al., 2008) that the ILD delivers the primary cue for horizontal localisation of sound sources in the high-frequency region and that the ITD delivers the primary cue for the localisation of low-frequency sounds. In this context, the split between high and low frequencies is approximately 1.5 kHz.
Research (Bogaerts et al., 2006) has shown that wearing a pair of uncoordinated hearing aids can have a destructive effect on the cues used for localisation and consequently reduce the localization abilities of the hearing aid user. However, with the new technology in CLEAR440, it is possible to preserve important psychoacoustic cues.
InterEar TruSound compression – preserving important localisation cues
In a normal compressions system, gain changes depending on the input level. Thus, when you are wearing an uncoordinated set of hearing aids, gain will be prescribed independently to the near and far ear on the basis of the input level at the individual ear. A sound coming from the side will have less intensity when it reaches the far ear than when it reaches the near ear. And because the sound is lower in intensity when reaching the far ear, more gain will be provided at the far ear than at the near ear. This means that the natural interaural level difference will be compromised. However, by coordinating the gain changes on the two sides, the natural ILD can be preserved. This is what the InterEar TruSound Compression does.
Specifically, in CLEAR440 hearing aids, the communication between two coordinated aids ensures that the input levels at the hearing aids are constantly compared (21 times per second), and that the compression response is adjusted accordingly to reflect the difference in input level at the two ears. In practical terms, this means that both ears will receive the same amount of gain, which will depend on the basis of the input level in the near ear.
InterEar volume shift and program control – for the preservation of important sound localisation cues and ease of use
The main purpose of coordinated compression is to preserve the ILD. However, if the user turns the volume up or down or changes program at one side only, the ILD can no longer be preserved. In order to prevent that the attempt to preserve the natural ILD is obstructed, InterEar volume shift and program control is introduced in CLEAR440 hearing aids. The InterEar volume shift and program control ensures that if the user changes the volume on one hearing aid the volume of the other hearing aid will also change accordingly. If the users changes program on one hearing aid, the same program will be chosen automatically by the other hearing aid. InterEar volume shift and program control is switched on as default in CLEAR440 hearing aids, but can be switched off by the hearing care professional in the fitting software.
Another major advantage of the InterEar volume shift and program control is that it makes it a lot easier for the hearing aid user to adjust the volume. Users wearing an uncoordinated pair of hearing aids have to adjust each hearing aid separately every time they need to turn the volume up or down or change program. With a pair of coordinated CLEAR440 hearing aids, the user only has to make the adjustment once, which is likely to be appreciated by many hearing aid users.
There may be situations where the hearing aid users would prefer to have different programs in their hearing aids. To accommodate this, a selection of compound packages is available in CLEAR440. These include:
- Master – Telecoil
- Master – Microphone+ Telecoil
- Master – Reverse focus
- Master – Zen
- Master – Audibility Extender
InterEar Speech Enhancer - Coordinated noise reduction in difficult listening situations
Communicating in a noisy environment is one of the most challenging situations for hearing aid users. Many users experience difficulties focusing on one speaker and leaving out the rest. The Speech Enhancer can be very helpful in that situation, and with the introduction of the WidexLink technology the situation can be improved even further.
The Speech Enhancer system is available in Widex high end hearing aids. The system is based on the standardised measure of SII (Speech Intelligibility Index) (ANSI S3.5). The system is unique in that it is able to take the hearing loss into account and optimise speech intelligibility by means of a constant calculation of the SII during noise reduction.
The Speech Enhancer contains a fast-acting mechanism which is able to add gain to frequency areas with speech to further optimise speech intelligibility. In CLEAR440 hearing aids, this fast-acting mechanism is coordinated to ensure that it is active on the side with the most dominant speaker. By exchanging important percentile data over the WidexLink, two CLEAR440 hearing aids are able to make a decision on whether to activate the fast-acting gain application mechanism and on which side to do this. This way, the Speech Enhancer systems will no longer base its decision to act on one-sided data input, but on data input on the sound environment on both sides of the head. Only on the side where speech is dominant will the speech enhancer actively work to preserve speech audibility by adding gain to frequency regions important for speech. On the opposite side the noise reduction system will work to keep the noise below the threshold of the listener using his or her hearing threshold data in the calculation. In practical terms, the coordination between hearing aids means that in a cocktail party situation with many speakers, the InterEar Speech Enhancer in CLEAR440 supports the singling out of the dominant speaker.
Digital Pinna – re-creating the natural pinna shadow effect
The ear has some natural directional characteristics, mainly due to the physical presence and shape of the pinna. One example is the pinna shadow effect. Sounds coming from the front reach the ear canal almost directly, whereas sounds coming from behind are obstructed by the pinna and are thus somewhat attenuated before they reach the ear canal. This pinna shadow effect mainly affects the region around 2 to 5 kHz, where sounds coming from behind are attenuated by 3-4 dB relative to sounds coming from the front. This natural 3-4 dB difference is an important cue for the listener to know whether a source is in front or behind.
