|Year : 2018 | Volume
| Issue : 2 | Page : 88-90
Effectiveness of multitalker babble over speech noise and its implications: A comparative study
Archana Gundmi, P Himaja, Alisha Dhamani
Department of Speech and Hearing, School of Allied Health Science, Manipal Academy of Higher Education, Manipal, Karnataka, India
|Date of Web Publication||4-Sep-2018|
Dr. Archana Gundmi
Department of Speech and Hearing, School of Allied Health Science, Manipal Academy of Higher Education, Manipal, Karnataka
Source of Support: None, Conflict of Interest: None
Introduction: Speech perception is a complex process. Many variables can affect speech perception; among that, background noise is one of the important factors where maximum interference for the speech perception happens. Perception of speech varies also with types of noise the individual is encountering. Aim: The present study was concentrated on effect of multitalker babble over speech noise in perceiving speech signal. Method: Twenty-four normal-hearing individuals participated in the present study were tested in two different conditions in different signal-to-noise ratio ratios. Result and Conclusion: Results revealed multitalker babble is more effective than speech noise thus can be used more relevantly in clinical practice.
Keywords: Multitalker babble, speech noise, speech perception
|How to cite this article:|
Gundmi A, Himaja P, Dhamani A. Effectiveness of multitalker babble over speech noise and its implications: A comparative study. Indian J Otol 2018;24:88-90
|How to cite this URL:|
Gundmi A, Himaja P, Dhamani A. Effectiveness of multitalker babble over speech noise and its implications: A comparative study. Indian J Otol [serial online] 2018 [cited 2019 Dec 5];24:88-90. Available from: http://www.indianjotol.org/text.asp?2018/24/2/88/240570
| Introduction|| |
Speech perception is “the process by which a perceiver internally generates linguistic structures believed to correspond with those generated by a talker.” Thus, for speech perception to take place, structurally and functionally, an intact auditory system is needed. Speech perception is a complex process that is affected by many variables. Reverberation time, distance from the speaker to the listener, auditory capabilities of the listener, cognitive abilities, speaking rate, dialect, and knowledge on the topic, level of the background noise, and level of the speech signal relative to the level of the background noise are the important factors which can speech perception.
Literature has thrown light that among these variables, signal-to-noise ratio (SNR) has a major impact in degrading the speech signals in both individual with normal hearing and also in individual with hearing impairment. When noise is present in a listening environment, it would mask the speech signal by concealing the less-intense portions of the speech signal, and therefore, consonantal phonemes are typically masked more than vowel phonemes because they contain less spectral energy. The result is a fall in the perception of speech under these conditions. Performance on speech perception tasks is best when the SNR of a listening environment is favorable (+10 dB). Performance on such tasks tends to decrease as the SNR becomes less favorable; similarly in clinical practice, we encounter that an overwhelming majority of patients who go to a health-care professional complain of difficulty in understanding speech in the presence of background noise. Hence, it becomes a challenging task for the audiologist during hearing aid selection decision, programming and fine tuning the hearing aid to satisfy individual needs.
Need of the study
In everyday environment, we encounter lot of noises and we are never under idyllic situations. Problems with speech in noise perception are associated with a great disruption of temporal encoding in noisy situation as noise causes excessive neural delay in person who has difficulty in hearing situation. Precise neural encoding of timing elements and robust pitch representation are essential for hearing in noise. Noise disrupts neural encoding to a greater degree in those with poor speech in noise perception. Locking onto stimulus regularities enables the listener to follow a specific voice in background noise. Hence, testing speech perception in the presence of noise should be included as a part of the routine audiological tests. One such test is speech perception in noise (SPIN) test that provides speech signals embedded in competing noises. It is also used in estimating SNR loss in hearing impaired individuals. The noise used in SPIN test is speech noise which is a broad band noise with a narrower frequency range extending from at least 250 to 4000 Hz. The use of speech noise in SPIN test has the limitation that it is less representative of everyday listening speech in the presence of background noise. Thus, using speech noise would not give us accurate results concerning the SNR loss. Hence, using effective noise to represent a realistic simulation of a social gathering, shopping mall, and restaurants is essential. Therefore, a need arises to document the need of using an effective noise such as multitalker babble which suits the environment which is more natural rather than using continuous speech noise.
Aim of the study
Effectiveness of multitalker babble over speech noise and its implications: A comparative study.
| Methods|| |
Twenty-four individuals participated in the study. Detailed case history was taken for each individual, and they reported of no significant medical history and no complaints of speech understanding. Modified Hughson-Westlake procedure was used to determine the pure-tone average (PTA) using GSI 61 clinical audiometer. Speech identification scores were obtained by conducting speech audiometry using clinical diagnostic audiometer under TDH 50 headphones for each ear independently. Twenty phonetically balanced words developed by Maya Devi (1982) were presented monaurally at 40 dB SL (above speech recognition threshold) or at most comfortable level, and number of words correctly identified was noted. Final scores will be calculated as percent. The participants had normal hearing (PTA <15) along with good speech discrimination scores and normal middle ear function were taken for the study.
