Some standard content:
GB/T176961999
This standard equivalently adopts the international standard ISO8253-3:1996 "Acoustic audiometry methods Part 3: Speech audiometry". This standard summarizes the long-term rich experience of various countries in language audiometry, stipulates the necessary conditions and examination methods of language audiometry, and makes the language audiometry technology have a norm to follow. It has a positive role in promoting the standardization of language audiometry methods, the development of audiology work in my country and international exchanges.
This standard, GR/16403-1996 and GB/T16296-1996 constitute the "audiometry method" standard series. This standard and GB/T15508-1995 "Acoustic speech intelligibility measurement method" can be used in appropriate circumstances. Appendix A and Appendix B of this standard can be used in conjunction with each other. This is an appendix to the instructions. This standard was proposed and approved by the National Technical Committee for Acoustic Standardization. The drafting units of this standard are the Institute of Otolaryngology, General Hospital of the People's Liberation Army, the Institute of Acoustics, Chinese Academy of Sciences, and the Institute of Psychology, Chinese Academy of Sciences.
The main drafters of this standard: Gu Rui, Zhang Ruwei, Fang Zhi, Chen Hongyi, Wang Naiyi: GB/T176961999
ISO Foreword
The International Organization for Standardization (ISO) is a worldwide joint organization composed of national standardization committees (ISO member countries). The formulation of international standards is usually completed by ISO technical committees. Each member country has the right to participate in a technical committee when it is interested in a standard determined by a technical committee. Governmental and non-governmental international organizations related to them also participate in this work. The International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC) maintain close cooperation in all aspects of electrotechnical standardization. Draft international standards adopted by each technical committee should be distributed to each member country. The draft international standard must be voted on by all member countries before it can be published as an international standard. International Standard IS082533 was drafted by ISO/TC43 Acoustics Technical Committee Basic Subcommittee. The general title of IS08253 is "Acoustics - Audiometry", which includes the following parts: Part 1: Pure tone air conduction and bone conduction audiometry; Part 2: Sound field audiometry using pure tone and narrowband test signals; Part 3: Speech audiometry.
This standard recommends the following annexes and appendices B Suggestive, for reference GB/T17696-1999
Speech audiometry is used for clinical diagnosis of hearing sensitivity, hearing rehabilitation and assessment of hearing impairment. The choice of language test materials and sound delivery methods depends on the specific purpose of the test.
1 Scope
National Standard of the People's Republic of China
Acoustic Audiometry Methods
Part 3: Speech Audiometry
Acoustics--Andiometric testmethods..Part 3: Speech audiometry
GB/T:17696- 1999
fqvIso8253-3:1996
This standard specifies the methods and necessary conditions for air conduction, bone conduction or sound field speech audiometry using test materials recorded by headphones, bone vibrators or physical sounders, and describes the methods for using noise to mask non-test noise or as competition. This standard also includes the conditions for recording speech materials and recommended methods for maintaining and calibrating speech audiometry equipment. This standard applies to Mandarin Chinese language audiometry and local dialect language audiometry. For some subjects, such as children, it may be necessary to use special examination methods, which are not described in this standard. This standard is not applicable to live voice language examination methods. In the informative appendix, some suggestions are given to help oral voice examinations obtain as reliable results as possible.
Due to the diversity of human languages: This standard does not make specific provisions on the content of language materials. Different types of examination materials have a significant impact on the results of language audiometry.
