EN ISO 8253-3:2012

SLOVENSKI STANDARD
01-september-2012
1DGRPHãþD
SIST EN ISO 8253-3:1999
Akustika - Avdiometrijske preskusne metode - 3. del: Govorna avdiometrija (ISO
8253-3:2012)
Acoustics - Audiometric test methods - Part 3: Speech audiometry (ISO 8253-3:2012)
Akustik - Audiometrische Prüfverfahren - Teil 3: Sprachaudiometrie (ISO 8253-3:2012)
Acoustique - Méthodes d'essais audiométriques - Partie 3: Audiométrie vocale (ISO
8253-3:2012)
Ta slovenski standard je istoveten z: EN ISO 8253-3:2012
ICS:
13.140 Vpliv hrupa na ljudi Noise with respect to human
beings
17.140.01 $NXVWLþQDPHUMHQMDLQ Acoustic measurements and
EODåHQMHKUXSDQDVSORãQR noise abatement in general
2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.

EUROPEAN STANDARD
EN ISO 8253-3
NORME EUROPÉENNE
EUROPÄISCHE NORM
March 2012
ICS 13.140 Supersedes EN ISO 8253-3:1998
English Version
Acoustics - Audiometric test methods - Part 3: Speech
audiometry (ISO 8253-3:2012)
Acoustique - Méthodes d'essais audiométriques - Partie 3: Akustik - Audiometrische Prüfverfahren - Teil 3:
Audiométrie vocale (ISO 8253-3:2012) Sprachaudiometrie (ISO 8253-3:2012)
This European Standard was approved by CEN on 29 February 2012.

CEN members are bound to comply with the CEN/CENELEC Internal Regulations which stipulate the conditions for giving this European
Standard the status of a national standard without any alteration. Up-to-date lists and bibliographical references concerning such national
standards may be obtained on application to the CEN-CENELEC Management Centre or to any CEN member.

This European Standard exists in three official versions (English, French, German). A version in any other language made by translation
under the responsibility of a CEN member into its own language and notified to the CEN-CENELEC Management Centre has the same
status as the official versions.

CEN members are the national standards bodies of Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia,
Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland,
Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland, Turkey and United Kingdom.

EUROPEAN COMMITTEE FOR STANDARDIZATION
COMITÉ EUROPÉEN DE NORMALISATION

EUROPÄISCHES KOMITEE FÜR NORMUNG

Management Centre: Avenue Marnix 17, B-1000 Brussels
© 2012 CEN All rights of exploitation in any form and by any means reserved Ref. No. EN ISO 8253-3:2012: E
worldwide for CEN national Members.

Contents Page
Foreword .3

Foreword
This document (EN ISO 8253-3:2012) has been prepared by Technical Committee ISO/TC 43 "Acoustics" in
collaboration with Technical Committee CEN/TC 211 “Acoustics” the secretariat of which is held by DS.
This European Standard shall be given the status of a national standard, either by publication of an identical
text or by endorsement, at the latest by September 2012, and conflicting national standards shall be
withdrawn at the latest by September 2012.
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent
rights. CEN [and/or CENELEC] shall not be held responsible for identifying any or all such patent rights.
This document supersedes EN ISO 8253-3:1998.
According to the CEN/CENELEC Internal Regulations, the national standards organizations of the following
countries are bound to implement this European Standard: Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech
Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia,
Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain,
Sweden, Switzerland, Turkey and the United Kingdom.
Endorsement notice
The text of ISO 8253-3:2012 has been approved by CEN as a EN ISO 8253-3:2012 without any modification.

INTERNATIONAL ISO
STANDARD 8253-3
Second edition
2012-03-01
Acoustics — Audiometric test methods —
Part 3:
Speech audiometry
Acoustique — Méthodes d’essais audiométriques —
Partie 3: Audiométrie vocale
Reference number
ISO 8253-3:2012(E)
©
ISO 2012
ISO 8253-3:2012(E)
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form or by any means,
electronic or mechanical, including photocopying and microfilm, without permission in writing from either ISO at the address below or ISO’s
member body in the country of the requester.
ISO copyright office
Case postale 56 • CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail copyright@iso.org
Web www.iso.org
Published in Switzerland
ii © ISO 2012 – All rights reserved

