1、ANSI S3.5-1997 Revision f ANSI S3.5-1969 (R 1986) Reaffirmed by ANSI April 22,2002 American National Standard Methods for Calculation of the Speech Intelligibility Index Secretariat Acoustical Society of America Approved 6 June 1997 American National Standards Institute, Inc. Abstract This Standard
2、defines a method for computing a physical measure that is highly correlated with the intelligibility of speech as evaluated by speech perception tests given a group of talkers and listeners. This measure is called the Speech Intelligibility Index, or SII. The SI1 is calculated from acoustical measur
3、ements of speech and noise. This standard is not a substitute for ANSI S3.2-1989 (R 1995) American National Standard Method for Measuring the IntelligibiMy of Speech over Communication Systems. Copyright Acoustical Society of America Provided by IHS under license with ASA Not for ResaleNo reproducti
4、on or networking permitted without license from IHS-,-,-AMERICAN NATIONAL STANDARDS ON ACOUSTICS The Acoustical Society of America (ASA) provides the Secretariat for Accredited Standards Committees S1 on Acoustics, S2 on Mechanical Vibration and Shock, S3 on Bioacoustics, and S12 on Noise. These com
5、mittees have wide represen- tation from the technical community (manufacturers, consumers, and general- interest representatives). The standards are published by the Acoustical Society of America through the American Institute of Physics as American National Stan- dards after approval by their respe
6、ctive standards committees and the American National Standards Institute. These standards are developed and published as a public service to provide standards useful to the public, industry, and consumers, and to Federal, State, and local governments. Each of the Accredited Standards Committees oper
7、ating in accordance with pro- cedures approved by American National Standards Institute (ANSI) is responsible for developing, voting upon, and maintaining or revising its own standards. The ASA Standards Secretariat administers committee organization and activity and provides liaison between the Acc
8、redited Standards Committees and ANSI. After the standards have been produced and adopted by the Accredited Standards Committees, and approved as American National Standards by ANSI, the ASA Standards Secretariat arranges for their publication and distribution. An American National Standard implies
9、a consensus of those substantially con- cerned with its scope and provisions. Consensus is established when, in the judgment of the ANSI Board of Standards Review, substantial agreement has been reached by directly and materially affected interests. Substantial agreement means much more than a simpl
10、e majority, but not necessarily unanimity. Consen- sus requires that all views and objections be considered and that a concerted effort be made towards their resolution. The use of American National Standards is completely voluntary. Their existence does not in any respect preclude anyone, whether h
11、e or she has approved the standards or not, from manufacturing, marketing, purchasing, or using products, processes, or procedures not conforming to the standards. NOTICE: This American National Standard may be revised or withdrawn at any time. The procedures of the American National Standards Insti
12、tute require that action be taken periodically to reaffirm, revise, or withdraw this Standard. Standards Secretariat Acoustical Society of America 120 Wall Street, 32nd Floor New York, New York 1ooO5-3993 Telephone: +i 212 248 0373 Telefax: +1 212 248 O146 E-mail: asastdsaip.org Q 1998 by the Acoust
13、ical Society of America. This Standard may not be reproduced in whole or in part in any form for sale, promotion, or any commercial purpose, or any purpose not falling within the provisions of the Copyright Act of 1976, without prior written permission of the publisher. For permission, address a req
14、uest to the Standards Secretariat of aie Acousticai Society of America. Copyright Acoustical Society of America Provided by IHS under license with ASA Not for ResaleNo reproduction or networking permitted without license from IHS-,-,-Contents Page . i Foreword 1 Scope. purpose. and applications 1 2
15、References 1 3 Definitions . 2 4 Methods for calculating Speech Intelligibility Index. SI1 9 procedures 11 the Speech Intelligibility Index 5 Methods for determining input variables for SI1 calculation 6 General relation between the intelligibility of received speech and Annexes 16 A SI1 for individ
16、uais with hearing loss . 17 8 Transferring SI1 to speech intelligibility 17 C Examples of SI1 computations . 21 Tables 1 Critical band SI1 procedurefrequencies. band importance function. standard speech spectra. internal noise. hearing threshold levels. and free-field to eardrum transfer function .