This standard also includes some special examination methods, such as directional and binaural listening examinations. 2 Referenced standards
The provisions contained in the following standards constitute the provisions of this standard through reference in this standard. When this standard was published, the version shown was effective. All standards will be revised. The parties using this standard should explore the possibility of using the latest version of the following standards. GB/T 3240-1982 Frequency band in acoustic measurement (eqIS(266-1975> GB/T 7341.1-1998 Listening Part 1: Pure audiometern
The process by which sound is transmitted from the outside to the inside of the air. 3.2 Bone conduction bone tonduclian
The mechanical vibration of the skull is stimulated to transmit sound to the inside. 3.3 Otalogically normal μerson Normal health condition, no ear disease, no blockage of the ear canal, no history of excessive noise exposure, no toxic drugs or family hearing loss Approved by the State Administration of Quality and Technical Supervision on March 8, 1999 and implemented on September 1, 1999
3.4 Free field frec[sound]field
GB/T 17696 --1999
A sound field in a homogeneous isotropic medium where the boundary effect can be ignored. 3.5 Quasi-free field quasi-free [sound]field A sound field where the room boundary has a moderate effect on the sound wave, which should meet the requirements of 5.3 in GB/T 1626---1996. 3.6 Diffuse [sound] field diffuse _sound field The energy density is uniform and the sound field is randomly distributed in all propagation directions. 3.7 Reference point reference point The midpoint of the straight line connecting the two external auditory canals of the subject when the subject is in the test position in the sound field: 3.8 Speech signal speech signal
The test signal produced by speech or synthetic speech. 3.9 Speech level spcechlevel
The sound positive level or vibration level of the speech signal measured in an appropriate coupler, artificial ear or in the sound field with specified frequency weighting and specified time weighting. For example, the speech level can be expressed as the equivalent continuous sound pressure level or vibration level calculated for the speech signal over a period of time using constant weighting. The silent intervals between each check item and the introductory sentence at the beginning of the word list are not included in this time schedule. For a single check item, the equivalent continuous sound pressure level can be estimated by subtracting 5% from the average of the maximum sound pressure levels obtained by C frequency weighting and 1 time weighting.
3. 10 Speech detection threshold level The language level of the test material when a specific subject can detect (not understand) a given test number in a specified transmission mode and a specified language pair. Note: Speech detection was once called speech awareness threshold. 3.1 Speech recognition rate Recognition score is the percentage of actual test items that can be correctly recognized at a specified language level using a specified language signal and a specified transmission method. Note: The language recognition rate was once called the "language discrimination rate (disrriminatinn)" \ 3.12 Maximum speech recognition score is the maximum speaker recognition rate obtained by subtracting 100 from the maximum speech recognition rate for a subject, regardless of the language level. The difference is the language recognition loss, which was previously called the "language discrimination loss". 3.13 Speech recognition threshold is the minimum language level at which a subject can obtain a language recognition rate of 50% using a specified language signal and a specified transmission method. Note: The language recognition threshold was once called the "language receptive exhaustion". 3. 14 Reference speech recognition threshold lcvel is the median value of the speech recognition level obtained by using a specified transmission method and a specified speech signal for a sufficient number of normal otological subjects of both sexes between the ages of 18 and 25, and the test materials are composed of the language that the subjects are proficient in. 3.15 Maximum speech level is the language level at which the maximum speech recognition rate is obtained for a subject with a specified speech signal and a specified transmission method. 3.16 Half-optimal speech level is the speech level at which the maximum speech recognition rate is obtained for a subject with a specified language signal and a specified transmission method. Note: The half-optimal language level is also called the "peak level". 3.17 Speech recognition curve is the function curve of the subject's individual speech recognition rate and language level obtained for a specified language signal and a specified transmission method. CB/T 176961999
Note: The line for speech recognition is a function of the definition function. 3.18 Reference speech recognitian curve The curve of the median speech recognition rate and speech level obtained from a sufficient number of normal otologists of both sexes aged 18 to 25 years for a specified speech signal suitable for the subject and a specified transmission mode. Note: The curve can be affected by using the method of calculating the reference speech recognition curve from a separate data set. 3.19 Hearing level for speech For a specified speech signal and a specified transmission mode, the speech level is subtracted from the corresponding reference speech recognition level. Note: The speech level is expressed as L and the hearing level is expressed as HL. 3.20 Competing sound level The sound pressure level or vibration level of the competing sound measured in the ear or in the sound field with a specified frequency weighting and a specified time weighting in an appropriate coupler.
