Is complex signal processing for bone conduction hearing aids useful? Martin Kompis, Anja Kurz, Flurin Pfiffner, Pascal Senn, Andreas Arnold, Marco Caversaccio Department of ENT, Head and Neck Surgery, Inselspital, University of Bern, 3010 Bern, Switzerland Objectives: To establish whether complex signal processing is beneficial for users of bone anchored hearing aids. Methods: Review and analysis of two studies from our own group, each comparing a speech processor with basic digital signal processing (either Baha Divino or Baha Intenso) and a processor with complex digital signal processing (either Baha BP100 or Baha BP110 power). The main differences between basic and complex signal processing are the number of audiologist accessible frequency channels and the availability and complexity of the directional multi-microphone noise reduction and loudness compression systems. Results: Both studies show a small, statistically non-significant improvement of speech understanding in quiet with the complex digital signal processing. The average improvement for speech in noise is +0.9 dB, if speech and noise are emitted both from the front of the listener. If noise is emitted from the rear and speech from the front of the listener, the advantage of the devices with complex digital signal processing as opposed to those with basic signal processing increases, on average, to +3.2 dB (range +2.3 … +5.1 dB, p ≤ 0.0032). Discussion: Complex digital signal processing does indeed improve speech understanding, especially in noise coming from the rear. This finding has been supported by another study, which has been published recently by a different research group. Conclusions: When compared to basic digital signal processing, complex digital signal processing can increase speech understanding of users of bone anchored hearing aids. The benefit is most significant for speech understanding in noise. Keywords: Bone anchored hearing aids, Complex signal processing
Introduction
almost intact, there is no need to compensate for recruitment and compression becomes less important. Nevertheless, complex signal processing may be useful for Baha. As Baha sound processors are worn behind the ear, sounds are picked up more easily from the rear than from the front. A directional compensation might bring its directional sensitivity closer to that of a normal human ear. Furthermore, complex signal processing may improve speech understanding in users with a combined hearing loss or in special indications, such as in single-sided deafness (Pfiffner et al., 2011b.) In this report, we compare speech understanding in quiet and in noise with four different Baha devices, two of which feature complex signal processing and two of which have only basic digital signal processing.
Bone anchored hearing aids (Bahas) have been successfully used for over 3 decades. For most of this time, Baha sound processors were relatively simple, basically analog, linear devices. Even after the introduction of complex digital sound processors for Baha, the question remains, whether they add a true benefit for the user or not. There are several reasons, why simple linear processing might seem adequate in Bahas, as opposed to conventional hearing aids. One way of reasoning is that already with simple linear signal processing, results with Baha tend to be rather good (Pfiffner et al., 2009) when compared to conventional hearing aids. This advantage, however, is probably an effect of the often well-preserved cochlear function in Baha users, as opposed to the sensorineural hearing loss in many users of conventional hearing aids. If the cochlear function is intact or
Materials and methods Study devices
Correspondence to: Martin Kompis, Department of ENT, Head and Neck Surgery, Inselspital, University of Bern, 3010 Bern, Switzerland. Email: [email protected]
Speech understanding was tested with four different Baha sound processors: Baha Divino, Baha BP100, Baha Intenso, and Baha BP110 power. Table 1
© W. S. Maney & Son Ltd 2014 DOI 10.1179/1467010014Z.000000000167
Cochlear Implants International
2014
VOL.
15
NO.
S1
S47
Kompis et al.
Is complex signal processing for bone conduction hearing aids useful?
