At the August 1986 Executive Board meeting, the tutorial paper entitled “Calibration of Speech Signals Delivered via Earphones” was accepted for dissemination. The document was prepared by the Committee on Audiologic Evaluation and Working Group on Electroacoustic Characteristics. The Working Group members were Gerald R. Popelka (chair), Michael E. Dybka and Terry Wiley. The Committee on Audiologic Evaluation members were Martin S. Robinette (chair), Margaret F. Carlin, John P. Durrant, Thomas A. Frank, Sandra M. Gordon-Salant, Carol Kamara (ex officio), and Michael P. Gorga. Monitoring vice president was Gilbert R. Herer, vice president for clinical affairs.
This material is reproduced with permission from American National Standard Specifications for Audiometers, ANSI S3.6-1969, copyright 1969 by the American National Standards Institute. Copies of this standard may be purchased from the American National Standards Institute, 1430 Broadway, New York, NY 10018.
WHEREAS, the Committee on Audiologic Evaluation is charged in part to review research evidence and current clinical practice and to evaluate the extent of consensus existing with respect to procedural techniques and electroacoustic characteristics of audiometric instrumentation pertinent to the clinical evaluation of auditory function; and to develop, when warranted, informational papers applicable to the practice of clinical audiometry for appropriate dissemination, and
WHEREAS, the committee is further empowered by LC 28-78 to appoint working groups (on approval by the Executive Board) to (1) review research and evaluate the degree of consensus and (2) compile a summary report dealing with a designated task, and
WHEREAS, the Committee on Audiologic Evaluation's Working Group on Electroacoustic Characteristics has fulfilled the above outlined responsibilities by submitting a needed tutorial paper entitled “Calibration of Speech Signals Delivered via Earphones,” and
WHEREAS, this tutorial paper is designed to assist audiologists in performing and/or supervising those measurements required to determine the calibration of an audiometer, and
WHEREAS, the Committee on Audiologic Evaluation reviewed and reached consensus concerning this tutorial paper, as well as submitted it to the select peer review process; therefore
RESOLVED, That the tutorial paper entitled “Calibration of Speech Signals Delivered via Earphones,” prepared by the Working Group on Electroacoustic Characteristics and submitted by the Committee on Audiologic Evaluation, be accepted for dissemination; and further
RESOLVED, That this tutorial paper be made available free to members and, if possible, published in an ASHA journal so it becomes part of the referenced literature.
Audiologists routinely use speech signals to draw inferences regarding speech processing ability of patients. The reliability and validity of the obtained estimates depend, in large part, on the performance of the audiometric equipment used for those measurements. The extent to which the performance of a given audiometer conforms to the American National Standards Institute, Inc. (ANSI) specifications for audiometers (S3.6-1969; R-1973), therefore, is an issue of major concern to the audiological community.
Among the responsibilities with which the Committee on Audiologic Evaluation has been charged are the following: (1) to review research evidence and current clinical practice and to evaluate the extent of consensus existing with respect to procedural instrumentation pertinent to the clinical evaluation of auditory function; and (2) to develop, when warranted by research evidence and/or by consensus, suggested procedures (guidelines) and informational papers applicable to the practice of clinical audiometry for appropriate dissemination on approval by the Legislative Council. In response to the charge of the Committee on Audiologic Evaluation, the committee's Working Group on Electroacoustic Characteristics has prepared this tutorial paper.
This tutorial paper is the second in a series designed to assist audiologists in performing and/or supervising those measurements required to determine the calibration of an audiometer. Specifically, this tutorial addresses the calibration of the audiometers for speech signals delivered via earphones. The order in which the measurements are presented is intended to reflect, in part, the frequency with which the measurements should be made as indicated in the Professional Standards Board (PSB) Accreditation Manual (1984) and, in part, the increased demands for specialized equipment and technical expertise associated with those measurements. It is assumed that the reader is familiar with the first tutorial paper in this series, published by ASHA: Calibration of Pure-Tone Air Conduction Signals Delivered via Earphones (1982). The first tutorial provides information about calibration instrumentation in greater detail. Copies may be obtained directly from ASHA.