: Competing sound is a sound signal sent to compete with the speech test signal during audiometry. 3.21 Speech noise level specchtonoiseleve!diffcrenec The difference between the speech and competing sound measured with the same weighting as 3.9 and 3.20. Note: This is called "speech noise level", and in this part of GBT16403 it is called "speech medium sound level difference". 3.22 Masking
a) The phenomenon that the hearing of a sound is increased due to the presence of another masking sound. b) When measuring the hearing of an ear, a method of adding noise to the ear to avoid affecting the ear. Note: From a quantitative point of view, the masking effect of speech is the percentage increase in the speech recognition level of an ear. 3.23 Effective masking level for speech The level of the masking sound that is equivalent to the speech hearing level of the speaker when the normal person's speech recognition level for a given speech signal is increased due to the presence of the masking sound. The normal person referred to here is one whose hearing meets the audiometry standard of ISU 389 and whose masking effect meets the requirements of ISO 4854.4. 3.24 Most comfort level for speech speech - the level of speech at which the subject considers the loudness of the speech signal to be most comfortable when the subject is given a specified mode of speech delivery. 3.25 connecledspeech
uninterrupted speech with natural intonation and coherent sentences with logical connection. 3.26 prndcc
a disyllabic word with equal intensity for each syllable
3.27 nonsense syllablelogatom
a monosyllabic or polysyllabic unit of speech without phonetic meaning to the listener. 3.28 checklist
a monosyllabic or polysyllabic word or nonsense syllable or a short segment of continuous speech of limited duration that is delivered and scored according to the correct rules in speech audiometry. Note: The entire test item or a part of the test item that is identified by the researcher may be used as the basis for scoring. 3.29 checklist li4t
A test item is composed of several selected test items as an independent unit for delivery and scoring. 3.30 A set of test items consists of at least one test item selected from a test list. 3.31 A phonrmirally halauced lesl list contains test items that have the same frequency of occurrence as the various phonemes that exist in normal oral communication in the same language. Note that the adjective "linguistically" has been used. 3.32 Guide words catricrphras
Phrases or phrases added before the language audiometry test items, whose structure or meaning does not affect the correct identification of the test items, GB/T17696-1999
3.33 Open test materials apen sct tus rualeriat The number of optional responses for each test contained in the test item group is unlimited. 4 Necessary items for the preparation of language materials
4. 1 Recording specifications
Each recording product shall include the following signals:
a) Language test materials.
6) Signals for the calibration of the language audiometer of not less than 605 The calibration signal shall be a weighted random sound in accordance with the provisions of GB/17341.2-199813.1, a noise band with a bandwidth of 1/3 octave centered at 1kHz, or? c) Signals for checking the frequency response of speech audiometers and recording and playback equipment. The duration should be no shorter than 15°, including 1/3-octave concave noise with a preferred 1/3-octave frequency range of 125-8000 Hz in accordance with GB3240. d) Signals for checking the harmonic distortion of speech audiometers are pure tones of 250 Hz, 500 Hz and 1000 Hz with a duration of no shorter than 60°, and the peak sound pressure level is equivalent to the highest peak sound pressure level of the recorded language inspection material. Note: The recording may also contain signals used as a part of the audio recording system as specified in Article 15 of this Convention. 4.2 Level
4.2.1 Language inspection material
The sound level of the language inspection material shall meet one of the following requirements! a) Equal language level method. The average language level of the proposed minimum check item group measured at the preset sound pressure level shall not differ from the average of the check items in the entire check list by more than ±1 dB. The language level of any individual check item shall not differ from the average of all the check items in the list by more than ±3 dB. The average language level of the check list shall not differ from the average of all the check items in the recording by more than ±1 dB. (b) Language recognition benchmark level method. The average language recognition benchmark level of the proposed minimum check item measured at the preset sound pressure level shall not differ from the average of all the check items in the list by more than ±1 dB. The language recognition benchmark level of any individual check item shall not differ from the average of all the check items in the list by more than ±3 dB. The average language recognition level of the check list shall not differ from the average of all the check items in the recording by more than dB.
4.2.2 Calibration signal The sound level of the calibration signal shall meet one of the following two requirements a) Language recognition method. The calibration signal level should not differ from the average level of all inspection items measured in 1.2.1a) by more than ±0.5dB. The calibration signal level should be measured with the same frequency and time weighting as the measured language level: b) Language recognition reference reading method. The relationship between the sound level of the calibration signal and the reference level of speech recognition should be the same as that specified by the manufacturer, with a tolerance of 2.5 dB. 4.3 Noise ratio When recording, the sound level of any inherent background noise measured at the same frequency and time as when measuring the speech level should be at least 40 dB lower than the calibration signal that meets 4.2-2. 4.4 Inspection room The ambient noise in the room where speech is recorded should be low enough, with a noise ratio of at least 40 dB. The reverberation time at any frequency in the range of 125 to 8000 Hz should be less than: 4.5 Frequency response of recording equipment The frequency response of recording equipment including microphones: The deviation in the frequency range of 125 to 8000 Hz is within 2 dB. 4.6 Interval time between consecutive inspection items The interval time between consecutive inspection items should be basically constant, with a variation ratio within 10%. Note: For monosyllabic microphones, the interval should not be less than 4%. 4.7 Instructions
Each copy of the recording should be accompanied by an instruction manual containing the following: a) the recording of the language material, b) the scoring method.