Microphone location is known to have a negative impact on the ability to determine if a sound is coming from in front or from behind. Front-back localisation is especially a problem for users with behind-the-ear (BTE) hearing aids, because the location of the microphone essentially offsets the normal localisation cues provided by the pinna (the outer ear).
For example, Westermann and Tøpholm (1985) found that BTEs give poorer localisation than ITEs, in particular with respect to front-back confusions. A BTE hearing aid captures the sound at the position of its microphone(s), i.e. above and behind the pinna. This means that the pinna shadow effect is not preserved in the signal that the HA provides, and consequently the HA user loses some of the ability to localise vertically, and to distinguish front and back.
Figure 4. The average pinna effect measured over 45 heads. The 3-4 dB difference between sound coming from behind and from the front is seen in the 2-5 kHz frequency region.
As mentioned, the spectral shaping provided by the pinna is an important cue for front/back localisation. One way to improve localisation for hearing aid users with BTEs, therefore, would be to attempt to re-create this shaping through processing of the input signal. This is precisely what Digital Pinna in CLEAR440 does.
A series of developmental experiments have shown that the pinna shadow effect can be simulated by introducing a restriction on the adaptive locator. More specifically, the natural attenuation of sounds coming from behind can be re-created by setting the frequency bands from 2 kHz and up (bands 10-15) in fixed directional mode (i.e. in the hypercardioid pattern, which picks up sound at the front and eliminates most sound from the sides and rear), while leaving the lower bands (1-9) in omni-directional mode.
Figure 5. The microphone patterns of Digital Pinna. The lower bands (1-9) are in omni-directional mode, while the upper bands (10-15) are in fixed directional mode (i.e., in hypercardioid).
The developmental experiments were carried out using four normal-hearing listeners as subjects. 12 loudspeakers were set up in a 360° horizontal circle around the test subject, with a radius of 1.2 meters and 30° between the loudspeakers. The setup is shown in the figure below.
Figure 6. The loudspeaker setup in the developmental experiments exploring Digital Pinna.
As stimuli, 4 different recordings of a wooden rod (120cm) that is dropped on the floor were used. Each of these 4 recordings was presented once from each of the 12 loudspeakers, giving a total of 4*12=48 presentations (in randomised order) for each localisation test. The test subjects were instructed to indicate from which loudspeaker each presentation was coming. The Digital Pinna was compared with the omni-directional mode and with a fixed locator (the one giving maximum directivity).
Figure 7 below shows the main results for the four different tests where head movements were not allowed. Digital Pinna resolved 80% of the front-back confusions that were made in omni-directional mode. This performance is slightly better than (or at least similar to) the fixed directional system. In addition, Digital Pinna did not degrade the horizontal localisation as the fixed directional system did, and Digital Pinna had the best horizontal localisation in the frontal plane.
Figure 7. Main result of a developmental experiment comparing front-back confusion with Digital Pinna, omni-directional and fixed directional mode. Head movements were not allowed in this condition. The left panel shows how many front/back confusions were made (in %) in each mode. The right panel shows by how many degrees horizontal (left-right) localisation errors deviated from the correct response angle on average.
When head movements were allowed, the Digital Pinna was clearly best overall. The results are displayed in figure 8 below. There were almost no front-back confusions with Digital Pinna compared to omni-directional and fixed directional mode, and the horizontal localisation did not show large errors in the back, like the fixed directional mode did.
Figure 8.
Main result of a developmental experiment comparing front-back confusion with Digital Pinna, omni-directional and fixed directional mode when head movements were allowed. The left panel shows how many front/back confusions were made (in %) in each mode. The right panel shows by how many degrees horizontal (left-right) localisation errors deviated from the correct response angle on average.
The results from the developmental studies indicate that Digital Pinna restores the ear's natural effect (pinna shadow) in BTEs, and thus the user's ability to distinguish between sources in front and back. The results also showed that it did so without degrading horizontal localisation (which the fixed directional system may). Furthermore, it enables the user to improve localisation by taking advantage of active head movements (which the fixed directional system does not).
Importantly, the microphone system remains adaptive when Digital Pinna has been activated. In quiet listening environments, the lower bands (1-9) will be in omni-directional mode, whereas the upper bands (10-15) will be in hypercardioid as shown in figure 5. In noisy listening conditions, the directionality of both lower and upper bands will increase to yield optimum speech intelligibility. The microphone mode of the lower bands may adapt any characteristic from omni-directional to bipolar, while the upper bands may move from hypercardioid towards bipolar as the signal-to-noise ratio worsens.
Advanced Sound Quality Features - Taking high sound quality a step further
A discussed above, one of the key elements in the CLEAR440 hearing aid is a collection of features designed to preserve important psychoacoustic cues. Another cornerstone in the CLEAR440 product family is a collection of enhanced sound quality features. These features will be described in more detail in the following.