The present document is limited to the electroacoustic calibration of speech-audiometer circuits. It applies to sections of the ANSI Standard S3.6-1969; R-1973 which relate to speech audiometers except those sections covered in the previous tutorial. Thus, only the following sections of the Standard will be covered: 5.2, 5.3, 5.4, 5.5, 5.6, 5.7, 5.8. Further, this tutorial deals only with the electroacoustic characteristics of speech signals through earphones. No attempt is made to offer recommended calibration procedures for speech signals in a soundfield. Further, this tutorial paper does not deal with the issue of the specification of the sound pressure level of speech. The reader should be aware that standards are subject to periodic review and revision. Indeed, a revision of the current ANSI Standard (S3.6) on specifications for audiometers is in preparation. However, the methods of checking audiometer performance described in this tutorial should prove to be applicable, in general, to subsequent versions of the Standard.
The format of this tutorial paper is that of a procedures manual for audiometer calibration using equipment available in most clinics. First, the relevant passage from the ANSI standard will be printed verbatim in bold type. Next, the recommended procedures will be given. The procedures described reflect the preferred practices of members of the Working Group on Electroacoustic Characteristics of the Committee on Audiologic Evaluation and of selected reviewers to whom earlier versions of this paper were submitted. The omission of other procedures does not imply that other procedures are less valid or are inappropriate.
All speech test material shall be so presented as to attain a standard reference level, as shown by a monitoring meter provided in the audiometer. The meter shall be connected ahead of the attenuator. For live voice tests, this shall be a meter with characteristics as described in American National Standard Volume Measurements of Electrical Speech and Program Waves, C16.5-1954 (specifications for a VU meter, Sections 3.2 to 3.5, inclusive). Provision shall be made in the amplifier for close adjustment of its gain to attain the desired reference level, and to accommodate differences of 20 dB in the absolute level of presented acoustic material, either recorded or live. If automatic alternative means are provided for governing the reference level, they shall produce the same result as would be obtained with the above arrangement of gain on the basis of a meter reading.
The monitoring meter referred to in this section of the standard is a volume unit (VU) meter. The speech audiometer VU meter pointer should reach 99% of the input signal value in 300 msec (tolerance ± 10%) and then overshoot by 1 to 1.5%. To check the response time of the VU pointer, Wayne O. Olsen suggests producing a calibration tape recording which consists of 1000 Hz tone of sufficient duration to adjust the pointer to 0 dB on the meter followed by 1000 Hz tone durations of 300, 270 and 330 msec (300 msec ± 10%) to ensure by visual inspection that the pointer accurately deflects to 0 dB on the meter (Wilber, 1985). Electronic switches set to instantaneous rise-fall times and interval timing devices may be used to produce the appropriate durations on the VU calibration recording. Accuracy of the VU meter scale at other decibel values may be checked by inserting an attenuator between the signal source and the audiometer input.
For the purpose of this standard, the sound pressure level of a speech signal at the earphone is defined as the rms sound pressure level (as defined in 2.6) of a 1000 Hz signal adjusted so that the VU meter deflection produced by the 1000 Hz signal Is equal to the average peak VU meter deflection produced by the speech signal. The level indication of the VU meter for a preliminary carrier phrase may be taken as the level indication of the immediately following speech material when that material is delivered in a natural manner at the same communication level as the carrier phrase.
