EN ISO 8253-3:1998

2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.Akustika - Avdiometrijske preskusne metode - 3. del: Govorna avdiometrija (ISO 8253-3:1996)Akustik - Audiometrische Prüfverfahren - Teil 3: Sprachaudiometrie (ISO 8253-3:1996)Acoustique - Méthodes d'essais audiométriques - Partie 3: Audiométrie vocale (ISO 8253-3:1996)Acoustics - Audiometric test methods - Part 3: Speech audiometry (ISO 8253-3:1996)17.140.01Acoustic measurements and noise abatement in general13.140Vpliv hrupa na ljudiNoise with respect to human beingsICS:Ta slovenski standard je istoveten z:EN ISO 8253-3:1998SIST EN ISO 8253-3:1999en01-november-1999SIST EN ISO 8253-3:1999SLOVENSKI
STANDARD
@ IS0 IS0 8253=3:1996(E) The interval between the test items depends upon the purpose of the speech test and the subject’s response mode and shall be stated. National authorities may determine the appropriate intervals (see 4.6). NOTE - The speech recognition score of a test subject may depend on the intervals between the test items. To make test results comparable to corresponding reference values, intervals should be used identical to those used during the determination of reference values. 11 Determination of speech detection threshold level 11 .l In monaural testing, start with the ear considered to be more sensitive. 11.2 Using connected speech as the test signal, let each test item be a segment of 1 s to 2 s duration, controlled by means of the interrupter switch of the equipment. The first item shall be presented at a sufficiently high level to evoke a definite response; e.g. 30 dB above the average of the subject’s pure tone hearing threshold levels at 500 Hz, 1 000 Hz and 2 000 Hz. 11.3 Reduce the level in steps of 20 dB until no response occurs. 11.4 Increase the level of the speech signal in steps of 5 dB and present one test item on each level until a response occurs. 11.5 After the response, decrease the level by 10 dB and then begin another ascent using steps of 5 dB until a response is given. 11.6 Repeat step 11.5 until three responses have occurred at the same level out of a maximum of five ascents. This level is the speech detection threshold level. 11.7 If less than three responses out of five ascents have been obtained at the same level, present a test item at a level 10 dB higher than the level of the last response. Then repeat steps 11.4 to 11.6 inclusive. 11.8 Proceed with the other ear if required. 11.9 A shortened version of this ascending method may be assumed to yield nearly equivalent results and may be appropriate in some cases. In this shortened version, continue the testing according to step 11.6 until at least two responses occur at the same level out of a maximum of three ascents. 12 Determination of speech recognition threshold level 12.1 General The procedures described below are based on two widely used step-sizes: 5 dB and 2 dB. Two alternative procedures are described: - a descending procedure using 5 dB steps, and - an alternative descending procedure using 2 dB or 5 dB steps. The procedures are expected to yield comparable results. However, experimental evidence for this is still unavailable. The procedures are identical from 12.1 .I to 12.1.3. Then, the descending procedure using 5 dB steps continues with 12.2 and the alternative procedure continues with 12.3. 9 SIST EN ISO 8253-3:1999