17、3 2 Equally-contributing (17 band) critical-band SI1 procedure . 4 3 One-third octave band SI1 procedure 5 4 Octave band SI1 procedure 5 B.l Critical band importance functions for various speech tests 18 8.2 One-third octave band importance functions for various speech tests 19 B.3 Octave band impor
18、tance functions for various speech tests 19 C.1 Worksheet for octave band SI1 calculation example . 21 C.2 Worksheet for one-third octave band SI1 calculation example 22 i Copyright Acoustical Society of America Provided by IHS under license with ASA Not for ResaleNo reproduction or networking permi
19、tted without license from IHS-,-,-Foreword mis Foreword is for information only and is not an integral part of ANSI S3.5-1997 Ameri- can National Standard Methods for Calculation of the Speech Intelligibility Index. This Standard defines a method for computing a physical measure that is highly corre
20、lated with the intelligibility of speech under a variety of adverse listening conditions, such as noise, filtering, and reverberation. It is a major revision of ANSI S3.5-1969 (R 1986), American National Standard Methods for the Calculation of the Aruculation Index. The most important changes in the
21、 present version of the Standard relate to the need to provide a general framework into which various methods for determining the input variables of the Speech Intelligibility Index model (e.g., the equivalent speech spectrum level, the equivalent noise spectrum level, and the equivalent hearing thr
22、eshold level) can be incorporated. For some applications these methods already exist (e.g., the modulation transfer function for determining the apparent speech-to-noise ratio in reverberation), while others still may be developed in future revisions of this Standard. In addition, the generality of
23、the Standard has been extended to include various measurement points (e.g., free-field for architectural acoustics or eardrum for telephony). The other changes of the Standard are due to new data which have been accumulated since 1969 for various parameters and procedures used in the calculations. T
24、hese new data include spread of masking, standard speech spectrum level, and relative impor- tance of various frequencies to speech intelligibility. Finally, the name has been changed from the Articulation Index to the Speech Intelligibility Index (SII). In this Standard, speech intelligibility refe
25、rs to how well an individual understands speech. It should be noted that SI1 should not be used as a substitute for determining speech intelligibility as described in ANSI S3.2-1989 (R 1995), American National standard Method for Measuting the Intelligibility of Speech over Communication Systems. Th
26、is Standard was developed under the jurisdiction of Accredited Standards Com- mittee S3, Bioacoustics, which has the following scope: StZmdats, specifications, methods of measurement and test, and terminology, in the fields of psychological and physiological amuscs, including aspeeis of general amus
27、 FAX: + 1 21 2 248 O1 46; E-mail: asastdsaip.org. iii Copyright Acoustical Society of America Provided by IHS under license with ASA Not for ResaleNo reproduction or networking permitted without license from IHS-,-,-AMERICAN NATIONAL STANDARD ANSI S3.5-1997 American National Standard Methods for Cal
28、culation of the Speech Intelligibility Index 1 Scope, purpose and applications 1.1 Scope The predictions of this Standard apply to listening conditions where the input variables of the Speech Intelligibility Index (SII) model can be accurately estimated. The input variables include the equiva- lent
29、speech spectrum level, the equivalent noise spectrum level, and the equivalent hearing thresh- old level. This includes the conditions where either speech or noise may not exist as directly measur- able physical quantities (e.g., conditions where speech correlated noise is present, such as rever- be
30、rated speech) but where equivalent speech spectrum level, equivalent noise spectrum level, and equivalent hearing threshold level can, never- theless, be calculated. The predictions made by use of this Standard are correct only on the aver- age, that is, across a group of talkers and a group of list
31、eners of both genders. The scope of the Standard is limited to natural speech, otologically normal listeners, and communication conditions which do not include multiple, sharply filtered bands of speech or sharply filtered noise. In addi- tion, the listeners should have no linguistic or cog- nitive
32、deficiencies with respect to the language used. 1.2 Purpose This Standard defines methods for computing a measure, called the Speech Intelligibility Index (Sll), that is highly correlated with the intelligibility of speech under a variety of adverse listening con- ditions, such as noise masking, fil
33、tering, and rever- beration. The SI1 is computed from acoustical measurements or estimates of speech spectrum level, from noise spectrum level, and from psy- choacoustical measurements or estimates of hear- ing threshold level. Various frequencies contribute different amounts to speech intelligibili