Table 1 Synopsis of the four Baha sound processors Medium power devices
High power devices
Device type
Divino
BP100
Intenso
Signal processing Adjustability
Basic digital Low cut, automatic gain control (AGC-O) 1
Complex digital Gain at 3 input levels
Basic digital Complex digital Low cut, Gain at 3 input levels gain 1 10
2 Yes (fixed)
3 Yes (fixed and adaptive) 13
Audiologist accessible channels Programs Multi-microphone noise reduction Battery size
13
provides a synopsis on the pertinent features of these processors. Two processors (Divino and BP100) are medium output power devices and the other two (Intenso and PB110 power) are high power devices. Two sound processors (Divino and Intenso) feature basic digital signal processing with only a single channel with the possibility to adjust the lowfrequency gain separately (Pfiffner et al., 2011b). Of these simpler processors, only the Divino has an optional dual microphone directional noise reduction feature. In contrast, the BP 100 and BP110 power feature substantially more complex signal processing. Gain can be adjusted by the audiologist at three different input levels and in 10 separate frequency channels. Fixed and adaptive directionality and different noise reduction modes can be chosen. In any microphone mode, a feature called ‘position compensation’ can be selected, which activates a small additional directionality to the front of the user, thus offsetting the drawback of the position of the Baha behind the ear.
10
2 No 675
BP110 power
3 Yes (fixed and adaptive) 675
For the second study (Kurz et al., 2013), 20 users of any Baha processor were recruited. They were first tested without Baha, then 10 of them were fitted with a Baha Intenso, the other 10 with a Baha BP110. After one month of use, speech tests in quiet and in noise were performed with the sound processor fitted in the first session and the other test processor was fitted. Another month later, tests were performed with the second sound processor. All tests included speech understanding in quiet under sound field conditions at 65 dB SPL with the German Freiburger Monosyllabic word test and speech tests in noise using the German Oldenburger Sentence Test (Wagener and Brand, 1999). All tests were performed in a sound proof room with speech emitted by a loudspeaker placed at a distance of 1 m in front of the listener. Noise was presented either from the same loudspeaker (spatial setting S0N0) or from a second loudspeaker placed 1m away in the rear of the listener (setting S0N180). All speech in
Study protocols Two separate studies were performed: one comparing the two medium output devices Baha Divino with the Baha BP100 (Pfiffner et al 2011a) and one comparing the two high-power devices, Baha Intenso and Baha BP110 (Kurz et al., 2013, in press). Both studies were approved by the local ethical committee performed in accordance with the declaration of Helsinki. In the first study (Pfiffner et al., 2011a), 20 experienced users of Baha Divinos were tested first without and then with their Baha. Tests included speech understanding in quiet and in noise, as described in detail below. They were then fitted with a BP100 and subsequently used this new processor for 3 months. After this test period they returned for speech tests with the BP100. As a part of this second session, some of the tests with the Divino were repeated to ascertain that differences in test results were due to the differences in the processor and not due to learning effects.
S48
Cochlear Implants International
2014
VOL.
15
NO.
S1
Figure 1 Speech understanding in quiet. Scores at a presentation level of 65 dB are shown. For medium and for high power devices, there is a statistically significant improvement from the unaided situation to any of the processors. An additional, smaller and statistically not significant improvement can be seen between basic and complex signal processing.
Kompis et al.
Is complex signal processing for bone conduction hearing aids useful?
the Intenso-BP110 power comparison, only the results within each of the studies, but not between the two studies, are directly comparable. For the medium power processors, the 20 subjects were 22 to 72 years (average: 51 years). Twelve subjects were implanted on the right side and 8 on the left side. For the high power processors, the 20 subjects were 20–74 years (average: 47 years). Ten were implanted on the right side, 10 on the left and had used a Baha for 1 to 17 years (average 9.1 years).
Results
Figure 2 Speech understanding in noise for speech and noise emitted by a single loudspeaker in front of the listener. The differences between the signal-to-noise ratio (SNR) in dB for 50% speech understanding between basic digital signal processing (Divino and Intenso) and complex digital signal processing (BP100 and BP110) is shown. Speech understanding is slightly better (+0.7 to +1.2 dB) with complex processing, only one of the differences is statistically significant (*P < 0.05).
noise tests were performed once in omnidirectional mode and once in fixed directional mode for those three devices, which feature a multi-microphone noise reduction system (cf. Table 1).
Subjects Twenty adult, unilaterally implanted, German-speaking Baha users participated in each of the two studies. All subjects had a bilateral conductive or mixed hearing loss. As different populations participated in the Divino-BP100 comparison from those in
Figure 3 Speech understanding of speech coming from the front in noise from the rear. The differences between the signal-to-noise ratio (SNR) in dB for 50% speech understanding between basic digital signal processing (Divino and Intenso) and complex digital signal processing (BP100 and BP110) are shown. Speech understanding is improved by +2.3 to +5.1 dB with complex processing (**P < 0.01).