ISO 8253-3:2012(E)
Contents Page
Foreword . v
Introduction .vi
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 Requirements for recording of speech material . 5
4.1 General requirements . 5
4.2 Reference recording . 5
4.3 Recording environment . 6
4.4 Frequency response of recording equipment . 6
4.5 Interval between successive test items . 6
4.6 Levels . 6
4.7 Signal-to-noise ratio . 6
4.8 Phonemic balance across test lists . 6
4.9 Perceptual balance across test lists . 7
5 Validation of speech material recordings . 7
5.1 General . 7
5.2 Determination of reference speech recognition curve . 7
5.3 Determination of perceptual equivalence of test lists . 8
5.4 Determination of repeatability of results (average test-retest reliability) . 8
5.5 Documentation . 9
6 Requirements for speech audiometry . 9
7 Ambient sound pressure levels in test room for speech audiometry . 9
8 Sound field speech audiometry . 9
9 Preparation and instruction of test subject .10
9.1 General .10
9.2 Preparation of test subject .10
9.3 Instruction of test subject .10
10 Subject’s response mode .10
11 Determination of speech detection threshold level . 11
12 Determination of speech recognition threshold level . 11
12.1 General . 11
12.2 Descending procedure using 5 dB steps .12
12.3 Alternative descending procedure .12
12.4 Adaptive procedure using fixed step sizes .13
12.5 Other adaptive procedure .14
13 Determination of speech recognition scores .14
14 Contralateral masking .14
15 Speech audiometry with competing sound .15
15.1 Type of competing sound .15
15.2 Presentation of competing sound .15
15.3 Speech and competing sound levels .15
15.4 Test procedure .15
16 Format of speech audiogram .17
17 Measurement uncertainty .17
18 Maintenance and calibration of equipment .18
18.1 General .18
ISO 8253-3:2012(E)
18.2 Intervals between tests .18
18.3 Stage A: Routine checking and listening tests .18
18.4 Stage B: Periodic electroacoustic tests .19
18.5 Stage C: Basic calibration tests .19
Annex A (informative) Example of speech materials .20
Annex B (informative) Examples of competing sound conditions .21
Annex C (informative) Typical results .22
Annex D (informative) Optimization of perceptual balance of test lists .24
Annex E (informative) Measurement uncertainty .25
Bibliography .30
iv © ISO 2012 – All rights reserved

ISO 8253-3:2012(E)
Foreword
ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies
(ISO member bodies). The work of preparing International Standards is normally carried out through ISO
technical committees. Each member body interested in a subject for which a technical committee has been
established has the right to be represented on that committee. International organizations, governmental and
non-governmental, in liaison with ISO, also take part in the work. ISO collaborates closely with the International
Electrotechnical Commission (IEC) on all matters of electrotechnical standardization.
International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2.
The main task of technical committees is to prepare International Standards. Draft International Standards
adopted by the technical committees are circulated to the member bodies for voting. Publication as an
International Standard requires approval by at least 75 % of the member bodies casting a vote.
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent
rights. ISO shall not be held responsible for identifying any or all such patent rights.
ISO 8253-3 was prepared by Technical Committee ISO/TC 43, Acoustics.
This second edition cancels and replaces the first edition (ISO 8253-3:1996), which has been technically revised.
ISO 8253 consists of the following parts, under the general title Acoustics — Audiometric test methods:
— Part 1: Pure-tone air and bone conduction threshold audiometry
— Part 2: Sound field audiometry with pure-tone and narrow-band test signals
— Part 3: Speech audiometry
ISO 8253-3:2012(E)
Introduction
Speech audiometry is used for the assessment of hearing in connection with diagnostic evaluation and
audiological rehabilitation.
The results of speech audiometry depend on the speech material and test method used. This part of ISO 8253
sets conditions for speech materials in order to assure minimum requirements of precision and comparability
between different tests using different speech materials including materials in different languages. It also
specifies procedures to be used when testing speech recognition.
vi © ISO 2012 – All rights reserved