GE/T 17696-1999
e) the reference speech recognition curves for each recording, under monaural and binaural listening conditions and the scoring method, including the range of variables, the source of the reference curves and detailed information on the measurement conditions (i.e. the number of subjects, age range and listening level range, and the speech audiometry equipment used).
d) the calibration signal, and the time course, spectrum and relative level of other signals recorded to check the performance required by 4.1 b), c) and d).
e) When the sound level of the language test material is adjusted according to the equal speech recognition level method, the relationship between the speech level of each test item and the total base level.
f) When the sound level of the language test material is adjusted according to the equal speech recognition level, the sound level of the calibration signal and its relationship with the reference speech recognition level.
g) The interval between consecutive test items. h) The frequency and time of the quantifiable speech level. 5 Language signal level
The language signal level of the specific test material used. It can be expressed in terms of speech level or speech listening level, see GB/T7341.2-1998 Chapter 3
6 Masking sound level
GB/T 7341.2-1998 Section 13 specifies the reference conditions for masking noise. The effective masking level of a masking sound with a certain spectrum and sound pressure level on speech depends on the properties of the speech signal. Therefore, the generally effective masking level cannot be determined. 7 Ambient sound pressure level in the examination room
The ambient sound pressure level in the examination room should not mask the language signal: The requirements for the ambient sound pressure level in the examination room are determined by the way the language signal is sent, that is, through headphones, cameras or speakers. Note: The requirements for ambient noise in the language test are not as strict as those in pure tone audiometry. It is suitable for pure tone audiometry in the entire frequency range down to a certain level. The examination room can be used for language audiometry down to the same hearing level. GB/T1619 and GB/T16296 respectively stipulate the maximum allowable ambient sound pressure level for absolute audiometry and acoustic audiometry using headphones and bone vibrators. 8 Sound field language audiometry
Sound field language audiometry can be added or not according to its examination purpose. It is recommended to place the language transmission speaker at a 0-degree angle in front of the receiver. If the test subject's speech audiometry results are to be compared with the reference speech recognition level or reference speech recognition curve, the speaker position should be the same as when the reference value was obtained. If noise is used as a noise source to compete with the speech signal, the noise source speaker position recommended in 15.2 should be used. The type of noise source (e.g., free field, neutral field, or diffuse field as described in Chapter 5 of GB/T16296-1996) and the position of the reference point should be stated. bzxZ.net
9 Preparation and guidance of test subjects
Generally, pure tone audiometry is performed before speech audiometry. 9.1 Preparation of test subjects
Recent exposure to noise may cause a temporary increase in the hearing threshold level. Therefore, obvious exposure to noise should be avoided before the audiometric examination, otherwise the following should be noted: To avoid excessive fatigue, the test subject should rest for at least 5 minutes before the test. Usually, a qualified person should perform an otoscope before the audiometry test. If there is any cervical congestion blocking the ear canal, it should be cleared and the audiometry should be delayed for a period of time: GB/T 17696-1999
Check whether the earphones are worn correctly. The ear-pressing type should not flatten the ear canal. Note: The requirements for qualified personnel shall be determined by the national competent authorities or other appropriate institutions. Before the test, the subject's ability to understand the test materials should be confirmed. If the test materials are responded to verbally, it should be confirmed that the subject can repeat the test materials.