The present study was carried out in two steps wherein multitalker babble stimulus was created in first step with 6 Kannada speakers who were asked to read random Kannada passages simultaneously and was recorded using CSL software (Computerized Speech Lab (CSL), Model 4500, PENTAX Medical acoustic products USA). The recorded material was then calibrated. Thereafter, in second step, SPIN test was carried out in 3 SNR conditions (1) 10 dB SNR, (2) 0 dB SNR, and (3) −10 dB SNR using two types of noise, i.e., speech noise and multitalker babble. In both the type of noises, standard 20 Kannada bisyllabic words were used as a target stimulus in random order. The bisyllabic words and the two types of noises were presented ipsilaterally. Scoring of 5% was given for each correct response. The obtained results were analyzed using statistical softwareStatistical Package for the Social Sciences (SPSS) v. 16, (IBM, Armonk, New York).
| Results|| |
The obtained results in [Table 1] are the average percentage scores for 24 participants obtained under the two noise conditions at 3 SNR conditions.
|Table 1: Speech identification scores for speech noise and multitalker babble|
Click here to view
They reveal that as SNR decreases, speech perception scores also decrease. At +10 dB SNR, participants were able to respond well at 100% whereas, at 0 dB SNR, the participants had some difficulty revealing a score of 85%–90%. However, at −10 dB SNR condition, speech perception scores were decreased significantly in both types of noises revealing a score of 15%–25%. Furthermore, it can also be seen that in −10 dB SNR, speech scores were decreased more evidently in multitalker babble compared to the scores obtained in speech noise revealing a score of 20% for multitalker babble and 28.5% in speech noise. Independent t-test was used, and it revealed that the difference obtained between speech noise and multitalker babble was found to be statistically significant (P = 0.01). Whereas no ear-specific and gender-specific responses were observed.
| Discussion|| |
It is an established fact that as SNR decreases, speech perception scores also decrease and that the problems with neural encoding of temporal and spectral information are associated with poor SPIN. However, from this study, it is interesting to note that as SNR decreases, speech perception scores in the presence of multitalker babble are poorer than the results obtained using speech noise. This can be attributed to multitalker babble being a more realistic simulation of a social gathering in which listener may tune out the target speaker and tune in one or more of the nearby talkers which is termed as “selective listening.” Furthermore, the use of speech noise eliminates temporary gaps that would put the auditory system into adaptation leading to better speech perception scores in the presence of continuous speech noise.
This study focuses on normal population; yet poor speech scores were obtained in these individuals as SNR was decreased in the presence of multitalker babble, and hence, it is logically attributed that individuals with hearing impairment would show more degraded scores with multitalker babble as a competing stimuli.
| Conclusion|| |
Multitalker babble is an effective noise compared to speech noise in predicting SNR loss while diagnosing a case as it simulates a realistic situation. Thus, it plays an effective role in assessment of actual hearing loss. Knowing the SNR loss makes it possible for the audiologist to recommend appropriate technology (e.g., omnidirectional microphones, directional microphones, array microphones, and FM systems) and during hearing aid selection that is required for the listener to understand speech in noisy situations. It also enables the audiologist to give patients realistic expectations for their potential improvement in noise with a given amplification technology. Estimating SNR loss would increase clinical efficiencies. They can be used for counseling patients suitably and guiding their expectations, which often reduces unnecessary visits for hearing aid readjustments. Furthermore, it can also be implemented in various steps of auditory training to rehabilitate individuals with hearing impairment across all age groups.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Boothroyd A. Assessment of speech perception capacity in profoundly deaf children. Am J Otol 1991;12 Suppl:67-72.
Humes LE, Christopherson L. Speech identification difficulties of hearing-impaired elderly persons: The contributions of auditory processing deficits. J Speech Hear Res 1991;34:686-93.
Humes LE, Roberts L. Speech-recognition difficulties of the hearing-impaired elderly: The contributions of audibility. J Speech Hear Res 1990;33:726-35.
Crandell CC, Smaldino JJ. Classroom acoustics for children with normal hearing and with hearing impairment. Lang Speech Hear Serv Sch 2000;31:362-70.
Bacon SP, Opie JM, Montoya DY. The effects of hearing loss and noise masking on the masking release for speech in temporally complex backgrounds. J Speech Lang Hear Res 1998;41:549-63.