INTERNATIONAL STANDARD ISO 8253-3:2012(E)
Acoustics — Audiometric test methods —
Part 3:
Speech audiometry
1 Scope
This part of ISO 8253 specifies basic methods for speech recognition tests for audiological applications.
In order to ensure minimum requirements of precision and comparability between different test procedures
including speech recognition tests in different languages, this part of ISO 8253 specifies requirements for the
composition, validation and evaluation of speech test materials, and the realization of speech recognition tests.
This part of ISO 8253 does not specify the contents of the speech material because of the variety of languages.
Furthermore, this part of ISO 8253 also specifies the determination of reference values and fulfilment
requirements for the realization and manner of presentation.
This part of ISO 8253 specifies procedures and requirements for speech audiometry with the recorded
test material being presented by air conduction through an earphone, or from a loudspeaker for sound field
audiometry. Methods for using noise either for masking the non-test ear or as a competing sound are described.
Some test subjects, for example children, can require amended test procedures not specified in this part of ISO 8253.
Specialized tests such as those used for evaluating directional hearing and dichotic hearing are outside the
scope of this part of ISO 8253.
2 Normative references
The following referenced documents are indispensable for the application of this document. For dated
references, only the edition cited applies. For undated references, the latest edition of the referenced document
(including any amendments) applies.
ISO 266, Acoustics — Preferred frequencies
ISO 8253-1, Acoustics — Audiometric test methods — Part 1: Pure-tone air and bone conduction audiometry
ISO 8253-2, Acoustics — Audiometric test methods — Part 2: Sound field audiometry with pure-tone and
narrow-band test signals
ISO/IEC Guide 98-3, Uncertainty of measurement — Part 3: Guide to the expression of uncertainty in
measurement (GUM:1995)
IEC 60645-1, Electroacoustics — Audiological equipment — Part 1: Pure-tone audiometers
IEC 60645-2:1993, Audiometers — Part 2: Equipment for speech audiometry
IEC 61672-1, Electroacoustics — Sound level meters — Part 1: Specifications
3 Terms and definitions
For the purposes of this document, the terms and definitions given in ISO 8253-1 and ISO 8253-2 and the
following apply.
ISO 8253-3:2012(E)
3.1
speech signal
acoustic signal which carries information in a given language
NOTE A speech signal may be a voice signal or an acoustic signal simulating a voice signal.
3.2
test item
particular monosyllabic or polysyllabic word or logatom, or sentence, or time-limited segment of connected
speech, used in accordance with defined rules of presentation and scoring in a speech audiometric procedure
NOTE Scoring may be based on a complete test item or parts thereof being correctly recognized.
3.3
speech material
entire set of test items which is used for speech recognition tests
NOTE Usually the speech material is subdivided into several test lists.
3.4
open-set test material
set of test items in which the number of alternative responses to each test item is unlimited
3.5
closed-set test material
set of test items in which the number of alternative responses to each test item is limited
3.6
phoneme
smallest recognizable unit of speech in a given natural language
3.7
phoneme class
subdivision of phonemes that show characteristic similarities in vocal production mode as well as in acoustical
signal properties
3.8
syllable
segment of speech which consists of a vowel with or without one or more accompanying consonants immediately
preceding or following
3.9
spondee
two-syllable word with equal stress on each syllable
3.10
logatom
syllabic unit that has no verbal meaning to the listener
NOTE A logatom is sometimes called a “nonsense syllable”.
3.11
test list
number of selected test items, presented and scored as a single unit
3.12
set of test items
selected number of test items from a test list
2 © ISO 2012 – All rights reserved

ISO 8253-3:2012(E)
3.13
phonemically balanced test list
test list which contains approximately the same proportions of the various phonemes as are present in typical
speech communication within a given language
3.14
perceptually balanced test list
test list which yield equivalent speech recognition score with minimized variance at a specified speech level
under equivalent conditions (specified manner of presentation and specified group of listeners)
NOTE Usually the specified group of listeners consists of otologically normal persons with an age from 18 years to
25 years inclusive. For special tests, different groups (e.g. otologically normal children within a specified age range) may
be used as well.
3.15
carrier phrase
sentence or phrase in which a test item is embedded such that the correct recognition of the test item is not
dependent upon the context or meaning of the sentence or phrase
3.16
reference recording of speech material
master recording
well-defined recording that represents the speech material and that is used for validation and application of the
speech material
3.17
speech level
equivalent continuous sound pressure level of the speech material as measured in an appropriate coupler, ear
simulator or in a sound field, with frequency weighting C according to IEC 61672-1
NOTE For speech test lists based on single test items separated by silent intervals, the integration does not include
these intervals. For test lists based on single test items with a carrier phrase, the integration includes the test items only. In
a digital recording, the silent intervals can be removed by editing. Numerical correction may also be made by determination
of the total duration of the test items and the total duration of the silent intervals.
3.18
equivalent continuous sound pressure level
L
p,eqT
ten times the logarithm to the base 10 of the ratio of the time average of the square of the sound pressure, p,
during a stated time interval of duration, T (starting at t and ending at t ), to the square of a reference value,
1 2
p , expressed in decibels
t
 2 
 