IS0 8253-3: 1996(E) NOTE - Spondees are often used as test items in many languages. 12.1 .l In monaural testing, start with the ear considered to be more sensitive. 12.1.2 The test subject shall be familiarized with the task prior to threshold determination by the presentation of a number of test items at a sufficiently high level to be clearly audible. A hearing level of speech of 20 dB to 30 dB above the average of the subject’s pure tone hearing threshold levels at 500 Hz, 1 000 Hz and 2 000 Hz is generally recommended. 12.1.3 Reduce the speech level in steps of 5 dB, presenting at least two test items on each level, until the test subject no longer responds correctly to all test items. Continue according to 12.2 or 12.3. 12.2 Descending procedure using 5 dB steps 12.2.1 Present a set of test items at the level where the subject ceased to respond correctly according to 12.1.3 and record the number of correct responses. A set of test items shall contain at least ten items. 12.2.2 If the subject scores at least 50 % on the set of test items, reduce the level in steps of 5 dB and present a new set of test items on each level until the subject scores less than 50 % on the set of test items. Usually one level is found to yield somewhat more than 50 % and the next lower level somewhat less than 50 %. If the subject scores less than 50 % on the set, increase the level in steps of 5 dB and present a new set of test items on each level until the subject scores more than 50 % on the set of test items. 12.2.3 The speech recognition threshold level for the ear is the integer value of the level corresponding to a score of 50 % correct as calculated by means of linear interpolation between the lowest level that yielded more than 50 % correct responses and the highest level that yielded less than 50 % correct responses. If a score of exactly 50 % is obtained at one speech level, that level represents the speech recognition threshold. 12.3 Alternative descending procedure 12.3.1 When one test item is missed, present a second at the same level. Continue this descending process in steps of 10 dB until a level is reached at which two consecutive test items are missed at the same level. 12.3.2 Increase the speech level by 10 dB. This defines the starting speech level, L,. 12.3.3 Present two test items at this starting speech level and at each successive 2 dB decrement if a step size of 2 dB is used. For a step size of 5 dB, see step 12.3.6. 12.3.4 Continue this process if at least five out of the first six test items are repeated correctly. If this criterion is not met, increase the starting speech level by 4 dB to 10 dB. 12.3.5 The descending series is terminated when the test subject responds incorrectly to five of the last six test items presented. 12.3.6 If a step size of 5 dB is used, five test items should be presented at each level and the test terminated when all test items at a single level are incorrectly recognized. 10 SIST EN ISO 8253-3:1999

@ IS0 IS0 8253=3:1996(E) 12.3.7 Calculate the speech recognition threshold level, T, in decibels, according to the following equation: T=L, - d-r/n + d/2 where Ls is the starting speech level, in decibels, according to step 12.3.2; r is the total number of correct responses given from the starting speech level according to step 12.3.3 up to the termination according to step 12.3.5 or 12.3.6; d is the step size, in decibels; n is the number of test items per level. 13 Determination of speech recognition scores If speech recognition scores are to be determined at more than one level (e.g. to determine a speech recognition curve), test items should not be repeated within the same session. If a test is not preceded by a determination of speech recognition threshold, the test subject shall be familiarized with the task prior to the determination of a score by the presentation of a number of test items at a sufficiently high level to be clearly audible. Usually, a hearing level for speech of 30 dB to 40 dB above the subject’s average pure tone hearing threshold levels at 500 Hz, 1 000 Hz and 2 000 Hz is a good choice. Set the audiometric equipment to the speech level required and present a complete test list at each level. NOTE - The use of only a part of a test list may reduce the reliability and the validity of the test. The choice of test level(s) is governed by the particular purpose of the test, as follows. a> b) c) To determine a maximum speech recognition score, a first score is normally determined at a speech level of 25 dB or 30 dB above the speech recognition threshold level. The level should then be increased in steps of 5 dB or 10 dB until a maximum score has been found or the subject reports discomfort or fatigue. If the score decreases at higher levels (the roll-over effect), the test shall be continued at lower levels. To determine a speech recognition score at the most comfortable loudness level, the test level shall be selected by presenting continuously a speech signal of the same type that is to be used for the actual test. Instruct the subject to report on its loudness (e.g.much too low, comfortable, too high, much too high). Often, halfway between too low and too high is a good initial choice for the most comfortable level. To determine a half-optimum speech level, the maximum speech recognition score first has to be determined. Reduce the level in steps of 5 dB or 10 dB and present a full test list on each level until one level yields a score above half of the maximum score and the next lower level yields a score below half of the maximum score. Determine the integer value of the half-optimum speech level by means of linear interpolation between the latter two test levels. Express the score obtained as a percentage and record the level at which it was achieved. 14 Contralateral masking To avoid hearing the speech signals in the non-test ear during monaural speech audiometry, it may be necessary to apply air-conducted masking noise to the non-test ear. Specifications for the masking noise are given in IEC 645-2: 1993, clause 13. 11 SIST EN ISO 8253-3:1999