9. 2 Instructions to the subject
In order to obtain reliable results, it is very important to instruct the subject on the test method with clear and easy-to-understand explanations. The instructions should be given in a language that the subject is familiar with. The content depends on the type of speech test to be conducted. Generally, it should explain: a) which ear to test first; b) how to respond to the type of test items; c) respond to speech in either ear, no matter how soft it sounds; d) respond immediately to each test item; e) if responding with spoken words, repeat the item clearly, even if you are not sure whether you heard the correct item, you should say it. Each test item can only be responded to once, and the subject should be warned to avoid unnecessary activities to avoid unnecessary noise. After giving the instructions, the examiner should check whether the subject has understood. If there is any doubt, the instructions should be repeated again. The subject should be aware that the test can be interrupted if there is any discomfort: 10. The subject's response method
The subject's response is usually spoken. But it can also be responded by writing or pressing keys. When the examiner and the subject are not in close contact, the subject's response is not clear.If the test is conducted in spoken language, a talkback system is required. The test subject's spoken response should be clearly understood by the examiner. If the test subject cannot speak clearly, another response method should be selected. The examiner should be familiar with the language of the test material and have sufficient hearing to understand the test subject's voice. The interval between test items determines the length of the language test and the test subject's response method. The interval time is explained to the test subject (see 4.6).
Note: The language recognition rate of the test is related to the time between the test items. If the test result is to be compared with the baseline value, an interval equivalent to the baseline value should be used.
11 Determination of speech perception level
11.1 When performing a single-audit test, start with the ear with better hearing. 11.2 When using continuous speech as a test signal, divide the test item into sections with a duration of 1 to 2 3. Use the device's group switch to control the sections. The first test item should be sent at a sound level high enough to cause a positive response; that is, 30 dB above the average pure tone listening level of the subjects at 500 Hz, 100 Hz and 2000 Hz. 11.3 Reduce the sound level in steps of 20 dB until there is no response. 11.4 Increase the speech signal level in steps of 5 dB, assign 5 dB to each test item, and make sure that the test item is responded to. 11.5 After a response is made, reduce the level by 10 dB and continue to increase it in steps of 5 dB until a response is made. 11.6 Repeat step 11.5 until a response is made at the same level 3 times out of a maximum of 5 increases. This is the speech perception level. 11.7 If there are less than 3 responses at the same level in 5 steps, add 10 dB to the level of the last response in the above test and repeat steps 11.4 to 11.5. 11.8 If necessary, continue to test for another year.
11.9 The simplified form of the rising method can obtain results close to the rising method and is suitable for general situations. This simplified form is that when the steps 11.6 are operated in sequence, there are at most 2 responses at the same level in 3 steps. 12 Determination of speech recognition level
12.1 Overview
GB/T176961999
The following steps are based on the two most commonly used grading levels, 5-dB and 2dB. This chapter introduces two optional procedures:
using the 5 dB step-down method;
using the improved step-down method in 2 dB or 5 dB steps. These two procedures should produce similar (comparable) results. However, they have not been confirmed by experimental records. The two procedures from 12.1 to 12.1.3 are the same. Then use the 5 dB sample method and continue with 12.2, while the improved method continues with 12.3. The construction of the language recognition reading level check uses the word field as the blank item. 12.1.1 For single ear examination, start with the better hearing environment: 12.1.2 Before starting to determine the reading value, use some sound levels that are high enough to be heard to let the subject know how to cooperate. Generally, the average of the subjects' pure tone hearing levels at 500Hz, 1000Hz and 2000Hz is used as the language hearing level, which is 20-30dB higher than the average value. 12.1.3 The language level is lowered in 5dB steps. At least two test items are given for each level, until the subjects can no longer respond correctly to all the test items given for that level. Then, according to 12.2 minus 12.3, and then 12.2 is lowered in 3dB steps. 12.2.1 For the language level at which the subjects can no longer respond correctly to all the test items in 12.1.3, a group of test items is given, and the number of correct responses is recorded. This group of test items must be at least 10. 12.2.2 If the subject scores more than 50 points for this group of test items, the language level is lowered in 5dB steps: a new group of test items is given for each level, until the subject scores less than 50 points for the group of test items. If the subject scores less than 50% on a set of items, the language level is increased by 5 to 15 steps and a new set of items is given each step until the subject scores more than 50% on a set of items. 12.2.3 The speech recognition grade is the integer that is calculated by linear interpolation between the lowest grade that produces more than 50% correct responses and the highest grade that produces less than 50% correct responses, which is equivalent to a score of 5% correct. If a score of 59% can be obtained accurately at the language level, this is the speech recognition grade. 12.3 Improved descending method
12.3.1 If a check item in 12.1.3 is missed, give the first check item at any level and continue descending in steps of 5dB/step until two consecutive check items are heard incorrectly at the same language level (end). This level is called the starting language level (L). 12.3.2 From the starting language level, increase in steps of 10dB. 12.3.3: Give two check items at the starting language level. If 2 dB steps are used, then decrease by 2dB/step. See 12.3.6.