Fig. 1 shows the results of the speech understanding tests in quiet at 65 dB. For both groups of participants, there is a substantial improvement in word understanding from the unaided condition to either of the two aided conditions. The difference ranges from 33.8 to 55.9% points and is statistically highly significant (P < 0.001; Wilcoxon matched pairs signedrank test). In addition, there is a smaller, statistically non-significant increase from the devices with a basic digital signal processing to the complex processing by 12.7% points for the BP100 vs. the Divino and by 1.3% points for the BP110 power vs. the Intenso. Fig. 2 shows the differences between devices with basic and complex signal processing for speech reception in noise, if speech and noise are emitted by the same loudspeaker in front of the listener (S0N0). In all four comparisons, there is a small gain of +0.7 to +1.2 dB in terms of signal-to-noise advantage. However, only the largest of these improvements is statistically significant. Also, the choice of the directional or the omnidirectional microphone has no influence, as the noise and the target signal sources are not spatially separated. Fig. 3 shows the results for speech from the front and noise from the rear (S0N180). There is a considerable (+2.3 to +5.1 dB) and statistically significant (P ≤ 0.0032) improvement for the sound processors with the complex digital signal processing. For the Divino/BP100 comparison, the advantage of the BP100 is the same for the directional or the omnidirectional mode, suggesting that the directional benefit is similar in both devices. For the Intenso/BP110 comparison, the difference is much larger in the directional mode, as the Intenso does not have this feature and all measurements had to be done in the omnidirectional mode.
Discussion and summary The studies presented suggest that complex digital signal processing is indeed beneficial for Baha users, when compared with basic digital signal processing. The benefit varies with the test setting. For speech in quiet, there may be an advantage, but it is relatively small and not statistically significant in
Cochlear Implants International
2014
VOL.
15
NO.
S1
S49
Kompis et al.
Is complex signal processing for bone conduction hearing aids useful?
our two groups of subjects. For speech in noise, where both speech and noise arrive from the same loudspeaker, there is, again, a small but consistent advantage in the order of magnitude of 1 dB, which is only for one comparison statistically significant. This is probably a beneficial effect of the multiband adjustable gain and compression, which allows, among others, a higher gain the high-frequency region (Pfiffner et al., 2011a; Kurz et al., 2013). In contrast, for speech from the front and noise from the rear, the benefit is much larger (+2.3 to +5.1 dB) and statistically significant for both pairs of speech processors. The position compensation feature, which is a small directionality feature present only in the devices with complex signal processing, is most probably the reason for the greater benefit even in the omnidirectional mode. Note that the Baha Intenso has no directional mode, thus the benefit is largest (+5.1 dB) when compared to the directional setting of the BP110 power. Very recently, another research group has published another study comparing the basic and complex signal processing in the Baha Intenso and the Baha BP110 power (Desmet et al., 2013). Their results are similar to those presented here: a small benefit of the device
S50
Cochlear Implants International
2014
VOL.
15
NO.
S1
with complex signal processing in quiet and in noise from the front, and a larger gain for noise from the rear, suggesting that complex signal processing can indeed be beneficial for Baha user.