The calibrated sound pressure level of speech refers to the rms SPL of a 1000 Hz signal which produces the same average peak VU meter deflection produced by speech. For prerecorded speech audiometer material, a 1000 Hz calibration tone that fulfills this criterion usually is provided directly on the tape or disc. There are a few instances where the calibration tone of prerecorded material is at a different level in relation to the speech. In these cases the sensitivity of the VU meter can be adjusted to compensate for the difference at the time of using the material. In such cases the level should be specified to the level of the stimulus, not to the level of the calibration tone. It is reasonable to assume that the majority of prerecorded material will have the correct relation between the level of the calibration signal and the level of the speech signal. However, it is recommended the intensity level relationship between the calibration signal and speech be checked (an oscilloscope or a level recorder would be appropriate for the measurement). A calibration signal from any prerecorded speech audiometric signal can be sent to zero VU and used for calibration of the speech circuit with a tape or disc input. For a microphone input, an external 1000 Hz tone must be provided. Methods of providing a 1000 Hz tone to the microphone are presented under the “compliance” heading following Section 5.4.
The standard reference threshold sound pressure level for speech shall be as a given in Table 2.[*] The relationship between the elements of the speech circuit shall be such that the scale markings of the attenuator will indicate zero hearing threshold level for speech when a calibrating tone of 1000 Hz brings the monitor meter to its standard reference deflection and simultaneously produces a sound pressure reference threshold sound pressure level for speech as given in Table 2.[*]
In a live-voice audiometer, the amplifier gain shall be adjustable so as to attain the standard reference level for the speech material used as indicated by the VU meter, when the operator speaks in a natural conversational voice, the relationship between the microphone and the mouth of the speaker being that recommended by the manufacturer as to distance and orientation.
The standard reference threshold sound-pressure level (SPL) for speech through an earphone given in Table 2[*] of the standard is 19 dB re 20 µPa. That is, 0 dB on the hearing-level dial corresponds to 19 dB SPL. This reference level is based on the use of the Western Electric, Type 705-A, earphone and the NBS 9-A coupler. Appropriate corrections must be applied to the output for other earphones. The existing ANSI standard does not give reference values for speech for other earphones. The proposed revision of the ANSI standard will likely solve this problem by defining the reference to be 12.5 dB above the pure-tone reference test threshold at 1000 Hz for the specific earphone. The reference thresholds for speech calculated in this manner for various earphones have been listed in Table 1. These were computed by adding 12.5 dB to the pure-tone reference test value at 1000 Hz for each earphone given in Table 1 of the first tutorial paper, Calibration of Pure-Tone Conducted Signals Delivered via Earphones (ASHA, 1982).
|Earphone Type||Reference Level for
1000 Hz dB SPL
|Computed Reference Level
for Speech in dB SPL
|WE 705-A||6.5 +||12.5||= 19.0|
|TDH-39, TDH-39p||7.0 +||12.5||= 19.5|
|TDH-49, TDH-49p||7.5 +||12.5||= 20.0|
|TDH-50, TDH-50p||7.5 +||12.5||= 20.0|
|TELEX 1470-A||6.5 +||12.5||= 19.5|
Reference-level calibration is performed by setting a 1000 Hz signal at 0 VU through the speech circuit and measuring the output SPL of the 1000 Hz signal at a hearing-level dial setting of 60 dB. A ±3 dB tolerance is allowed according to the ANSI S3.6-1969; R-1973 standard. Thus, a range of 76 to 82 dB SPL (79 ±3 dB) is acceptable for type 705-A earphones.
The calibration equipment can be the same as that used for SPL measurements with tones. Table 2[**] can be used to record the data. It is important that the 1000 Hz reference signal be derived from the source being used for speech audiometry. These fall into two categories: microphone (live voice), or prerecorded (tape or disc).
In the case of a microphone circuit, the external calibration tone should be applied within an anechoic space. This requirement (specified later in Section 5.6.1) was established to ensure measurement of the audiometer characteristics independent of the acoustic environment in which the microphone is placed. Clearly, this requirement is difficult to fulfill in many clinical environments because a pure sine-wave field (anechoic space) is not always available. This problem underscores the need to use recorded speech signals whenever feasible. Although there is no clear alternative to this requirement, there are two nonstandard approaches which could be used to calibrate a microphone circuit.