pt()dt
∫
T 
t
 1 
L =10 lg dB
pT,eq
 
p
 
 
 
 
where the reference value, p , is 20 µPa
NOTE Because of practical limitations of the measuring instruments, p is always understood to denote the square
of a frequency-weighted and frequency-band-limited sound pressure. If a specific frequency weighting as specified in
IEC 61672-1 and/or specific frequency bands are applied, this should be indicated by appropriate subscripts; e.g. L
p,A,10 s
denotes the A-weighted time-averaged sound pressure level over 10 s.
[3]
[ISO/TR 25417 , 2.3]
3.19
speech detection threshold level
for a given test subject, for specified speech material and a specified manner of signal presentation, the speech
level at which it is detected (not necessarily understood) in 50 % of the trials
ISO 8253-3:2012(E)
3.20
speech recognition score
for a given test subject, for specified speech material, a specified manner of signal presentation and at a
specified speech level, the percentage of correctly recognized test items or scorable items if the scoring
method is not based on whole test items
3.21
maximum speech recognition score
for a given test subject, for specified speech material and a specified manner of signal presentation, the
maximum speech recognition score obtained regardless of speech level
3.22
speech recognition threshold level
for a given test subject, for specified speech material and a specified manner of signal presentation, the lowest
speech level or speech-to-noise ratio at which the speech recognition score is equal to 50 %
NOTE Speech recognition threshold has been called “speech reception threshold”.
3.23
reference speech recognition threshold level
for specified speech material and a specified manner of signal presentation, the median value of the speech
recognition threshold levels of a sufficiently large number of otologically normal persons, of both sexes, between
18 years and 25 years inclusive for whom the test material is appropriate
3.24
optimum speech level
for a given test subject, for specified speech material and a specified manner of signal presentation, the speech
level or range of speech levels at which maximum speech recognition score is obtained
3.25
half-optimum speech level
for a given test subject, for specified speech material and a specified manner of signal presentation, the speech
level at which half of the maximum speech recognition score is obtained
3.26
speech recognition curve
for a given test subject, for specified speech material and a specified manner of signal presentation, a curve
that describes the speech recognition score as a function of speech level
3.27
reference speech recognition curve
for specified speech material and a specified manner of presentation, a curve that describes the median speech
recognition score as a function of speech level for a sufficiently large number of otologically normal persons of
both sexes, aged between 18 years and 25 years inclusive and for whom the test material is appropriate
3.28
hearing level for speech
for specified speech material and a specified manner of signal presentation, the speech level minus the
appropriate reference speech recognition threshold level
3.29
competing sound
additional sound that is presented during speech recognition tests
3.30
competing sound pressure level
equivalent continuous sound pressure level of a competing sound as measured in an appropriate coupler, ear
simulator or in a sound field, with frequency weighting C according to IEC 61672-1
4 © ISO 2012 – All rights reserved