References Desmet J.B., Bosman A.J., Snik A.F., Lambrechts P., Hol M.K., Mylanus E.A., De Bodt M., Van de Heyning P. (2013) Comparison of Sound Processing Strategies for Osseointegrated Bone Conduction Implants in Mixed Hearing Loss: Multiple-Channel Nonlinear Versus SingleChannel Linear Processing. Otology and Neurotology, 34(4): 598–603. Kurz A., Caversaccio M., Kompis M. (2013) Hearing performance with 2 different high-power sound processors for osseointegrated auditory implants. Otology and Neurotology, 34(4): 604–610. Pfiffner F., Caversaccio M., Kompis M. 2011a. Comparisons of Sound Processors Based on Osseointegrated Implants in Patients With Conductive or Mixed Hearing Loss. Otology and Neurotology, 32: 728–735. Pfiffner F., Kompis M., Flynn M., Åsnes K., Arnold A., Stieger C. 2011b. Benefits of Low Frequency Attenuation of Baha® in Single Sided Sensorineural Deafness. Ear and Hearing, 32(1): 40–45. Pfiffner F., Kompis M., Stieger C. 2009. Bone Anchored Hearing Aids: Correlation between pure tone thresholds and outcome in three users groups. Otology and Neurotology, 30: 884–890. Wagener K., Brand T.K.B. 1999. Entwicklung und Evaluation eines Satztestes für die deutsche Sprache Teil III: Evaluation des Oldenburger Satztestes. Audiologic Acoust, 38: 86–95.
Introduction
almost intact, there is no need to compensate for recruitment and compression becomes less important. Nevertheless, complex signal processing may be useful for Baha. As Baha sound processors are worn behind the ear, sounds are picked up more easily from the rear than from the front. A directional compensation might bring its directional sensitivity closer to that of a normal human ear. Furthermore, complex signal processing may improve speech understanding in users with a combined hearing loss or in special indications, such as in single-sided deafness (Pfiffner et al., 2011b.) In this report, we compare speech understanding in quiet and in noise with four different Baha devices, two of which feature complex signal processing and two of which have only basic digital signal processing.
Bone anchored hearing aids (Bahas) have been successfully used for over 3 decades. For most of this time, Baha sound processors were relatively simple, basically analog, linear devices. Even after the introduction of complex digital sound processors for Baha, the question remains, whether they add a true benefit for the user or not. There are several reasons, why simple linear processing might seem adequate in Bahas, as opposed to conventional hearing aids. One way of reasoning is that already with simple linear signal processing, results with Baha tend to be rather good (Pfiffner et al., 2009) when compared to conventional hearing aids. This advantage, however, is probably an effect of the often well-preserved cochlear function in Baha users, as opposed to the sensorineural hearing loss in many users of conventional hearing aids. If the cochlear function is intact or
Materials and methods Study devices
Correspondence to: Martin Kompis, Department of ENT, Head and Neck Surgery, Inselspital, University of Bern, 3010 Bern, Switzerland. Email: [email protected]
Speech understanding was tested with four different Baha sound processors: Baha Divino, Baha BP100, Baha Intenso, and Baha BP110 power. Table 1
© W. S. Maney & Son Ltd 2014 DOI 10.1179/1467010014Z.000000000167
Cochlear Implants International
2014
VOL.
15
NO.
S1
S47
Kompis et al.
Is complex signal processing for bone conduction hearing aids useful?
Table 1 Synopsis of the four Baha sound processors Medium power devices
High power devices
Device type
Divino
BP100
Intenso
Signal processing Adjustability
Basic digital Low cut, automatic gain control (AGC-O) 1
Complex digital Gain at 3 input levels
Basic digital Complex digital Low cut, Gain at 3 input levels gain 1 10
2 Yes (fixed)
3 Yes (fixed and adaptive) 13
Audiologist accessible channels Programs Multi-microphone noise reduction Battery size
13
provides a synopsis on the pertinent features of these processors. Two processors (Divino and BP100) are medium output power devices and the other two (Intenso and PB110 power) are high power devices. Two sound processors (Divino and Intenso) feature basic digital signal processing with only a single channel with the possibility to adjust the lowfrequency gain separately (Pfiffner et al., 2011b). Of these simpler processors, only the Divino has an optional dual microphone directional noise reduction feature. In contrast, the BP 100 and BP110 power feature substantially more complex signal processing. Gain can be adjusted by the audiologist at three different input levels and in 10 separate frequency channels. Fixed and adaptive directionality and different noise reduction modes can be chosen. In any microphone mode, a feature called ‘position compensation’ can be selected, which activates a small additional directionality to the front of the user, thus offsetting the drawback of the position of the Baha behind the ear.