An acoustic signal can be delivered to the microphone with a hearing-aid analysis system. Provided the anechoic test chamber is sufficiently large, the microphone of the audiometer can be placed in the test chamber and a test signal derived from the loudspeaker of the test box can be directed through the audiometer microphone. The control microphone of the hearing aid test box ensures that the level of the test tone at the audiometer microphone is correct.
An acoustic signal can be delivered to the microphone from an auxiliary source. A portable audiometer or external oscillator connected to an earphone or preferably a speaker can be used to deliver a calibration signal to the microphone. The transducer should be at approximately the same distance and orientation relative to the microphone as the talker/microphone arrangement used in speech testing. In addition, the signal source should be relatively free of distortion and noise. The obvious disadvantage of this technique is that the electroacoustic measurements are not being made in a pure sine-wave field. Therefore, the acoustic environment (test room) can influence the output measurements and the measurements of the speech circuit will reflect the characteristics of the acoustic environment as well as those of the audiometer. On two-channel audiometers which have the capability of being set up simultaneously for pure-tone testing on one channel and for speech audiometry on the second channel, one channel may be used as the external source while calibrating the other channel. This arrangement may be reversed so that both channels may be calibrated. A convenient arrangement is to set one channel for live voice testing and the other channel to produce a 1000 Hz tone directed at a high level to the monitor earphone. The monitor earphone is then placed in front of the microphone at the desired distance. The 0 VU level can then be set by adjusting either the level in the monitor circuit, the level of the microphone circuit, or, if necessary, the distance between the monitor earphone and the microphone.
When conducting monitored live-voice testing, acoustic environment of the test microphone should be compatible with the environment for pure-tone threshold testing, i.e., meet the requirements of ANSI S3.1-1977, American National Standard Criteria for permissible ambient noise during audiometric testing.
The frequency response characteristic of a live-voice audiometer shall be such that with the microphone in an acoustic sine-wave field of a given sound pressure level (approximately 74 dB), end in the relationship to the acoustic source which is recommended by the manufacturer as to distance and orientation, the sound pressure level developed by the audiometer at each of the frequencies 200, 300, 400, 700, 1500, 2000, 3000, 4000 Hz does not differ from that at 1000 Hz by more than ±5 dB.
The lack of the anechoic space which caused difficulty in accomplishing the measurement mentioned above in the discussion of Section 5.4 causes the same problems in accomplishing the measurements for compliance with this section. The solution is to use the same schema used in that section. All that is necessary is to implement the measurement described in Section 5.4 above, change the frequency from 1000 Hz to each of the other frequencies (200, 300, 400, 700, 1500, 2000, 3000, and 4000 Hz) and determine if the output is within ±5 dB of the output at 1000 Hz. If a source is used which does not have continuous frequency adjustments (most portable audiometers and some hearing aid test systems), it is probably sufficient to use the closest frequency available (250 Hz for both 200 and 300 Hz, 500 Hz for both 400 and 700 Hz). To ensure the input at each frequency is approximately 74 dB, a monitor microphone should be placed near the audiometer test microphone.
The frequency response characteristic of a recorded speech audiometer shall he such that when used with an appropriate sine-wave recording, as described below, the sound pressure level developed by the audiometer throughout the frequency band from 200 to 4000 Hz shall not differ from that at 1000 Hz by more than ±5 dB, and shall not rise at any frequency outside of this band by more than 10 dB.
Speech audiometers using disc records shall be designed for use with recordings made in accordance with Record Industry Association of America Standards, Standard Recording and Reproducing Characteristics and Dimensional Standards—Disc Phonograph Records for Home Use, Bulletin No. E-4.
