HEARING AID SYSTEM USING

MATLAB
Rahil Arora 17BEC0817 E-mail: [email protected]

Kumaran Karthikeyan 17BEC0732 E-mail: [email protected]

P Dinesh Kumar 17BEC0415 E-mail: [email protected]

Naveen Kumar 17BEC0206 E-mail: [email protected]

Gurubalaji Ramachandran 17BEC0050 E-mail: [email protected]

Abstract: Traditional analog hearing aids look like a belong to the hearing aid associated with the hearing aid
simple radio. They can be adjusted for volume, bass and support, customer dissatisfaction with equipment is not
treble. But hearing loss is not just a loss of technical meet expectations and costs new digital versions of hearing
volume. Rather hear deficiency can increase sensitivity and aids Hearing protection is typically measured when moving
reduce tolerance to certain sounds while decreasing the auditory threshold relative to normal ears for some with
sensitivity to other. For example, digital technology can hearing loss. However, only a part use hearing aids. This is
make the difference between speech noise and background due to a number of factors that belong to the hearing aid
noise, allowing one while filtering the other. About 10% of associated with the hearing aid support, customers'
the world's population suffers from some type of hearing dissatisfaction with equipment is not meet expectations and
loss, but only a small percentage of this statistic uses a costs new digital versions of hearing aids . Hearing
hearing aid. Stigma associated with the wearing of a protection is usually measured when moving
hearing aid, customer dissatisfaction with the performance the auditory threshold compared to normal ears
of that device and The costs associated with a high-
performance solution are all causes of low market Classification of Hearing Loss
penetration. Through the With digital signal processing, the Hearing level
digital hearing aid now offers what the analog hearing aid
can not offer. It proposes the possibility of improving the Normal hearing- -10 dB – 26 dB
signal-to-noise ratio, the flexible treatment of the gain, the
reduction of return, etc. In this article the simulation of a mild hearing loss 27 dB - 40 dB
simple digital hearing aid has been developed using
MATLAB programming language. Setting up this
configurable digital hearing aid (DHA) The system Moderate hearing 40 dB - 70 dB
includes the noise reduction filter, the frequency shaping loss
function and the amplitude compression function. This
digital hearing system is designed to adapt to patients with severe hearing loss 70 dB - 90 dB
mild or moderate hearing loss since different gains can be
defined to map different levels of hearing loss. profound hearing 90dB
greater than loss

Keywords: Hearing aid, Gaussian Noise, Signal Table 1: Different degree of Hearing Loss
Processing, Fourier transform, Signal to Noise Ratio.
Basically, all hearing aids used analog technology for
I. Introduction sound processing. Improvements have been made using the
development of digital technology sound processing for the
Hearing aids are one of the most important problems for effectiveness of hearing aids. Nowadays, digital hearing
humans. They are a small electronic instrument that makes aids are small, which means can be hidden inside the ear
the sound louder and makes speech easier to hear and and have almost perfect sound reproduction. Research on
understand. The hearing aid is designed to capture sound digital hearing aids has been growth and now a small
waves with a tiny microphone, turn the weaker sounds into programmable computer are able to amplify millions of
louder sounds and send them to the ear by a small speaker. different sounds signals had been built into the devices,
With the microchips available today, hearing aids have improve the hearing ability of the hearing impaired
become smaller and their quality has improved people. The first digital hearing aids were launched in the
considerably. About 10% of the world's population bears mid-80s, but these early models were slightly not very
bears some with hearing loss. However, only part practical. Ten years later, the digital audience the helpers
use hearing aids. This is due to a number of factors that
were really successful, with small digital devices placed Noise Reducing Filter
inside or discreetly behind the ear. Today, digital Major anxiety for the hearing impaired is the ability of a
technology is really part of everyday. Most households hearing aid to differentiate speech signal in a noisy
have a variety of digital products, such as phones, video environment. Therefore, eliminate noise, a reduction filter
recorders and personal computers. Hearing aids were also function is used in this design. To suppress the noise in the
benefited from the emergence of digital technology. signal, The wavelet filtering function is used.
Among the advantages of digital signal processing
this allows hands-free use. Ugly automatically adjust the b. Frequency Shaper
volume and height on it's own. It makes thousands of
adjustments per second which causes a reduction of the A major complaint of hearing aid users is that the hearing
background noise, better listening in noisy situations, his aid amplifies all signals rather than the significant signal
quality and multiple settings of the program [4]. The user that they wish to hear. More hearing impaired has difficulty
can switch between the varieties of programs for different hearing loud frequency signal. Therefore, the formatting
listening situations. frequency is designed to correct hearing loss to some
frequencies. It applies a high gain for higher frequencies
and vice versa. The typical transfer function of the
II. Methodology frequency amplifier is shown in Figure 2.
Below is a block diagram for the MATLAB c. Amplitude Compression
implementation of Digital Hearing Aid System. The input
speech signal takes the form of human voice. The input Basically, the amplitude compression function is the task of
speech signal will pass through several functions i.e noise controlling the overall gain of a speech amplification
addition, noise reduction filter, frequency shaper and system. The amplitude compression will be ensure that the
amplitude compression before producing an adjusted output amplified signal does not exceed saturation power. The
speech signal which is audible to the hearing impaired saturation power is where the sound signal begins to
person. become uncomfortable.