ISO 8253-3:2012(E)
3.31
speech-to-noise level difference
speech-to-noise ratio
SNR
difference between the speech level and the competing sound level
3.32
effective masking level for speech
level of a specified masking sound which is numerically equal to that hearing level for speech to which the
speech recognition threshold level for a specified speech signal would be raised by the presence of that
masking sound
3.33
most comfortable level for speech
for a given test subject and a specified manner of presentation, the speech level at which the loudness of the
speech signal is judged by the test subject to be the most comfortable
3.34
connected speech
running speech
continuous speech with natural intonation, consisting of consecutive sentences with logical connections
4 Requirements for recording of speech material
4.1 General requirements
Each copy of the recording shall contain the following signals:
a) the speech material;
b) a signal for the calibration of the equipment for speech audiometry, of a duration not less than 60 s; the
calibration signal shall be a weighted random noise, e.g. as specified in IEC 60645-2:1993, 13.1, a band
of noise centred at 1 kHz and having a bandwidth of one-third octave, or a frequency-modulated tone at
1 kHz having a bandwidth of at least one-third octave — the modulating signal shall be either sinusoidal or
triangular with a repetition rate in the range from 4 Hz to 20 Hz;
c) signals for testing the frequency response of the equipment to be used for speech audiometry including
the playback equipment and the recording — such signals shall have a duration of not less than 15 s and
consist of one-third-octave bands of white noise centred at the preferred one-third-octave frequencies in
accordance with ISO 266 in the frequency range 125 Hz to 8 000 Hz.
NOTE For speech material based on other time and/or frequency weightings than C-weighted equivalent sound
pressure level, correction factors can be determined describing the difference related to C-weighted equivalent sound
pressure level for speech level, calibration signal level and competing noise level.
4.2 Reference recording
A suitable speaker, either male or female, should have a normal and clear articulation without pronounced
dialect. The speaker should be instructed to maintain clarity, natural pace, and vocal effort and avoid emphasis
on key words.
A reference recording mainly consists of the elements specified in 4.1. The reference recording shall not be
modified. All copies shall be made from the reference recording. Any changes of the reference recording (e.g.
a new recording or changes of levels or frequency content) require a new validation (see Clause 5).
For speech audiometry intended to be performed in noise the reference recording shall also include the
associated competing noise or noises, either on the same or on separate channels.
ISO 8253-3:2012(E)
4.3 Recording environment
The room in which the speech is recorded shall be sufficiently quiet to provide a signal-to-noise ratio of at least
40 dB and shall have a reverberation time of less than 0,5 s at any frequency in the range from 125 Hz to 8 000 Hz.
4.4 Frequency response of recording equipment
The frequency response of the recording equipment including the microphone used shall be flat within 2 dB in
the frequency range from 125 Hz to 8 000 Hz.
NOTE The quality of the recording and negligible loss of quality after repeated use make digital recordings superior.
4.5 Interval between successive test items
The intervals between the successive test items of the recording should be specified and constant within ±10 %.
For single test words, an interval of not less than 4 s is suggested.
4.6 Levels
4.6.1 Speech material
The average reference speech recognition threshold level of the smallest set of test items to be used at a
predetermined sound pressure level shall not deviate by more than ±1 dB from the average of all test items
within the list. The reference speech recognition threshold level of any single test item shall not deviate by
more than ±3 dB from the average reference speech recognition threshold level of all test items of the list.
The average speech recognition threshold level of test lists shall not differ from the average reference speech
recognition threshold level of all test items recorded by more than ±1 dB, see Annex D.
4.6.2 Calibration signal
The level of the calibration signal shall not deviate by more than ±0,5 dB from the average speech level of all
test items of the specific speech material. The level of the calibration signal shall be measured using the same
frequency weighting as when measuring the speech level.
4.7 Signal-to-noise ratio
The level of any inherent background noise on the recording shall be at least 40 dB below the level of the
calibration signal in accordance with 4.6.2 as measured using the same frequency weighting as when measuring
the speech level.
4.8 Phonemic balance across test lists
All test lists of a specific speech material shall be phonemically balanced.
In some cases, it may not be possible to achieve perfect phonemic balance. In such cases, the test lists shall
be phonemically balanced based on phoneme classes, i.e. voiced and unvoiced plosives and fricatives, nasals,
long and short vowels.
The distribution of phonemes of the speech material should approximate the phoneme distribution of the
test language as closely as possible. Speech materials consisting of a large amount of items allow a closer
approximation than smaller speech materials. In some cases, the possibility of achieving a close approximation
to the phoneme distribution is limited, e.g. in screening tests or tests for children, because the speech material
has to be limited to a relatively small number of test items.
6 © ISO 2012 – All rights reserved