It is implied that the pickup will be provided with an appropriate network, such that the overall system shall have the characteristics specified in the first paragraph of 5.6.2, when used with a record in which the radial velocity level of the modulated groove as a function of frequency has the following relative values:
|Frequency, Hz||Velocity Level, dB|
It is very difficult to determine if prerecorded speech material on discs complies with the standard. The ANS1-1969; R-1973 standard does not mention standards for taped material. A reasonable approach in view of this is to purchase prerecorded audio discs or tapes which have calibration tones at the appropriate frequencies. Calibration then involves playing these calibration tones and making measurements similar to that for the microphone circuit.
The standard does require that levels of components lower than 200 Hz or higher than 4000 Hz be no greater than 10 dB above the values between these two frequencies. This can be checked only if the measurement device can be set to measure in different frequency regions. A sound level meter with one octave or with one-third octave filters is sufficient for this purpose.
In a speech audiometer, with a pure-tone input having no harmonic less than 40 dB below the fundamental, the fundamental of the output signal shall be at least 25 dB above the level of any higher harmonic when the output of the amplifier is 6 dB above the standard reference deflection of the VU meter. Test for conformance with this requirement shall be made with the attenuator set to its minimum attenuation or so as to produce a sound pressure level of 120 dB referred to 0.0002 microbar. The tests shall be conducted either at the set of frequencies of 200, 400, 700, 1000, 2000, end 4000 Hz or at the alternative set of frequencies of 250, 500, 1000, 2000, and 4000 Hz. A recorded speech audiometer shall tested with a suitable test recording, such as those described by the Record Industry Association of America. A live-voice audiometer shall be tested with pure tones of the above frequencies supplied to the microphone at a sound pressure level of 74 dB referred to 0.0002 microbar.
As with pure-tone audiometers, it is important to evaluate the overall distortion present in the speech circuit of an audiometer. Like other electroacoustic measurements for the speech circuit, distortion measurements should be made independently for each speech source (microphone, prerecorded tape, or disc) because different circuitry may be involved dependent on the input source.
It is important that the input signal used to measure distortion through a speech audiometer be relatively free of distortion. ANSI specifications require that the pure-tone input have no harmonic greater than 40 dB below the fundamental. Distortion measurements are made with minimal attenuation on the intensity level dial or at 120 dB SPL re 20 Pa. ANSI specifications for overall distortion through a speech audiometer require that the level of the fundamental frequency be at least 25 dB or more above the level of any higher harmonic.
Audiometers can be tested for distortion using commercially available tape or disc recordings. A microphone circuit is tested with pure tones at an input level of 74 dB SPL re 20 Pa. The scheme used should be that suggested for reference-level and frequency-response measurements. A sound level meter with one octave or one-third octave filters is sufficient so long as filter roll-off is greater than 25 dB per octave (preferable equipment would include a spectrum or wave analyzer). Adjust the HL dial so that each tone produces 120 dB SPL. Then read the levels at each of the harmonic intervals above the test frequency. These levels should be 95 dB SPL (120 minus 25) or less.
Dial readings of the hearing threshold levels for speech shall extend at least from 0 dB to 100 dB in steps of 2.5 dB or less, with the 0 dB setting corresponding to the reference speech threshold level as given in Table 2.[*] The measured difference between two successive designations shall not differ from the dial-indicated difference by more than 1 dB. Measurements for compliance with this requirement shall be made electrically at the input of the earphone, with the earphone attached to the coupler, using a pure-tone signal of 1000 Hz, or the earphone may be replaced by a dummy load which simulates the earphone electrically.
