Download Frequency Compression through Beltone Sound Shifter

Frequency Compression through
Beltone Sound Shifter
Tammara Stender, Au.D.
Astrid Haastrup, M.A.
Hearing loss that slopes to greater degrees in the high frequencies is very common; however, it often presents
The success of this approach often depends on the sever-
Of the frequency lowering approaches used today by
problems with individuals’ ability to discriminate speech. In some cases, today’s hearing instruments may not
ity of the hearing loss in the high frequencies; for greater
the hearing instrument industry, the most commonly fit-
be able to provide adequate audibility for these high frequencies. Frequency lowering strategies have been
degrees of high-frequency hearing loss, traditional amplifi-
ted strategy is frequency compression. Just as WDRC
used in the hearing instrument industry as an alternative for clients who do not benefit from traditional amplifi-
cation may have limited success in providing audibility for
and other types of amplitude compression “move” low-
cation in the high frequencies. However, many of these strategies also introduce significant distortion by means
these high frequency sounds.
intensity sounds to higher intensity levels for the individual,
of this feature. Beltone Sound Shifter is a new frequency lowering approach that improves high frequency audibility, with minimal degradation of sound quality.
INTRODUCTION
frequency compression “moves” high-frequency sounds
There are a number of reasons why a traditionally-fitted
to a lower, more aidable region. The ultimate goal of both
hearing instrument may not be able to effectively amplify
amplitude and frequency compression is the same: audi-
more severe degrees of high frequency hearing loss. The
bility. To this end, Beltone introduces its new frequency
most common culprits include feedback from high gain
compression feature: Sound Shifter.
Hearing loss that slopes to greater degrees of severity for
The loss of certain phonemes produced at typical speech
requirements in this frequency region, and limitations of
the high frequencies is commonly encountered in hearing
levels may present significant problems for the client. For
the receiver. The physical properties of larger receivers,
clinics. This common hearing loss configuration, however,
example, the /s/ phoneme is necessary in the English lan-
which are used in high power hearing instruments, result in
Beltone Sound Shifter,
Revealed
presents its own challenges, both for the client and for the
guage to distinguish plurality (e.g., “bag” versus “bags”).
a lower-frequency roll-off than what is observed for lower-
In addition to the shared goal of audibility, Beltone Sound
hearing care professional. For the client, many consonant
The sentence “Please bring the bags” will have a different
power hearing instruments.
Shifter employs much of the same terminology as ampli-
sounds important for accurate speech discrimination lie in
meaning to the client if she only hears “Please bring the
the high frequency region. For example, the /s/, /th/ and
bag_”.
/f/ phonemes may be rendered inaudible with only a mild
tude compression. Frequency compression changes the
Aside from the technical issues in providing this large
input to a more amenable output above a cut-off thresh-
amount of high-frequency amplification, there may be a
old, or knee point. The degree of frequency compression
degree of high frequency hearing loss. Figure 1 shows
A hearing loss such as the one shown in Figure 1 affects
physiological issue to consider as well. In some cases,
applied can be expressed as the frequency compression
a bilateral steeply sloping high frequency hearing loss. A
the audibility for environmental sounds as well as speech.
providing high frequency amplification for these severe
ratio. Sound Shifter compresses the frequencies above the
speech banana is superimposed on the audiogram, which
For this client, environmental sounds such as birdsong will
hearing losses may not be beneficial. It has been sug-
knee point, effectively moving them closer together. The
indicates the common distribution of speech sounds in the
be inaudible without some form of high frequency amplifi-
gested that the corresponding high frequency areas on
input frequencies are amplified in a narrower frequency
frequency-intensity realm over time (aka, the long term av-
cation (Figure 2).
the cochlear basilar membrane may represent a “dead
region that is more usable to the hearing instrument user.
erage speech spectrum, or LTASS). Individual phonemes
region,” characterized by non-functioning inner hair cells.
indicate on the audiogram their specific frequency and
Applying amplification to these “dead regions” may result
Like amplitude compression, an input/output function can
intensity ranges in everyday speech. Sounds below the
in poorer speech understanding.
be used to describe how Sound Shifter works (Figure 3).