12.3.4 If at least one of the first six check items is recited correctly, continue this process. If this point cannot be reached, increase the starting level by 4~-10 dB.
12.3.5 If five of the last six check items are heard incorrectly, stop the descending series. 12.3.6 If 5 dB is used, 5 test items are given in each level. If all the test items in a level are recognized incorrectly, the test is terminated:
12.3.7 Calculate the decibel value of the language recognition level (1) according to formula (1): Tl-dxE+%
Where: - Starting language level (see 12.3.2): - The total number of correct responses in the entire process from the starting language level in 12.3.3 to the termination of the test in 12.3.5 or 12.3.6; - The number of correct responses in each level. dB:
- The number of tests in each level:
13 Determination of language recognition rate
If the language recognition rate is to be measured at multiple language levels (for example, to measure the language recognition threshold), the required test items should not be repeated in the same test.
GR/T 17696--1999
If the speech recognition threshold has not been checked before, the test items should be presented to the subject at a sufficiently high speech level that the subject can hear clearly before measuring the recognition rate to familiarize the subject with the test method. Usually, a speech level is selected that is 30-40 dB above the subject's average pure tone hearing threshold level at 500Hz, 7000Hz and 2000Hz.
Adjust the audiometer output to the required speech level and present a complete test form at each language level. In general, using only a portion of the test items in the test form will reduce the reliability and validity of the test. The choice of test level is determined by the following specific test conditions: a) Determine the maximum speech recognition rate, usually the percentage of the first score measured at a speech level 25-30dB above the speech recognition threshold level. Increase the speech level in 5dB or 10dB steps until the maximum recognition rate is measured or the subject feels uncomfortable. Until you feel comfortable or tired. If the percentage decreases at a higher language level (fallback effect), continue at a lower spectrum level: b) Determine the language recognition rate at the best loudness level. The method for selecting the inspection level is to continuously send the type of language signal used in the actual inspection, and let the test subject say what he hears (i.e.: too low, just right, too high, too loud). The midpoint between "too low" and "too high" is usually the preferred best level.
e) Determine the semi-optimal language level. First measure the maximum language recognition rate, and then increase the level every 5dB or 10dB. -- Reduce the speech sound level by one level, and send a whole test sheet every level until a percentage slightly higher than half of the maximum recognition rate and a percentage slightly lower than half of the maximum recognition rate are obtained. Use linear interpolation to obtain the integer value of the half-optimal speech level between the speech levels that obtain the above two percentages, and record this result (%) and the speech level that obtains this result. 14 Masking
In monaural speech audiometry, in order to prevent the non-test ear from hearing the speech signal, air conduction masking noise can be added to the non-test ear. Chapter 18 of GB/T7341.2--1998 stipulates masking noise. If the speech level expressed by the speech audiometer is more than 40dB higher than the lowest average bone conduction hearing threshold level of the opposite side at two frequencies of 510Hz, 1000Hz and 2000Hz, masking should be added in the air conduction examination, and in the monaural bone anomaly examination, masking should also be added. The minimum masking level L for air conduction examination (expressed in effective masking level) is! Lm = L - 40 + (MAm - MB.)