The sound pressure levels produced by the earphone as referred to the standard reference threshold level shall not differ from the indicated values by more than 3 dB. Measurements for compliance with this requirement shall be made at 1000 Hz, and may be made by combining an acoustical measurement of sound pressure level at a 60 dB dial setting with the results of interval measurements made under 5.7.1
As with the pure-tone circuit of an audiometer, the attenuator of a speech circuit must have a linear characteristic. The changes in the output level of the speech audiometer shall correspond closely to the level increments indicated on the intensity-level dial. The speech audiometer hearing-level dial must extend at least from 0 to 100 dB in steps of 2.5 dB or less. ANSI specifications require that the dial increments and measured increments shall not differ by more than 1 dB. One should note that attenuation steps are 2.5 dB or less and therefore the linearity requirements of “successive designations” (steps) being accurate within ± 1 dB is not a more stringent requirement than that for pure-tone circuits. Measurements of attenuator linearity for speech are performed with a 1000 Hz reference signal using the same procedure and instrumentation used for evaluating the hearing-level attenuator of a pure-tone audiometer.
Most audiometers use the same attenuator for both pure-tones and speech signals. If this is the case and the attenuator has been calibrated for pure-tones as specified in the first tutorial paper, Calibration of Pure-Tone Air Conducted Signals Delivered via Earphones, then it is not necessary to recheck the attenuator for speech signals.
The electrical background noise from all sources other than surface noise of the recordings shall be at least 50 dB below the level of the signal. This measurement for noise shall be made in the following manner:
The amplifier shall be adjusted so that the meter indicates the reference level when the input is a 1000 Hz signal from a record or in the case of a live-voice speech audiometer, when there is delivered to the microphone a 1000 Hz signal at a sound pressure level of 85 dB. The output sound pressure level of the audiometer shall be measured with the above input signal and with the attenuator at the 100 dB hearing threshold level setting.
Secondly, the output sound pressure level shall be measured with no signal input to the audiometer, and with the attenuator set at the 100 dB hearing threshold level. In testing an audiometer with a mechanical reproducer, the pick-up is placed in the “rest” position, but the turntable is allowed to revolve. In testing with a magnetic-tape playback, the mechanism is activated but no tape is run across the pick-up. In testing with a live-voice type of audiometer, the microphone shall be protected as completely as is feasible from an acoustic input, or a dummy microphone may be used.
The pressure level measured under condition (2) above shall be at least 50 dB below the pressure level measured under condition (1).
A speech audiometer shall also meet the requirements of 4.4 as far as these are applicable to a speech audiometer.
The outputs of the speech circuit should be free of background noise. ANSI specifications require that the background noise, other than the surface noise of recordings, shall be at least 50 dB below the signal level. Measurements are made with and without a 1000 Hz input signal. The gain control is adjusted so that the VU meter indicates the reference level for the 1000 Hz signal. For a microphone circuit, the 1000 Hz signal delivered to the microphone is set at 85 dB SPL. Signal to noise ratios are determined by comparing the output levels with input and with no input. In both cases, the intensity-level dial is to be set at 100 dB. With the signal off (turned off for microphones, tone arm lifted for discs, tape transport on “stop” for tape), the output level should be 70 dBA or less.
American Speech-Language-Hearing Association. (1984). Professional services board standards for accreditation on professional service programs in speech-language pathology and audiology. Rockville, MD: Author.
International Electrotechnical Commission. (1980). Magnetic tape recording and reproducing systems. Part III: Methods of measuring the characteristics of recording and reproducing equipment for sound on magnetic tape (IEC 94-3-1980). Geneva. Switzerland: Author.
American Speech-Language-Hearing Association Working Group on Electroacoustic Characteristics of the Committee on Audiometric Evaluation. (1982). Calibration of pure-tone air-conducted signals delivered via earphones. Rockville, MD: Author.
Index terms: instrumental assessment
Reference this material as: American Speech-Language-Hearing Association. (1987). Calibration of speech signals delivered via earphones [Relevant Paper]. Available from www.asha.org/policy.
© Copyright 1987 American Speech-Language-Hearing Association. All rights reserved.
Disclaimer: The American Speech-Language-Hearing Association disclaims any liability to any party for the accuracy, completeness, or availability of these documents, or for any damages arising out of the use of the documents and any information they contain.