IS0 8253-3: 1996(E) @ IS0 Masking shall be used in air-conduction testing if the speech signal level, expressed as the hearing level for speech, exceeds by 40 dB or more the average bone-conduction hearing threshold levels for two of the frequencies out of 500 Hz, 1 000 Hz and 2 000 Hz of the contralateral ear showing the lowest hearing threshold levels. Masking shall always be used in monaural bone-conduction testing. The minimum masking level for air conduction testing L, 1 expressed as the effective masking level, shall be: L,=h-40 dB+ (MAm -MB,,.,) Lt is the speech level, expressed as the hearing level for speech; MAm is the average pure tone air-conduction threshold level of the ear to be masked at two frequencies out of 500 Hz, 1 000 Hz, and 2 000 Hz, having the lowest hearing threshold levels, expressed as the hearing level; MBfll is the average pure tone bone-conduction threshold level of the ear to be masked at two frequencies out of 500 Hz, 1 000 Hz, and 2 000 Hz having the lowest hearing threshold levels, expressed as the hearing level; NOTES 1 In some cases, the masking level may have to be increased to assure sufficient masking of the non-test ear. 2 In bone-conduction testing, the minimum masking level is 40 dB higher than L,. 15 Speech recognition with competing sound 15.1 Type of competing sound A frequency-weighted unmodulated random noise in accordance with IEC 645-2:1993, subclause 13.1 may be used as a competing sound. If an amplitude-modulated or any other noise is used, its characteristics shall be specified. NOTE - It is recognized that competing sounds using recordings of human speech may also be used, such as multitalker babble (cocktail-party noise) or connected speech from a single speaker. With such types of noise, larger variation in test results may occur than with weighted random noise. 15.2 Presentation of competing sound In earphone testing, the competing sound is presented by means of the same earphone as used for the speech test signal. In sound field audiometry, the recommended position of the speech loudspeaker is frontal to the listener at an angle of incidence of 0’. It is recommended that the competing sound should be generated from two loudspeakers symmetrically located at the angles of incidence of - 45’. If the competing sound is a random noise, presented from two or more lateral loudspeakers, the noise from the loudspeakers shall be non-coherent. If, for particular purposes, be clearly specified. other angles of incidence ( cl e. . above the head of the test subject) are used, they shall 12 SIST EN ISO 8253-3:1999

@ IS0 IS0 8253-3: 1996(E) 15.3 Speech and competing sound levels The levels of speech and competing sound shall be measured as specified in 3.9 and 3.20, using the same frequency and time weightings, respectively. The recommended speech level is 65 dB, which approximately corresponds to normal speech level in conversation. If other speech levels are used, they shall be clearly stated. The level of the competing sound may be either fixed or variable. If fixed, the recommended noise level is 60 dB. If varied, it shall be changed in steps of 5 dB or less. 15.4 Test procedure The test may be performed either to determine a speech recognition score at one or more fixed speech-to-noise ratios or to determine the speech-to-noise ratio at the speech recognition threshold. 15.4.1 Speech recognition score at fixed speech-to-noise ratio 15.4.1 .l Set the audiometric equipment to the required speech level. 15.4.1.2 Familiarize the test subject with the task by setting the competing sound level low (e.g. at 20 dB below the speech level) and presenting a number of test items which are clearly audible. 15.4.1.3 Set the competing sound level to the value required for the test and present a complete test list. Calculate the score as a percentage. 15.4.2 Speech recognition threshold with a competing sound 15.4.2.1 General In the procedures described, increasing the level of the competing sound may be replaced by decreasing the level of the speech signal and vice versa. The test result may be influenced by the alternative chosen. Alternative procedures are described for determining the speech-to-noise ratio that yields the speech recognition threshold at a given speech or competing sound level. They are expected to yield comparable results. However, experimental evidence for this is still unavailable. The procedures are identical from 15.4.2.1 .I to 15.4.2.1.2. Then, the procedure to decrease speech-to-noise ratio using 5 dB steps continues in 15.4.2.2 and the alternative procedure continues in 15.4.2.3. 15.4.2.1.1 Follow steps 15.4.1.1 and 15.4.1.2. 15.4.2.1.2 Increase the competing sound level in steps of 5 dB or less, and present at least two test items at each level until the test subject incorrectly recognizes one test item. 15.4.2.2 Procedure to decrease speech-to-noise ratio using 5 dB steps 15.4.2.2.1 Present a set of test items at this speech-to-noise ratio and record the numbe
...