34、ty, and, within a certain range, a higher speech-to-noise ratio contributes to intelligibility. By measuring the speech-to-noise ratio in each contributing fre- quency band and adding the results, the intelligi- bility of a speech communication system can be predicted. 1.3 Applications SI1 procedure
35、s in this Standard consist of several parts. Clause 4 specifies calculation methods when the input variables (Le., equivalent speech spectrum level, equivalent noise spectrum level, and equivalent hearing threshold level) are known. The application domain of this framework is quite general and exten
36、ds to all listening conditions, within the scope of the Standard, where adequate methods for specifying these input variables exist. Measurement and calculation procedures for specifying the input variables with which to calcu- late SI1 for a number of conditions encountered in practice, such as ext
37、ernal noise masking, rever- berant speech, monaural listening, and some con- ditions of binaural listening are provided in clause 5. Extension of the SI1 calculation methods to indi- viduals with hearing loss is contained in annex A. Annex B provides procedures for taking into ac- count the content
38、of speech materials in the as- sessment of speech intelligibility. Examples of the basic SI1 computational proce- dures applied to octave and one-third octave fre- quency band procedures are contained in annex C. 2 References The following Standards contain provisions which, through reference in thi
39、s text, constitute provisions of this American National Standard. At the time of approval by the American National Standards In- stitute, Inc. (ANSI), the editions indicated were valid. Because Standards are revised from time to time, users should consult the latest version ap- proved by the America
40、n National Standards Insti- tute. For purposes of this Standard, the use of the latest revision of a referenced Standard is not mandatory. Information on recent editions is avail- able from the ASA Standards Secretariat. 2.1 Normative references l ANSI S3.2-1989 (R 1995) American National Standard M
41、ethod for Measuring the Intelligibility of Speech over Communication Systems. 1 O 1998 Acoustical Society of America Copyright Acoustical Society of America Provided by IHS under license with ASA Not for ResaleNo reproduction or networking permitted without license from IHS-,-,-ANSI 53.5-1997 2 ANSI
42、 S3.6-1996 American National Standard Specifications for Audiometers, 3 ANSI s3.20-1995 American National Standard Bioacoustjcal Terminology. 4 ANSI S3.21-1978 (R 1997) American National Standard Method for Pure-tone Threshold Audi- ometry. 5 ANSI S3.35-1985 (R 1997) American National Standard Metho
43、d of Measurement of Petfotmance Characteristics of Hearing Aids under Simulated in situ Working Conditions. 2.2 Informative references 6 Bentier, R. and Pavlovic, C.V. (1989) “Transfer functions and correction factors used in hearing aid evaluation and research,” ar and Hearing 10, m Dugal, R.L., Br
44、aida, L.D., and Durlach, N.I. (1 978) “Implications of previous research for the selection of frequency-gain characteristics,” in Acoustical Factors Affecting Hearing Aid Pehr- mance, edited by B.A. Studebaker unit, decibel (dB). 3.23 equivalent hearing threshold level. For monaural listening condit
45、ions, at a specified fre- quency, hearing threshold level arithmetically av- eraged across the group of ears of the listeners for whom the SI1 calculations are performed. For bin- aural listening conditions, equivalent hearing threshold level may, in an appropriate manner, ac- count for the hearing
46、threshold levels of both the right and left ears. Unit, decibel (dB). 3.24 reference internal noise spectrum level. Spectrum level of a fictitious internal noise in the ear of the listener, which, if it were an external masker, would give rise to the reference pure-tone threshold. The reference pure
47、-tone threshold is obtained in the free-field where the listener faces the source and is measured at the center of the listeners head (mid-point between the ears; lis- tener absent) using either the psychoacoustical method of constant stimuli (50% identification rate, 1 dB step size) or the psychoac
48、oustical method of limits (1 dB step size). NOTES 1 Reference internal noise spectrum level is given in tables 1-4 for different SI1 procedures of this Stan- dard. 2 In earlier documents reference internal noise spectrum level has been labeled threshold for sounds having continuous spectra (referenc
49、e 9), or O 1998 Acoustical Society of America 7 Copyright Acoustical Society of America Provided by IHS under license with ASA Not for ResaleNo reproduction or networking permitted without license from IHS-,-,-ANSI s3.5-1997 threshold spectrum densify (reference 11). The refer- ence internal noise spectrum level listed in tables 1-4 is derived from reference 11, and is equal to the ref- erence pure-