a. Noise summation

As the input speech signal for this system is a clear signal,

noise is added to simulate a real situation. In this system,
adaptive white Gaussian noise (AWGN) and random noise
are added to the input speech signal using MATLAB a
function. Noise (AWGN) has a continuous and uniform
frequency spectrum over a specified period frequency band
and has an equal power per Hertz of this B: and. It consists
of all frequencies at the same intensity and has a normal
probability density (Gaussian) a function.

Figure 2: Typical Frequency transfer function

III. Implementation & Simulation

The code, written in MATLAB, loads the input wave

signal, takes the sampling frequency and the number bits of
this signal. Next, adaptive white Gaussian Noise (AWGN)
and random noise are added to the signal before they are
processed by various MATLAB function for audible output
to the person with hearing loss. For the purposes of
analysis, a sample of speech signal is selected. The sample
is a man speaking: "I am Mohd Syahril Nizar, student in
last year at the UIA " This signal is added by Adaptive
Figure 1: System Block Diagram White Gaussian Noise (AWGN) and random noise.
For simplicity, a graphical user interface (GUI) has been
created built to make this digital hearing system work demo How the hearing instruments work
simulation. To run the demo successfully, you have to
necessary to enter all the parameters that are maximum Digital makes a difference
gain to apply, saturation power and four frequency values
where the gain changes. Figure 3 below shows the The revolutionary digital signal processing revolutionary
graphical interface of this system. hearing aids enable scientists and manufacturers to write
intelligent software and develop sophisticated algorithms
that lead to new benefits such as:

• Understanding better speech in noisy

environments
• Increased profit without any feedback
• Increased listening comfort and speech detection
Ability to customize instrument settings to meet the user's
need
More detailed orientation features replaced the amplifier
and,
In their simplest form they think miniature hearing
instruments with four basic components:

• Microphone
• Amplifier
• Speaker (Receiver)
• Power supply (batteries)
Figure 3: GUI of digital Hearing Aids Whatever style or size, all hearing aids consist of these four
parts.
In this simulation, a sample of hearing impaired patients is
obtained from "Jabatan Audiologi dan Sains Pertuturan,
Microphones and receivers are converters, that is, they
Fakulti condemns Kesihatan Bersekutu, Universiti
convert energy from one form to another. The microphone
Kebangsaan Malaysia ". This patient suffers from moderate
collects acoustic energy (sound) and converts it into an
hearing loss characterized by:
electrical signal. The receiver collects electrical signals
- Hearing threshold at 40 dB. from the amplifier and converts them to acoustic energy
(sound).
- Threshold of pain at 90 dB.
Between the microphone and the receiver, the amplifier
- You have difficulty hearing high frequencies increases the amplitude of the signal transmitted by the
microphone before transmitting it to the receiver that sends
A. Equations and Formulae Used for FFT:
it to the inner ear.
The resulting sequence is interpreted as follows:

X[k]=∑𝑁−1
𝑛=0 𝑥[𝑛]𝑒
−𝑗2𝜋𝑛/𝑁

The zero frequency corresponds to n=0, positive

frequencies 0<f<Fs/2 correspond to values 1<=n<=N/2-1
while negative frequencies correspond to N/2+1<=n<=N-1.
Here, Fs denotes sampling frequency which is 22050 Hz
here. The matrix obtained after performing this function
contains frames of the original speech signal filtered by
hamming filter and transformed with FFT. The elements of
the matrix are complex numbers and symmetrical because
FFT was used to transform. By calculating DFT we can
obtain the magnitude spectrum.
For AWGN:

Y=s+10^(-Eb_N0_dB/20)*n;

where ‘s’ is the transmitted sequence, Eb_N0_dB is the

SNR and ‘n’ is the Additive White Gaussian Noise.