1
hearing thresholds on the figure are audible to the client,
However, instead of plotting input and output levels, inIf adequate gain cannot be provided by the compressor,
amplification, the /s/, /th/, /f/ and /k/ phonemes are inau-
or if the client does not receive benefit in speech under-
dible for this individual client.
standing from this degree of high-frequency amplification,
an effective alternative to provide audibility may be a freFigure 2. Steeply sloping high frequency hearing loss, with typical
environmental sounds superimposed to indicate areas of audibility
and inaudibility for the client.
quency lowering approach. Frequency lowering strategies
work through the displacement of signals from higher frequencies to lower frequencies, where beneficial amplifi-
put and output frequencies are used. The blue part of the
High
Output Frequency
and sounds above the thresholds are inaudible. Without
No Sound Shifter (“Off”)
Compressed region
Uncompressed
region
cation may be achieved. Although all frequency lowering
Figure 1. Steeply sloping high frequency hearing loss, with the
speech banana (LTASS) and English phonemes superimposed to indicate areas of audibility and inaudibility for the client.
Traditional amplification using wide dynamic range com-
approaches introduce some degree of distortion due to
pression (WDRC) provides greater audibility for low-level
the artificial movement of some frequencies to a different
sounds in the necessary frequency regions, in an at-
region, the benefit in audibility may result in a favorable
tempt to raise the level of speech and other environmental
compromise for some individuals.
sounds to an audible level per the individual’s hearing loss.
Frequency
Compression
Ratio
Frequency
Compression
Knee Point
Low
Low
Input Frequency
High
Figure 3. Input/output function describing the relationship between
input and output frequencies with Beltone Sound Shifter. Frequency
compression occurs only above the frequency compression knee
point. Beneath the knee point, no frequency compression is performed.
curve is the frequency region in which Sound Shifter is not
loss may not distinguish a sound quality difference for a
and blue indicating the least intense. Both hearing instru-
compression knee point as close to 2 kHz as possible,
active. The frequency relationships here are not changed
range of frequency lowering settings.2 Sound quality for
ments were programmed with amplitude compression for
with the same 2:1 compression ratio as used by the “mod-
by Sound Shifter. The purple part of the curve is where
music is generally preferable for normal hearing and hear-
a mild-to-moderate hearing loss, with similar frequency
erate” setting.
Sound Shifter is active, and demonstrates how the input
ing impaired listeners without frequency lowering, or with
compression knee points and ratios.
frequency corresponds to a lower output frequency. The
up to moderate frequency lowering settings. Results and
input frequency at which frequency compression begins
considerations from the literature were instrumental in de-
Panel A shows the signal processed through Beltone
60 dB SPL (RMS) were recorded through each hearing
is the knee point. The relationship between the input fre-
veloping Sound Shifter and its settings, in order to provide
WARP 17 amplitude compression alone, and then with
instrument for each of these hearing losses. The record-
quency and the output frequency in the purple part of the
the desired audibility for high frequency sounds with the
Sound Shifter activated. In Panel B, the same signal is
ings were then analyzed through the Hearing Aid Speech
curve is expressed as the frequency compression ratio.
least amount of impact on the sound quality.
processed first through the competitor’s wide-dynamic
Quality Index (HASQI),4 which measures the “coherence”
range amplitude compression, and then also through the
between the input and output signals of the hearing aid. A
Identical 10-second segments of pop music presented at
3
Sound Shifter was Designed
with Sound Quality in Mind
To provide maximal listening benefits with minimal sig-
competitor’s non-proportional frequency compression.
higher HASQI score predicts better sound quality, as the
nal distortion, Sound Shifter is a proportional frequency
The frequency compression knee points are shown as
output sound will be more similar to the input sound, with
Frequency lowering is a unique hearing instrument fea-
compression strategy. This approach was chosen over a
the point where the slope of the response changes. (Note
less added noise and distortion relative to the original input
ture in that there is no general consensus in the literature
non-proportional frequency compression technique, cur-
there no change in slope for the response when no fre-
sound. The HASQI has been found to be a valid measure
about its benefits to groups of people. Some individuals
rently offered by another hearing aid manufacturer. To illus-
quency compression is activated).
