(Design & construction of a low-cost audiometer with built-in assistive diagnostics)

Designing Audiometer with Enhanced Diagnostics

(Design & construction of a low-cost audiometer with built-in assistive diagnostics)

Ashok Kumar Dhingra

M Tech (VLSI & Embedded Systems), SY
MIT WPU, School of Electronics & Communication Engineering

Guide: Prof. Dr. Arti Khaparde, HOS

Abstract
A low-cost Audiometer was designed and constructed for carrying out
primarily, Pure Tone and Speech Audiometry. The Audiometer also
included other standard subjective behavioural tests: Tone Decay, ISIS,
Recruitments (ABLB, LDL), and Stenger’s. Other standard features offered
were: output options: Air Conduction Headphones, Bone Vibrator, and
Free Field output, Contralateral Masking, Audiogram (during Pure Tone
test) on instrument’s LCD display, battery operation.

In addition, the Audiometer offered assisted diagnostics. It analysed and

interpreted various tests conducted on a patient, based on current
knowledge, to point out which part of auditory system may be
malfunctioning (site of lesion). The diagnosis was suggestive rather than
prescriptive, with the aim of assisting the audiologist or clinician
conducting the tests. In a simplest test inference, audiometer classified
type of hearing loss: Mild, moderate, severe, profound or total.

Keywords: Audiometer, Hearing Impairment(loss), Parts of the ear,

Hearing Tests, Audiologist, auditory system, Diagnostics, Hearing Aids

1. Introduction
Hearing impairment in humans can be due to problems with any part of
our ears: the outer ear (auricle, ear canal, ear drum outer layer), the
middle ear (ear drum inner layer, the ossicles), or the inner ear
(cochlea)– the auditory system – or even beyond in the nerves leading to
brain. Hearing loss can simply be age related, or noise induced,
temporary or permanent, general or specific at certain high or even low
frequencies, abnormally non-linear.
Many of these problems can be brought out by professionals using just a
tuning fork or a range of electronic instruments like Audiometer and
Tympanometer, based on ingenious testing methods invented by
researchers over the years.
Audiometer generates essential auditory test signals which are applied to
patient’s ears or mastoid through transducers like air conduction

Page 1 of 6
(Design & construction of a low-cost audiometer with built-in assistive diagnostics)

headphones, or bone vibrator or simply broadcasted through

loudspeakers. A large range of sophisticated Audiometers are available in
the market.
This project concerns design and construction of a low-cost Audiometer
which is used for Pure Tone & Speech Audiometry and which interprets
the audiometry tests and points out (diagnoses) possible malfunction (site
of lesion/pathology) of a part of auditory system. The interpretation is
based on current knowledge in the field of audiometry, and
interpretations given by researchers.

2. Design & Construction:

Block diagram: Fig 1 below shows simplified block diagram of the

Audiometer

Fig. 1: Block Diagram of Audiometer

Page 2 of 6
(Design & construction of a low-cost audiometer with built-in assistive diagnostics)

Test Signals Generated by the Audiometer:

a. Pure Sinewave b. Warble Tone

C. Masking Noise d. Pulsed waveform

Fig. 2: Test Signals generated by audiometer

All blocks in the block diagram i.e. Sinewave generator, Masking Noise
generator, Attenuator, Output Amplifiers, Electronic Switching, Op Amps,
Rectifier and Voltage Regulator were implemented by commercially available
integrated circuits.

All transducers: Microphones, AC Headphones, and Bone Conductors, and LCD

displays were of standard makes available in the market.

Two PCBs were designed, one for accommodating touch keys plus display, and
the other a mixed signal PCB for all analog and digital components.

a b
Fig 3: a. Audiogram Display, b. Touch keys

Page 3 of 6
(Design & construction of a low-cost audiometer with built-in assistive diagnostics)

Calibration: A software based system to bring consistency of displayed output

and actual output was designed and implemented.