10
2 No 675
BP110 power
3 Yes (fixed and adaptive) 675
For the second study (Kurz et al., 2013), 20 users of any Baha processor were recruited. They were first tested without Baha, then 10 of them were fitted with a Baha Intenso, the other 10 with a Baha BP110. After one month of use, speech tests in quiet and in noise were performed with the sound processor fitted in the first session and the other test processor was fitted. Another month later, tests were performed with the second sound processor. All tests included speech understanding in quiet under sound field conditions at 65 dB SPL with the German Freiburger Monosyllabic word test and speech tests in noise using the German Oldenburger Sentence Test (Wagener and Brand, 1999). All tests were performed in a sound proof room with speech emitted by a loudspeaker placed at a distance of 1 m in front of the listener. Noise was presented either from the same loudspeaker (spatial setting S0N0) or from a second loudspeaker placed 1m away in the rear of the listener (setting S0N180). All speech in
Study protocols Two separate studies were performed: one comparing the two medium output devices Baha Divino with the Baha BP100 (Pfiffner et al 2011a) and one comparing the two high-power devices, Baha Intenso and Baha BP110 (Kurz et al., 2013, in press). Both studies were approved by the local ethical committee performed in accordance with the declaration of Helsinki. In the first study (Pfiffner et al., 2011a), 20 experienced users of Baha Divinos were tested first without and then with their Baha. Tests included speech understanding in quiet and in noise, as described in detail below. They were then fitted with a BP100 and subsequently used this new processor for 3 months. After this test period they returned for speech tests with the BP100. As a part of this second session, some of the tests with the Divino were repeated to ascertain that differences in test results were due to the differences in the processor and not due to learning effects.
S48
Cochlear Implants International
2014
VOL.
15
NO.
S1
Figure 1 Speech understanding in quiet. Scores at a presentation level of 65 dB are shown. For medium and for high power devices, there is a statistically significant improvement from the unaided situation to any of the processors. An additional, smaller and statistically not significant improvement can be seen between basic and complex signal processing.
Kompis et al.
Is complex signal processing for bone conduction hearing aids useful?
the Intenso-BP110 power comparison, only the results within each of the studies, but not between the two studies, are directly comparable. For the medium power processors, the 20 subjects were 22 to 72 years (average: 51 years). Twelve subjects were implanted on the right side and 8 on the left side. For the high power processors, the 20 subjects were 20–74 years (average: 47 years). Ten were implanted on the right side, 10 on the left and had used a Baha for 1 to 17 years (average 9.1 years).
Results
Figure 2 Speech understanding in noise for speech and noise emitted by a single loudspeaker in front of the listener. The differences between the signal-to-noise ratio (SNR) in dB for 50% speech understanding between basic digital signal processing (Divino and Intenso) and complex digital signal processing (BP100 and BP110) is shown. Speech understanding is slightly better (+0.7 to +1.2 dB) with complex processing, only one of the differences is statistically significant (*P < 0.05).
noise tests were performed once in omnidirectional mode and once in fixed directional mode for those three devices, which feature a multi-microphone noise reduction system (cf. Table 1).
Subjects Twenty adult, unilaterally implanted, German-speaking Baha users participated in each of the two studies. All subjects had a bilateral conductive or mixed hearing loss. As different populations participated in the Divino-BP100 comparison from those in
Figure 3 Speech understanding of speech coming from the front in noise from the rear. The differences between the signal-to-noise ratio (SNR) in dB for 50% speech understanding between basic digital signal processing (Divino and Intenso) and complex digital signal processing (BP100 and BP110) are shown. Speech understanding is improved by +2.3 to +5.1 dB with complex processing (**P < 0.01).