Where: --- Speech level expressed in speech level: h+r++. 2
MA---The average pure tone air conduction auditory level at the two frequencies with the lowest air conduction auditory level at 500Hz, 1000Hz and 2000Hz to be masked, expressed in hearing level; MB.- The average pure tone bone conduction hearing threshold level of the ear to be masked at the two frequencies where the bone conduction hearing threshold level is the lowest, 20 Hz, 0 Hz and 20 Hz, expressed in hearing level. Notes
1 In some cases, the masking level must be increased to achieve adequate masking for the test. 2 In the bone conduction test, the minimum knock level is 40 dB higher than L. 15 Speech recognition with competing sounds
15.1 Types of competing sounds
Frequency-weighted non-modulated random noise in accordance with GB/T7341.2--1998+13.1 may be used as competing sounds: if amplitude-modulated noise is used, its characteristics should be noted. Note: Recordings of human speech may also be used as competing sounds, such as the cacophony of multiple people (cocktail party noise) or the continuous speech of a single speaker. When this type of noise is used, the difference in the test results is smaller, while the difference in the results obtained with weighted random noise is larger. 15.2 Transmission of competing sounds
During headphone testing, the competing sound shall be transmitted through the same headphones used to transmit the speech test signal. In sound field audiometry, it is recommended that the speech transmission be located in front of the listener with an angle of incidence of . If two speakers with two-way tubes are used to transmit the competing sound, it is recommended that the angle of incidence be 45°. If the competing sound is random noise, it shall be transmitted through two or more side-channel speakers. The noise emitted from the speakers shall be non-uniform. GB/T17696-1999
If other angles of incidence (for example, above the subject's head) are used in specific tests, they should be clearly stated. 15.3 Speech and competing sound levels
shall be as specified in 3.93.20: Speech and competing sound levels shall be measured using the same frequency and time weighting. It is recommended to use 65d (this angle is . (roughly equivalent to the level of speech in ordinary conversation). If other languages are used, this should be clearly stated. The competition level can be fixed or variable. If fixed, a noise level of 60 dB is recommended. If variable, the noise level can be changed in 5 dB (or less) steps. 15.4 Examination procedures
The examination can measure the language recognition rate of one or more fixed speaker-to-voice ratios or the speech-to-noise ratio at the speech recognition threshold: 15.4.1 Speaker recognition at a fixed speech-to-noise ratio: 15.4.1.1 Adjust the audiometer to the required language level. 15.4.1.2 Use a low competition level (i.e. 20 dB below the language level) and give some audible test items to familiarize the subject with the test. 35.4.1.3 Adjust to the required competition level for the test and send a complete test form. Calculate the percentage of the score. 15.4.2 Speech recognition with competition
15. 4.2.1 Overview
In the following steps, increasing the competing sound level can be replaced by reducing the speaker's signal level, and vice versa. The results of the examination may be affected by this alternation.
Another procedure for determining the speech recognition threshold at a known speech or competing sound level is also described. They may produce comparable results. However, there is no experimental evidence to support this. The steps in 15.4.2.1.1 to 75.4.2.1.2 are the same: while in 15.4.2.2+ the speech-to-throat ratio is reduced in 5 dB steps, and 15.1.2.3 is followed by another alternative method, 15.4.2.1.1 is repeated according to 15.4.1.1 and 15.4. 1.2. 15.4.2.1.2 Reduce the speech noise by 5 steps (or 53 steps less), and send at least 2 test items at each step until the subject hears one of the test items incorrectly.
15.4.2.2 Reduce the speech noise ratio by 5 steps. 15.4.2.2.1 In this group of test items, count the number of test items that are correctly identified, and at least 10 test items are correct. 15.4.2.2.2 If the subject hears one of the test items incorrectly, If the score of a check item is at least 50%, the competition sound is increased by 5cB (or less than 5B) and a new set of check items is sent at each speech noise ratio until the score of a set of check items is less than -50%. Between the lowest level greater than 0% and the highest level less than 50%, the speech noise ratio of the speech recognition reading is determined by linear interpolation. If the speech noise ratio score of a station is exactly 50, then this is the speech recognition threshold level for that signal ratio. 15.4.2.3 Reduce the side noise. This is the next step. 15.4.2.3.1 When one check item is wrong, give the second one at the same competitive sound level. Continue to increase the level by -10dB. Then, give the next two check items wrong. This level is called the "starting level" of the competitive sound. Is.s15.4.2.3.2 From the starting level of the competitive sound, increase it by 10dB. 15.4.2.3.3 Give two check items at this starting level. Then increase it by 2dB (if 2 dB is used). If 5 dB is used, see 15, 1, 2, 3, 6.
15.4.2.3.4 If 5 out of the first 6 check items are correct, continue this step. If this indicator cannot be achieved, then start from the "starting level" and screen 4-10 dB.
15.4.2.3.5 Reduce the speech-to-noise ratio in a series (add competing sounds) until 5 out of the last 6 check items are misheard.
15.4.2.3.6 If the 5rB level is used, 5 check items should be sent at each level, and the check should be terminated when all the check items are misheard at a certain level. 15.4.2.3.7 Calculate the sound pressure level at which 50% of the check items can be correctly recognized in the selected speech using formula (3): 1. 1as + d × + sign
Where: 1. Competing sound level for 0% correct speech recognition rate 3Ln.s - - Non-initial level of competing sound, dB, (see 15.4. 2. 3. 2) * (3)Until the subject mishears one of the test items.