ISO 8253-3:2012(E)
4.9 Perceptual balance across test lists
All test lists of a specific speech material shall be perceptually balanced, i.e. the result of the speech recognition
test shall be independent of the choice of test list.
Perceptually balanced test lists shall fulfil the requirements given in 4.6.1. Additional requirements for
perceptually balanced test lists are given in the following.
The speech recognition curves for all test lists shall have been determined in such a way that the speech
recognition threshold level can be determined with an expanded uncertainty not exceeding 1 dB (the width
of the 95 % confidence interval shall not exceed 2 dB); see Annex E. To achieve this, speech recognition
measurements at different presentation levels (or speech-to-noise ratios in case of a speech recognition test
in noise) shall be performed with at least 10 otologically normal persons whose native language is the same
as the language of the speech material. The speech recognition measurements shall be performed within
the normal context of the speech stimulus (e.g. presentation of sentences when using a sentence test). The
speech recognition of each particular test item shall then be scored and analysed independently (e.g. scoring
of particular words of the sentences).
In order to achieve an expanded uncertainty not exceeding 1 dB, extreme care is necessary to reduce the
potential uncertainty sources.
In order to increase the perceptual homogeneity of the speech material, speech recognition differences
between the particular speech items may be minimized (see Annex D).
5 Validation of speech material recordings
5.1 General
For the validation, equipment fulfilling the requirements of IEC 60645-2:1993, Clauses 6 to 17, type A, A-E,
B or B-E shall be used. Testing shall be performed monaurally. If equipment of type A-E or B-E is used, the
evaluation shall include using both uncorrected earphone and earphone corrected for free-field equivalent
output. Only complete test lists shall be used.
NOTE It can be expected that the use of equipment of Type A-E or B-E gives identical results for both binaural
earphone and loudspeaker listening due to their free-field equivalent frequency response.
5.2 Determination of reference speech recognition curve
The reference speech recognition curve for each speech material and manner of presentation shall be
determined by performing speech recognition tests using a sufficiently large group (at least 25) of otologically
normal persons whose first language is the language of the speech material. The minimum requirements for the
pure-tone hearing threshold levels of test subjects shall be specified by the laboratory carrying out the validation.
It is recommended that the following criterion based on the pure-tone audiogram be used: the hearing threshold
levels of the test subjects should be at or below 10 dB at the frequencies 250 Hz, 500 Hz, 750 Hz, 1 000 Hz,
1 500 Hz, 2 000 Hz, 3 000 Hz, 4 000 Hz, 6 000 Hz, and 8 000 Hz. At a maximum of two frequencies the hearing
threshold level may be 15 dB.
If the speech recognition test is intended to be performed with children, the reference speech recognition curve
shall be determined using a sufficiently large group (at least 25) of otologically normal children of the same age
range as the speech material is intended to be used for.
If the speech recognition test is intended to be used in silence, the reference speech recognition curve shall
be determined in silence.
If the speech recognition test is intended to be used with a competing noise, the reference speech recognition
curve shall be determined using this noise. If the test is intended to be used with different noises, the reference
speech recognition curve shall determined using each particular noise.
ISO 8253-3:2012(E)
The reference speech recognition curve shall be given as a table which describes the relation between speech
level (or speech-to-noise level difference) and speech recognition score. The speech levels (or speech-to-
noise level differences) which are required to achieve the speech recognition scores 30 %, 40 %, 50 %, 60 %,
70 %, 80 %, and 90 % shall be specified. These values can be calculated using an adequate interpolation
between the measured data.
If persons show a large variability of speech recognition results (e.g. children), it can prove necessary to
measure the reference speech recognition curve using a larger number of persons.
If it is difficult to determine the whole reference speech recognition curve from 0 % to 100 %, ensure that the
speech recognition curve is determined in the range of interest. If the speech recognition test is intended to
be used for the determination of speech recognition threshold levels, the reference speech recognition curve
should be measured at values near the speech recognition threshold level (including values above and below
this value) using a sufficient number of measurements. If the speech recognition test is intended to be used for
the determination of the optimum speech level, the reference speech recognition curve should be measured at
values from 50 % (or less) to values near to 100 % using a sufficient number of measurements.
5.3 Determination of perceptual equivalence of test lists
The perceptual equivalence of test lists of speech materials fulfilling the requirements specified in 4.9 shall be
determined by performing speech recognition tests using a sufficiently large group (at least 10) of otologically
normal persons as described in 5.2.
If the speech recognition test is intended to be used in silence, the perceptual equivalence of test lists shall be
determined in silence.
If the speech recognition test is intended to be performed with a competing noise, the perceptual equivalence
of test lists shall be determined using this noise. If the test is intended to be performed with different noises, the
perceptual equivalence of test lists shall be determined using each particular noise.
If persons show a large variability of speech recognition results (e.g. children), it may be necessary to determine
the perceptual equivalence of test lists curve using a larger number of persons.
The perceptual equivalence of test lists shall be specified by the average 95 % confidence intervals of the test
lists. If the speech recognition test is intended for the determination of speech recognition threshold levels, the
equality of test lists shall be specified as the average 95 % confidence interval for the test lists of the speech
recognition threshold level when using different test lists with the same group of test subjects. If the speech
recognition test is intended for the determination of the optimum speech level, the equality of test lists shall be
specified as the average 95 % confidence intervals for the test lists of the speech recognition scores 50 %,
60 %, 70 %, 80 %, and 90 % when using different test lists with the same group of test subjects.
That means that the squared differences between the results using the different test lists should be averaged
across listeners. The square root of this value multiplied by 2 gives the desired value.
5.4 Determination of repeatability of results (average test-retest reliability)
The repeatability of a speech recognition test result is quantified by the average test-retest reliability within one
subject. Generally, different test lists shall be used in test and retest because the responses of the initial test
may be memorized by the subject which may influence the result of the retest.
If the speech recognition test is intended to be performed in silence, the average test-retest reliability shall be
determined in silence.
If the speech recognition test is intended to be performed with a competing noise, the average test-retest
reliability shall be determined using this noise. If the speech recognition test is intended to be performed with
different noises, the average test-retest reliability shall be determined using each particular noise.
The average reliability (test versus retest) shall be specified by the average 95 % confidence intervals across subjects.
That means that the squared differences between the results of the test and the retest measurements should
be averaged across listeners. The square root of this value multiplied by 2 gives the desired value.
8 © ISO 2012 – All rights reserved