InterEar feedback cancelling
To be successful, a feedback system must be effective in terms of eliminating feedback. The precision with which it can determine if the signal really is a feedback signal is also important.
Experience has shown that the Multi-directional active feedback cancelling system is extremely efficient in controlling dynamic feedback problems. No matter if the hearing aid user is talking on the phone, hugging a friend, or putting on a hat, the Multi-directional active feedback cancelling system has been designed to ensure that whistling does not occur.
With the introduction of CLEAR440, we have managed to make our industry leading system even more precise. Specifically, when an external, autocorrelated sound like a whistle or an alarm is picked up by the hearing aid, the InterEar coordination between the hearing aids means that they are able to compare detected sound from both sides of the head. If the feedback-like signal is the same on both sides, it can be deducted that it is an external sound rather than a feedback signal which has been detected. Thus, with InterEar feedback cancellation, we are able to avoid gain regulation when it is not necessary as a result of "false positives". However, if a feedback-like sound is only found on one side of the head, the system will deduct that it is feedback which needs to be handled.
Enhanced bandwidth
Audio bandwidth is one of the key factors in maintaining a high sound quality. Thanks to the new technology in the CLEAR440 product range, we are able to offer an exceptionally broad audio bandwidth in models with a Clearband receiver, stretching from 70 Hz to 10.5 kHz in the music program, and 100 Hz to 11.2 kHz for digitally transmitted sound. This is industry leading.
One area where a broad audio bandwidth makes a clear difference for hearing aid users is when they listen to music. The high frequencies provide ambience and brilliance to the sound. Thus, the sound experience will be somewhat richer with an upper bandwidth of 10.5 kHz when listening to a hi-hat or cymbals, for instance. Similarly, an audio bandwidth stretching as far down as 70 Hz will produce a fuller bass.
TruSound Softener
The advanced sound quality features of 3D TruSound also include the TruSound Softener. The purpose of the TruSound Softener feature is to make impulse sounds, such as rattling porcelain or hammer blows, less annoying without removing them from the surrounding sound environment or making them unnaturally soft. The TruSound Softener is described in more detail in a separate whitepaper entitled TruSound Softener: A new algorithm for detecting and handling impulse sounds.
Summary
With 3D TruSound, Widex takes sound processing a step further. 3D TruSound introduces a new dimension in sound processing – the preservation of the fundamentals of a natural sound experience. This is achieved by taking signal analysis data from the opposite hearing aid into account in the processing.
The new proprietary WidexLink technology in the CLEAR440 product family makes it possible to preserve a number of important psychoacoustic cues used for determining where a sound is coming from.
Our attempt to provide a realistic listening experience for the hearing aid user has lead to the development of a comprehensive collection of features, including
- Digital Pinna, which has been developed to support front-back localisation
- InterEar TruSound compression, and InterEar volume shift and program control, which contribute to the preservation of psychoacoustic cues used for localising sound coming from the sides.
- InterEar Speech Enhancer, which may help hearing aid users focus on the dominant speaker in noisy situations
Another cornerstone in the CLEAR440 product family is a collection of enhanced sound quality features. The result of our latest effort to provide the best possible sound quality include
- The TruSound Softener, which has been designed to detect and handle impulse sounds
- Extended bandwidth in both high and low frequencies
- InterEar feedback cancelling which minimises the risk of continuous steady sounds being attenuated by the feedback system because they are mistaken for feedback
References
ANSI S3.5. 1997. American National Standard: Methods for the calculation of the Speech Intelligibility Index.
Algazi, V. R.; Duda, R. O.; Thompson, D. M. & Avendano, C. (2001). The CIPIC HRTF Database.
In proceedings of IEEE Workshop on Applications of Signal Processing to Audio and Acoustics, 99-102.
Middlebrooks, J. C., & Green, D. M. (1991). Sound localization by human listeners. Annual Review of Psychology, 42, 135-159.
Plack, C. J. (2005). The sense of hearing. New Jersey: Lawrence Erlbaum Associates
Schub, D. E., Carr, S. P., Kong, Y., & Colburn, H. S. (2008). Discrimination and identification of azimuth using spectral shape. Journal of the Acoustical Society of America, 124(5), 3132-3141.
Van den Bogaert, T., Klasen, T., Moonen, M., Van Deun, L., & Wouters, J. (2006). Horizontal localization with bilateral hearing devices: Without is better than with. Journal of the Acoustical Society of America, 119(1), 515-526.
Westermann, S. & Tøpholm, J. (1985). Comparing BTEs and ITEs for localizing speech. Hearing Instruments, 36(2), 20-24.
Wightman, F. L. & Kistler, D. J. (1992). The dominant role of low-frequency interaural time differences in sound localisation. Journal of the Acoustical Society of America, 91(3), 1648-1661.