of subjective sound quality judgments by people with nor-
will benefit from it while others will not. In addition, there
trate the differences in the processing of a signal, a 90 dB
is no clear way to identify prior to the fitting which clients
SPL pure tone sweep was processed through Beltone’s
The results for Beltone Sound Shifter show a clean re-
will derive benefit from the frequency lowering approach.
proportional frequency compression and the competitive
sponse; minimal differences in energy were present in the
Results for all three types of hearing losses showed the
Finally, the effect of this type of processing on sound qual-
non-proportional frequency compression to create spec-
output above the frequency compression knee point as
same trends; for simplicity, the average of these results
ity is not generalizable. Individuals with milder or normal
trograms (Figure 4). Each spectrogram is an input/output
compared to the output below the knee point, where fre-
for each frequency compression setting are depicted in
hearing thresholds may distinguish a difference in sound
function, showing the input frequency on the x-axis and
quency compression was not active. In contrast, the re-
Figure 5. The highest HASQI scores relative to the input
quality with frequency lowering as compared to traditional
the output frequency on the y-axis. Colors indicate the in-
sults for the competitive frequency compression feature
signal were obtained when frequency compression was
processing, while those with greater degrees of hearing
tensity of the signal, with red indicating the most intense
were not as clean, as indicated by the jagged response
deactivated (the “off” setting). For both frequency com-
above the frequency compression knee point. This is due
pression techniques, the HASQI score was inversely re-
to the non-proportional relationship of the competitor’s fre-
lated to the aggressiveness of the algorithm. This indi-
quency compression. More distortion products are gener-
cates that stronger frequency compression settings are
ated by this approach. These distortion products manifest
predictive of poorer sound quality. When proportional and
themselves on the spectrogram with a blurred quality of
non-proportional frequency compression were compared,
the response curve above the frequency compression
however, the predicted sound quality of Sound Shifter was
knee point.
better.
Sound Shifter On
Output Frequency (Hz)
Output Frequency (Hz)
PANEL A
Sound Shifter Off
Input Frequency (Hz)
Input Frequency (Hz)
mal hearing and with mild to moderate hearing loss.
Sound Quality for Music
0,8
Sound quality for music was also evaluated for propor-
0,7
tional and non-proportional frequency compression. In this
comparison, each hearing instrument was programmed
for 3 hearing losses: mild flat, mild sloping to moderatelysevere, and mild sloping to profound. Gains were set according to the default prescriptive targets in the respective
manufacturer’s fitting software. Frequency compression
was set to “off,” “moderate” and “strong.” The “moderate”
Input Frequency (Hz)
Figure 4. Spectrograms for a swept pure tone were measured with
conventional amplitude compression (left) and with both amplitude
and frequency compression activated (right). Panel A shows the results obtained with Beltone Sound Shifter. Panel B shows the results
Input Frequency (Hz)
for a competitor’s hearing instrument. These spectrograms show the
difference in distortion generated by Beltone’s proportional frequency
compression strategy, compared to that generated by the competitor’s non-proportional frequency compression (FC) approach.
setting had a frequency compression knee point as close
to 3 kHz as possible, and a compression ratio as close
to 2:1 as possible. The “strong” setting had a frequency
0,6
HASQI Score
Competitor FC On
Output Frequency (Hz)
Output Frequency (Hz)
PANEL B
Competitor FC Off
0,5
0,4
0,3
0,2
0,1
0
Off
Moderate
Strong
Frequency Compression Setting
Sound Shifter
Competitor FC
Figure 5. Sound quality for music, as predicted by the HASQI, revealed better
sound quality preservation for Sound Shifter than for a non-proportional frequency
compression strategy.
Development of Sound
Shifter Settings
Objective measures during the fitting appointment includ-
sults of other studies using frequency compression, group
How to Fit Sound Shifter
ed real ear insertion gain (REIG) measurements for Ling
data did not indicate optimal frequency compression set-
Sound Shifter settings can be found in the “Features” sec-
All frequency lowering techniques introduce a compromise
sounds /s/ and /sh/. Forty ears were evaluated for each of
tings nor average benefit. Also as expected, certain indi-
tion of Solus Pro (Figure 8). As need for frequency com-
between audibility and sound quality. Some users will ben-
the 8 frequency compression settings plus the “off” condi-
viduals did obtain improved outcome measure results and
pression is not easily generalizable to particular hearing
efit greatly from the increased audibility of high-frequency
tion. Results revealed 3 distinct grouping of REIG meas-
expressed preference for frequency compression process-
losses or client characteristics, the feature is set to a de-
sounds, while others may be bothered by the feature’s in-
urements among the conditions (Figure 6).