IV. Working
Early devices, such as trumpets and horns, were passive osition',P4,...
amplification cones designed to collect sound energy and 'PlotAsTwoSidedSpectrum',false);
direct it to the ear canal. Modern devices are computerized
SpectroGraph.SpectrumType =
electroacoustic systems that transform environmental sound
into audible sound, according to audiometric and cognitive 'Spectrogram';
rules. Modern devices also use sophisticated digital signal
processing to improve speech intelligibility and user while numplays>0
comfort. This signal processing includes feedback
management, extended dynamic range compression,
directivity, lowering of frequency, and noise reduction. data=step(mic);
step(audplyer, double(data));%Play
V. MATLAB code: the Output Signal
%step(scope,data);
%Measure Screen Size of the device step(SpectroGraph,data);
%Calculate position values of figure numplays = numplays-1;
windows end
scrsz = get(0,'ScreenSize');
%P1 = [50 300 scrsz(3)/2 scrsz(4)/2]; release(mic);
%P2 = [50 80 scrsz(3)/3 scrsz(4)/3]; release(audplyer);
%P3 = [620 500 scrsz(3)/3 scrsz(4)/3]; release(scope);
P4 = [620 80 scrsz(3)/2 scrsz(4)/2]; release(SpectroGraph);

SamplesPerFrame=1024;
Fs=44100;
T=10;%Specify Algorithm run time in VI. Results
seconds here Figure 4 below is the original speech signal which is plot
numplays = (T*Fs)/SamplesPerFrame; on time versus amplitude axis.

mic=dsp.AudioRecorder();
mic.SamplesPerFrame=SamplesPerFrame;
mic.DeviceName='Microphone (Realtek
High Definition Audio)';
mic.NumChannels=1;
mic.SampleRate=Fs;
mic.OutputDataType='double';
mic.BufferSizeSource='Property';
mic.QueueDuration=2;
mic.BufferSize=128;

audplyer =
dsp.AudioPlayer('SampleRate',Fs);
Figure 4: Original speech signal
scope =
dsp.TimeScope('SampleRate',Fs,'TimeSpan Then adaptive white Gaussian noise is added to the original
',0.1,... wave signal. The purpose of this addition just to simulate
'Position',P1,'YLimits',[-1 1]); noises in the real situation. Figure 5 shows the signal after
noise addition.
scope.ShowGrid = 1;
scope.ShowLegend = 1;

SpectroGraph =
dsp.SpectrumAnalyzer('SampleRate',Fs,'P
 Figure 7: Spectrogram of Original and Adjusted Signal

Figure 5: Corrupted speech signal

Afterward, the denoising process takes place which

removes most of the noise in the signal as shown in figure
6.

Comparing the spectrograms of the original signal and the Figure 8

filtered signal, we can see that the amplitude of the noise in
the signal was noticeably reduced as shown in figure 7.

Figure 6: Signal after denoising

Figure 8 & 9: MATLAB Simulation

However, the adjusted signal strength is not increase as

your expectations. Perhaps the cause of this the error is due
to the badly involved gain function.

VII. Conclusion
New digital aids offer more capacity to tune the sound
without altering the quality and helps the listener. In this
digital hearing aid implementation of the system using
MATLAB, its the treatment is digitized. Thus, it is possible
to refine the sound signal, for example by reducing noise
and improvement of speech signals. In addition, using
digital technology, amplification can be done only at
frequencies that the user must amplify. This will eliminate
the problem with conventional amplifier that amplified the
entire signal, including noise. In digital hearing aids in
general, when the signals are converted into digital signals.
This is the digitization makes it possible to specify analyze
& filter the signals. The signals can be processed in one or
more frequency channels. At the end, the digital signal is
again converted to its analog form. The benefits of using
digital aids could
Improve the quality of life by improving the sound quality,
Greater comfort of listening, better communication in noisy
environment, better speech intelligibility group
conversations and more flexibility in case of Progressive
hearing less.

VIII. References

[1] Frost and Sullivan, World Audiology Products Markets,

1997.

[2] G. J. Proakis and G. D. Manolakis, Gigital Signal

Processing ; principles, algorithm, and application, 3rd

edition, Prentice Hall, New Jersey, 1996.

[3] http://www.hearingresearch.org/ross.html

[4] http://www.helpinguhear.com/ hearing_aids.html

[5] https://en.wikipedia.org/wiki/Hearing_aid

[6] Trudy Stetzler, Neeraj Magotra, Pedro Gelabert, Prethi

Kasthuri, Sridevi Banglore, Low Power Real Time
Programmable DSP Development Platform for Digital
Hearing Aids System, Texas Instrument.
[7] https://www.starkey.com/improve-your-
hearing/frequently-asked-questions/how-hearing-aids-
work?video=y&vidid=image4114710062001