15.4.2.2 Method of reducing the speech-to-noise ratio by 5cB steps. 15.4.2.2.1 In this set of test items, count the number of test items that are correctly recognized, with at least 10 test items for each test item. 15.4.2.2.2 If the subject scores at least 50% on a set of test items, increase the competition sound by 5cB steps (or less) and send a new set of test items at each set of speech-to-noise ratio until the score on a set of test items decreases by 50%. Between the lowest level that yields a correct score greater than 0% and the highest level that yields a correct score less than 50%, use linear interpolation to determine the speech-to-noise ratio for speech recognition. If the score for the speech-to-noise ratio is exactly 50, then this is the speech recognition threshold level for that signal-to-noise ratio. 15.4.2.3 Reduce the side noise. This is the next step. 15.4.2.3.1 When one check item is wrong, give the second one at the same competitive sound level. Continue to increase the level by -10dB. Then, give the next two check items wrong. This level is called the "starting level" of the competitive sound. Is.s15.4.2.3.2 From the starting level of the competitive sound, increase it by 10dB. 15.4.2.3.3 Give two check items at this starting level. Then increase it by 2dB (if 2 dB is used). If 5 dB is used, see 15, 1, 2, 3, 6.
15.4.2.3.4 If 5 out of the first 6 check items are correct, continue this step. If this indicator cannot be achieved, then start from the "starting level" and screen 4-10 dB.
15.4.2.3.5 Reduce the speech-to-noise ratio in a series (add competing sounds) until 5 out of the last 6 check items are misheard.
15.4.2.3.6 If the 5rB level is used, 5 check items should be sent at each level, and the check should be terminated when all the check items are misheard at a certain level. 15.4.2.3.7 Calculate the sound pressure level at which 50% of the check items can be correctly recognized in the selected speech using formula (3): 1. 1as + d × + sign
Where: 1. Competing sound level for 0% correct speech recognition rate 3Ln.s - - Non-initial level of competing sound, dB, (see 15.4. 2. 3. 2) * (3)Until the subject mishears one of the test items.
15.4.2.2 Method of reducing the speech-to-noise ratio by 5cB steps. 15.4.2.2.1 In this set of test items, count the number of test items that are correctly recognized, with at least 10 test items for each test item. 15.4.2.2.2 If the subject scores at least 50% on a set of test items, increase the competition sound by 5cB steps (or less) and send a new set of test items at each set of speech-to-noise ratio until the score on a set of test items decreases by 50%. Between the lowest level that yields a correct score greater than 0% and the highest level that yields a correct score less than 50%, use linear interpolation to determine the speech-to-noise ratio for speech recognition. If the score for the speech-to-noise ratio is exactly 50, then this is the speech recognition threshold level for that signal-to-noise ratio. 15.4.2.3 Reduce the side noise. This is the next step. 15.4.2.3.1 When one check item is wrong, give the second one at the same competitive sound level. Continue to increase the level by -10dB. Then, give the next two check items wrong. This level is called the "starting level" of the competitive sound. Is.s15.4.2.3.2 From the starting level of the competitive sound, increase it by 10dB. 15.4.2.3.3 Give two check items at this starting level. Then increase it by 2dB (if 2 dB is used). If 5 dB is used, see 15, 1, 2, 3, 6.
15.4.2.3.4 If 5 out of the first 6 check items are correct, continue this step. If this indicator cannot be achieved, then start from the "starting level" and screen 4-10 dB.
15.4.2.3.5 Reduce the speech-to-noise ratio in a series (add competing sounds) until 5 out of the last 6 check items are misheard.
15.4.2.3.6 If the 5rB level is used, 5 check items should be sent at each level, and the check should be terminated when all the check items are misheard at a certain level. 15.4.2.3.7 Calculate the sound pressure level at which 50% of the check items can be correctly recognized in the selected speech using formula (3): 1. 1as + d × + sign
Where: 1. Competing sound level for 0% correct speech recognition rate 3Ln.s - - Non-initial level of competing sound, dB, (see 15.4. 2. 3. 2) * (3)
Tip: This standard content only shows part of the intercepted content of the complete standard. If you need the complete standard, please go to the top to download the complete standard document for free.