ing, although variability was observed. Figure 7 shows the
fault “off” setting in each program.
herent distortion and perceptual sound quality. As previ-
distribution of individual scores with and without Sound
Effect of Different Settings on Ling /s/
ously stated, there are no defined guidelines about which
lowering. Further, there is no “cookbook approach” to
fitting the technology. For instance, two individuals with
similar hearing losses may derive the greatest benefit from
different frequency compression knee points and ratios, or
Insertion Gain (IG)
individual clients will be the best candidates for frequency
from no frequency lowering at all. For these reasons, the
best approach to applying and fitting frequency compres-
Shifter for the UWO Plurals Test. Subjects who obtained
55
50
45
40
35
30
25
20
15
10
5
0
-5
-10
-15
-20
positive benefit performed better with Sound Shifter on
off
5000_1.33
4000_1.33
4000_2
3500_2
3000_2
2500_2
2250_2
2000_2
100
sion is often a conservative one.
1000
Frequency (Hz)
10000
this test than with conventional amplification. Two of the
16 subjects achieved better scores with conventional
amplification than frequency compression, which is indicated by the negative benefit. Test subjects who preferred
Sound Shifter processing over conventional amplification
indicated that the sound quality was crisper and clearer.
Benefit Obtained with Sound Shifter
5
Effect of Different Settings on Ling /sh/
found through internal testing to provide the greatest audibility benefit while also minimizing deleterious effects on
the sound quality. Results of initial testing indicated that
the compression frequency knee point had a greater effect
on the feature than the compression ratio.
Insertion Gain (IG)
designed with a limited number of settings, which were
Figure 8. Sound Shifter can be activated and set in the “Features” section.
4
off
5000_1.33
4000_1.33
4000_2
3500_2
3000_2
2500_2
Number of Subjects
As such, Sound Shifter in Solus Pro fitting software was
55
50
45
40
35
30
25
20
15
10
5
0
-5
-10
-15
-20
3
To activate Sound Shifter, choose a setting in the dropdown menu. The recommended setting based on the
2
client’s audiogram will be indicated in “bold” font style.
1
When frequency compression is not expected to provide
2250_2
2000_2
100
1000
Frequency (Hz)
10000
0
-2
-1
0
1
2
3
4
5
6
7
UWO Plurals Test Score With Sound Shifter Minus
Score Without Sound Shifter
To determine the most effective and differentiated com-
optimal results for the client (e.g., flat hearing losses, noise
notches, reverse slope hearing losses), the recommended
setting will be “Off.” The criteria for recommending “Mild”,
Figure 6. REIG measurements among frequency compression settings for the /s/
and /sh/ phonemes.
Figure 7. Most individuals obtained improved scores on the UWO Plurals Test when
Sound Shifter was activated.
“Moderate” and “Strong” settings are as follows:
quency compression settings in addition to the “Off” set-
For the field trial component of this testing, frequency
The overall conclusion from these tests was that the fitting
• A “Mild” setting is recommended if the audiogram has
ting (Table 1). Seventeen test subjects with steeply slop-
compression settings were determined for 20 of the 37
process could be simplified to a smaller number of set-
a slope of 10 dB or greater per octave frequency, and
ing hearing loss and 20 with severe-to-profound hearing
participants based on a real ear measurement protocol to
tings. Since REIG measurements revealed three grouped
the slope begins at 4000 Hz or higher (Figure 9).
loss participated in the trial.
optimize audible bandwidth. The remaining 17 participants
responses across the eight settings evaluated, and more
• A “Moderate” setting is recommended if the audiogram
were fitted according to which frequency compression
than one setting was found to provide comparable benefit
has a slope of 10 dB or greater per octave frequency,
setting enabled them to perform best on the UWO Plu-
for individuals, three Sound Shifter settings were derived:
and the slope begins at 2000 Hz (Figure 10).