3. Assistive Diagnostics and Classification [3]:

This is based on following table of audiometric (subjective, behavioural and
Speech) tests and inferences, compiled from references given at the end.
Inferences based on only one test are not fool-proof. Actual diagnosis is based
on correlating multiple tests.
A more exhaustive, authoritative assistive diagnostic system can be built from
current knowledge and use of AI:

Test Possible Inference if adverse score or observation

Pure Tone Average ( of 1 to 3 KHz) Threshold
 91 dBHL : Profound Hearing Loss
71 to 90 dBHL: Severe Hearing Loss
41 to 70 dBHL: Moderate Hearing Loss
26 to 40 dBHL: Mild Hearing Loss
< 26 dBHL: Normal hearing

A notch centered at about 4000 Hz: possible noise

induced hearing loss
Tone Decay High Levels of TD indicate problems beyond cochlea (Retro-
Cochlear pathology)
0.5 dB in 60 s : Normal
10 to 15 dB in 60 s: Mild
20 to 25 dB in 60 s: Moderate
SISI Score > 70% +ve: inner ear problem (Cochlear pathology)
Score < 30% - ve, but if any problem in PT, disorder
elsewhere than in inner ear
Recruitment: Cochlear problem
ABLB
Bekesy Type 1: Normal hearing or conductive loss
Type 2: Cochlear loss
Type 3: 8th nerve lesion
Speech: SRT If it does not correlate with average PT threshold, possible
post-cochlear pathology
Speech: SDS Sensorineural loss or post-cochlear lesions (SDS < 30%)

Page 4 of 6
(Design & construction of a low-cost audiometer with built-in assistive diagnostics)

4. Results:
The Audiometer correctly generated essential test signals and applied on patient
under test. Pure Tone, Speech Tests, and other special tests like Tone Decay,
SISI, ABLB could be done successfully and correct results obtained.
Built in diagnostics also worked as designed, though at this stage it was too
simplistic.

5. Conclusions and Recommendations:

A low-cost audiometer was designed and constructed. It also featured built in
assistive diagnostics to help audiologist or clinician to localise problem in
auditory system.
To make the audiometer more useful,
1. More specialized hearing tests need to be included
2. More exhaustive use to be made of current knowledge in the field
3. More use of AI tools to be made to correlate various tests and arrive at
more accurate diagnosis.

References:
1. Jenifer J Lents, Audiometric Testing with Pulsed, Steady, and Warble
Tones in Listeners with Tinnitus and Hearing Loss. American Journal of
Audiometry, 2017 Sept., 26(3) 328 - 337
2. J. Majumdar and LK Sharma. Application of Data Mining Techniques to
Audiometric data among professionals in India. J. of Scientific Research &
Report 3(23):2960-2971, 2014

3. Integrated Audiometric Tests for Cochlear and Retro-cochlear Pathologies.

Jyotsna Nadkarni
Audiologist and Speech Pathologist, Bombay Hospital Medical Research
Centre, Mumbai 400 020.
4. Introduction to Audiometry, J Martin
5. www.drbalu.co.in, of Dr T Balasubramanium

6. Marco Cox. A Bayesian binary classification approach to pure tone

audiometry. arXiv:1511.08670v2[stat.AP]16 Mar 2016

Page 5 of 6
(Design & construction of a low-cost audiometer with built-in assistive diagnostics)

7. Anwar MN. Clustering audiology data. 19th Machine Learning conference of

Belgium, 2010
8. Xinyu Song. Fast, Continuous Audiogram Estimation using machine
learning. Ear & Hearing 36(6) June 2015
9. Xinyu Song. Improving pure tone audiometry using probabilistic machine
learning. Washington University
10.Jacob Grdner. Bayesian active model selection with an application in
automated audiometry.
11.I. Kononenko. Machine learning for medical diagnosis. Artificial
Intelligence in Medicine, 23(I):89-109, 2001
12.Barbour D. (2015). Optimizing Pure Tone Audiometry using Gaussian
Processes. Association for Research in Otolaryngology
13.D Bing. Predicting the hearing outcome in sudden sensorineural hearing
loss via machine learning. Jan 2018, Clinical Otolaryngology.
14.Mark dRuiter, Basic Audiometry Learning Manual (Machine Learning in
Audiometry). Book.

Page 6 of 6