Home' Accord : Accord March 2017 Contents Feature 9
of hearing. Weaker parts of speech such as the consonants
are turned up so that they are audible, while the louder
parts of speech such as the vowels are turned down so that
they are within comfortable loudness levels. Some form
of compression is also used to keep the amplified signal
below the saturation limit of the device to avoid distortion.
Essentially, with WDRC, soft and loud sounds are kept within
the working auditory space (dynamic range) of the user and of
the device. However, there is a downside to WDRC; it always
distorts the signal it compresses (Souza, 2002). Despite
this, designers of amplification systems have almost entirely
focused on achieving audibility. Maximizing speech fidelity
was hardly a top priority. For the users of hearing devices, this
basically meant accepting distortion as an inevitable side
effect of the amplification. An alternative to wide dynamic
range compression is linear gain whereby equal amounts of
gain are applied to a range of inputs. Linear gain processing
was the amplification strategy of choice before wide
dynamic range came into being in hearing devices. It has a
distinct advantage of preserving the amplitude variations
inherent in the input speech signal as much as possible.
In other word, there is minimal signal distortion with linear
amplification (Arehart K, Kates J and Anderson, 2010).
Another key parameter that is often overlooked or given much
less attention is the speed at which the amplification system
responds to fluctuations in the input signal. This speed is
commonly referred to as time constants in clinical audiology.
A fast response system can make softer sounds that follow
an intense sound in speech audible but it has a tendency
to distort the speech signal through frequent gain change
(Souza, 2002; Stone and Moore, 2007). A slow response
system helps to preserve the amplitude/time structure of
the original signal in the longer term by keeping the gain
constant, and is less distorting to the speech signal (Plomp,
1988). Overall, better speech fidelity can be achieved with
a slow response system (Schum and Sockalingam, 2010).
Achieving the required audibility with minimal distortion
is a huge challenge from a signal processing standpoint.
It requires an amplification system that is able to compute
the right amount of gain and the right response speed for a
given input at a given point in time. For developers of Oticon
Medical’s Ponto sound processors however, this challenge
is now a thing of the past. Ponto sound processors feature
a novel amplification system called Speech Guard that is
designed to achieve just that audibility without the level
distortion that is commonly associated with WDRC systems.
The Structure and Function of Speech Guard
Speech Guard, as its name suggests, is designed to
“protect” those intensity variations naturally inherent in
speech as much as possible. The key to achieving this lies
in the time constants Speech Guard uses to respond to
fluctuations in the input signal. Speech Guard comprises
two systems constantly interacting with each other and
responding to fluctuations in the input signal in real time.
It is a highly adaptive system, capable of responding either
slowly or quickly depending on changes in input level. There
is essentially a slow response system and a fast response
Figure 1: Components of the Speech Guard Amplification System
Figure 1 shows the two systems in Speech Guard working in tandem to respond at appropriate speeds
to fluctuations in the input signal. The ultimate goal of these two systems is to ensure that the dynam-
ics of the output speech signal resembles that of the input signal as much as possible.
SLOW RESPONSE SYSTEM
FAST RESPONSE SYSTEM
• Monitors input and updates gain slowly
• Keeps processing near linear protecting signal fidelity
• Monitors rapid & large level changes in input
• Responds with instantaneous, fast gain adjustment
Bone anchored sound processors have been used successfully for many years now, but with technology
that lags behind that of modern air conduction hearing instruments. With continual developments in signal
processing technologies in air conduction hearing instruments, there is a strong need for incorporating
newer technologies into bone anchored sound processors so that patients wearing these processors can
benefit from their superior performance to the same extent that wearers of advanced air conduction hearing
instruments currently do. An amplification scheme that is designed to minimize distortion of speech
signal even during the occurrence of a sudden extraneous noise signal has been shown to benefit users
of Oticon’s advanced hearing instruments in terms of speech understanding and sound quality (Bruun
Hansen et al., 2010). This amplification scheme is known as Speech Guard. Given the reduced dynamic
range of hearing of wearers of bone anchored sound processors and the relatively small dynamic range of
the processors themselves, Speech Guard avoids the common pitfall of compression that is widely used in
today’s hearing aids and bone anchored processors: signal distortion. This paper describes how Speech
Guard is designed to maintain the amplitude variations between sounds and preserve the natural details
and nuances of speech.
The heart of any hearing device is its amplification system.
Though hearing devices are often simply considered to be
amplifiers, there is a lot more than simple amplification that
goes on in the amplification system. A complex series of
signal processing takes place to ensure that the signal that
reaches the user is audible, comfortable and intelligible.
A fundamental task of the amplifier is to respond to
intensity fluctuations in the input signal in such a way that
the integrity and fidelity of the output signal is maximally
preserved. Achieving this is rather challenging for designers
of hearing devices as any signal processing that is employed
to maximize audibility is bound to result in some sort of
distortion. While there is no arguing that the audibility is the
basic fundamental function of any hearing device, audible
distortion in the signal can potentially influence the way
end users accept and use hearing devices in the long term.
The Mark Track Survey (Kochkin, 2010) has clearly revealed
that there is more to patient satisfaction than merely
making sounds audible. Just how natural or undistorted
the signal from the device is can significantly impact
patient satisfaction. There is also a cognitive argument
for preserving the natural dynamics of speech as much
as possible. It has been shown that a natural undistorted
signal is the easiest for the brain to process (Lunner, 2010,
Behrens, 2010). It is quite clear that preserving the integrity
of the input signal is a key to achieving patient satisfaction
and long term use.
There are several parameters of amplification that have to be
taken into account when designing an amplification system.
For one thing, the amount of gain prescribed for signal
inputs across a range of frequencies is a key consideration.
Typically, for sensorineural hearing loses, this is achieved
through wide dynamic range compression (WDRC). WDRC is
employed to varying extents in almost all hearing devices to
squeeze a range of inputs into the patient’s dynamic range
Minimizing Signal Distortion through
a Novel Amplification Scheme in
Bone Anchored Sound Processors
Ravi Sockalingam, PhD, FAAA, Aud (C)
Director of Clinical Research and Professional Relations, Oticon Medical
Links Archive Accord December 2016 Accord June 2017 Navigation Previous Page Next Page