Fig. 1 shows the results of the speech understanding tests in quiet at 65 dB. For both groups of participants, there is a substantial improvement in word understanding from the unaided condition to either of the two aided conditions. The difference ranges from 33.8 to 55.9% points and is statistically highly significant (P < 0.001; Wilcoxon matched pairs signedrank test). In addition, there is a smaller, statistically non-significant increase from the devices with a basic digital signal processing to the complex processing by 12.7% points for the BP100 vs. the Divino and by 1.3% points for the BP110 power vs. the Intenso. Fig. 2 shows the differences between devices with basic and complex signal processing for speech reception in noise, if speech and noise are emitted by the same loudspeaker in front of the listener (S0N0). In all four comparisons, there is a small gain of +0.7 to +1.2 dB in terms of signal-to-noise advantage. However, only the largest of these improvements is statistically significant. Also, the choice of the directional or the omnidirectional microphone has no influence, as the noise and the target signal sources are not spatially separated. Fig. 3 shows the results for speech from the front and noise from the rear (S0N180). There is a considerable (+2.3 to +5.1 dB) and statistically significant (P ≤ 0.0032) improvement for the sound processors with the complex digital signal processing. For the Divino/BP100 comparison, the advantage of the BP100 is the same for the directional or the omnidirectional mode, suggesting that the directional benefit is similar in both devices. For the Intenso/BP110 comparison, the difference is much larger in the directional mode, as the Intenso does not have this feature and all measurements had to be done in the omnidirectional mode.
Discussion and summary The studies presented suggest that complex digital signal processing is indeed beneficial for Baha users, when compared with basic digital signal processing. The benefit varies with the test setting. For speech in quiet, there may be an advantage, but it is relatively small and not statistically significant in
Cochlear Implants International
2014
VOL.
15
NO.
S1
S49
Kompis et al.
Is complex signal processing for bone conduction hearing aids useful?
our two groups of subjects. For speech in noise, where both speech and noise arrive from the same loudspeaker, there is, again, a small but consistent advantage in the order of magnitude of 1 dB, which is only for one comparison statistically significant. This is probably a beneficial effect of the multiband adjustable gain and compression, which allows, among others, a higher gain the high-frequency region (Pfiffner et al., 2011a; Kurz et al., 2013). In contrast, for speech from the front and noise from the rear, the benefit is much larger (+2.3 to +5.1 dB) and statistically significant for both pairs of speech processors. The position compensation feature, which is a small directionality feature present only in the devices with complex signal processing, is most probably the reason for the greater benefit even in the omnidirectional mode. Note that the Baha Intenso has no directional mode, thus the benefit is largest (+5.1 dB) when compared to the directional setting of the BP110 power. Very recently, another research group has published another study comparing the basic and complex signal processing in the Baha Intenso and the Baha BP110 power (Desmet et al., 2013). Their results are similar to those presented here: a small benefit of the device
S50
Cochlear Implants International
2014
VOL.
15
NO.
S1
with complex signal processing in quiet and in noise from the front, and a larger gain for noise from the rear, suggesting that complex signal processing can indeed be beneficial for Baha user.
References Desmet J.B., Bosman A.J., Snik A.F., Lambrechts P., Hol M.K., Mylanus E.A., De Bodt M., Van de Heyning P. (2013) Comparison of Sound Processing Strategies for Osseointegrated Bone Conduction Implants in Mixed Hearing Loss: Multiple-Channel Nonlinear Versus SingleChannel Linear Processing. Otology and Neurotology, 34(4): 598–603. Kurz A., Caversaccio M., Kompis M. (2013) Hearing performance with 2 different high-power sound processors for osseointegrated auditory implants. Otology and Neurotology, 34(4): 604–610. Pfiffner F., Caversaccio M., Kompis M. 2011a. Comparisons of Sound Processors Based on Osseointegrated Implants in Patients With Conductive or Mixed Hearing Loss. Otology and Neurotology, 32: 728–735. Pfiffner F., Kompis M., Flynn M., Åsnes K., Arnold A., Stieger C. 2011b. Benefits of Low Frequency Attenuation of Baha® in Single Sided Sensorineural Deafness. Ear and Hearing, 32(1): 40–45. Pfiffner F., Kompis M., Stieger C. 2009. Bone Anchored Hearing Aids: Correlation between pure tone thresholds and outcome in three users groups. Otology and Neurotology, 30: 884–890. Wagener K., Brand T.K.B. 1999. Entwicklung und Evaluation eines Satztestes für die deutsche Sprache Teil III: Evaluation des Oldenburger Satztestes. Audiologic Acoust, 38: 86–95.