rals Test. If equivalent scores were obtained for more than
“Mild,” “Moderate” and “Strong.” Table 2 shows the cor-
• A “Strong” setting is recommended if the audiogram
one setting, the most conservative setting (i.e., the highest
responding frequency compression knee points and ratios
has a slope of 10 dB or greater per octave frequency,
compression knee point and lowest ratio) was chosen.
for each of these settings.
and the slope ends at 2000 Hz (Figure 11).
pression knee points and ratios, a second internal test was
conducted. The goal of this trial was to evaluate 8 fre-
Test
Setting
Frequency Compression
Knee Point
Frequency Compression
Ratio
”Off”
1 (weakest)
2
3
4
5
6
7
8 (strongest)
n/a
5000
4000
4000
3500
3000
2500
2250
2000
n/a
1.33
1.33
2
2
2
2
2
2
Table 1. Eight frequency compression settings evaluated during the internal trial.
5
Post-field trial measures included clarity ratings, sound
quality ratings for male and female speech and music,
speech testing, and the Speech, Spatial and Qualities of
Hearing Scale (SSQ)6. As was expected, based on the re-
Sound Shifter Frequency Compression
Setting
Knee Point
Mild
Moderate
Strong
4000 Hz
3500 Hz
2500 Hz
Frequency Compression
Ratio
1.33
2
2
Table 2. Frequency compression knee points and ratios for each setting of Sound
Shifter.
REFERENCES
1. Moore B. Dead Regions in the Cochlea: Conceptual
Foundations, Diagnosis, and Clinical Applications. Ear
and Hearing, 2004; 25 (2), 98-116.
2. Parsa V, Scollie S, Glista D, Seelisch A. Nonlinear frequency compression: effects on sound quality ratings
of speech and music. Trends in Amplification. 2013;
17(1): 54-68.
Figure 9. Example of audiometric
thresholds when a “Mild” Sound Shifter setting is recommended.
3. Bentler R. Hearing aids: Clarity or controversy? Audiology Online Hearing Aid Conference, 2012. Course
Figure 12. When Sound Shifter is turned on, a solid vertical line and a shaded region
indicate the range of frequencies that are compressed.
21332.
Upon choosing a Sound Shifter setting, a solid dark verti-
4. Kates JM, Arehart KH. The Hearing-Aid Speech Quality
cal line and a gray area will appear in the graph. The verti-
Index (HASQI). Journal of the Audio Engineering Soci-
cal line indicates the frequency compression knee point,
ety, 2010; 58(5), 363-381.
and the gray area denotes the frequency range that is
compressed (Figure 12).
5. Glista D, Scollie S. Development and evaluation of an
English language measure of detection of word-final
Recommendations for Sound Shifter are specific to the
plurality markers: The University of Western Ontario
ear’s audiometric thresholds. It is possible that different
Plurals Test. American Journal of Audiology, 2012. 21:
Sound Shifter settings for each ear may be recommended
76-81.
in the case of asymmetric hearing losses. Further, Sound
Figure 10. Example of audiometric
thresholds when a “Moderate” Sound
Shifter setting is recommended.
Shifter may be activated in one or more programs, binaurally or monaurally, per the individual’s needs.
Summary
Sound Shifter provides yet another tool for hearing care
professionals to improve high frequency audibility, especially for individuals who do not receive optimal audibility through traditional amplification. For certain individuals,
Sound Shifter can improve audibility for high frequency
speech sounds, such as /s/ and /th/, while maintaining
sound quality to a greater degree than can be offered by
a non-proportional frequency compression strategy. Research revealed that a limited number of settings were appropriate to fit Sound Shifter for a wide range of individuals
who may benefit from this feature. Thus, Beltone Sound
Figure 11. Example of audiometric
thresholds when a “Strong” Sound
Shifter setting is recommended.
Shifter is an option that enables the hearing care professional to better accommodate the needs of individual clients.
6. Gatehouse S, Noble W. The speech, spatial and qualities of hearing scale. Int J of Aud, 2004; 43(2):85-99.
M200993-GB-13.09-Rev.A
Worldwide Headquarters
Beltone A/S
Lautrupbjerg 7
DK-2750 Ballerup
Denmark
Tel.: + 45 45 75 11 11
Fax: + 45 45 75 11 19
www.beltone-hearing.com