ISO 8253-3:2012(E)
If the speech recognition test is intended for the determination of speech recognition threshold levels, the test-
retest reliability shall be assessed for the speech recognition threshold level. If the speech recognition test is
intended for the determination of the optimum speech level, the test-retest reliability shall be specified for the
speech recognition scores 50 %, 60 %, 70 %, 80 %, and 90 %.
5.5 Documentation
Each copy of the recording shall be supplied with documentation containing the following information:
a) a printed version of the recorded speech material;
b) the method(s) of scoring, e.g. by sentence, word, key word, or phoneme;
c) the reference speech recognition curves as specified in 5.2 for each method of scoring, the origin of
the reference curves and details of measuring conditions (e.g. number, age range and range of hearing
threshold levels of test subjects and type of equipment for speech audiometry used as well as the
arrangement of loudspeakers for competing noise, if applicable);
d) the repeatability of results as specified in 5.4;
e) the durations, frequency spectra and levels relative to speech level of the calibration signal and other
recorded signals for the purpose of testing performance specifications in accordance with 4.1 b) and c);
f) the level of the calibration signal and its relationship to the reference speech recognition threshold level;
g) the intervals between successive test items;
h) the characteristics of any accompanying competing noise signal.
6 Requirements for speech audiometry
Speech audiometry shall be performed using equipment fulfilling the requirements of IEC 60645-2. Testing
may be performed either by earphones monaurally or binaurally, by bone vibrator, or in a sound field. The
level of a speech signal shall be expressed in terms of either speech level or hearing level for speech for the
particular test material being used.
7 Ambient sound pressure levels in test room for speech audiometry
The ambient sound pressure levels in the test room shall not mask the speech signals. The requirements for
the ambient sound pressure levels in the test room depend on the mode of presentation of the speech signal,
i.e. through an earphone, bone vibrator or loudspeaker.
NOTE Requirements for ambient noise during speech audiometry can be less stringent than those for pure-tone
threshold audiometry. If a test room is appropriate for pure-tone threshold audiometry down to a certain hearing level value
over the whole frequency range, the room can also serve for speech audiometry
...