Speech Transmission Index Research Articles

Simplified expressions of uncertainty for single number quantities in ISO 3382-3 (2022).

The ISO 3382-3 (2022) standard defines single number quantities (SNQs) characterizing the spatial decay of speech (D2S, LpAS4m, rC) and of its intelligibility (rD). The standard assesses the accuracy of the measurement using a unique uncertainty value for each SNQ, relying on a round-robin test performed in a single office. To make this assessment more accurate, analytical expressions have been established by applying the law of propagation of uncertainty. These expressions, although precise, are too complex to be included in the standard. This work consists of their simplification, relying on 95 measurements conducted in 36 companies in France and Germany. First, the terms associated with distance measurement uncertainties were found to be negligible. Second, the estimation of the uncertainties of A-weighted speech levels and Speech Transmission Index, based on the measurements made at the workstations constituting the 95 paths, enabled further simplifications. Finally, simplified expressions turn out to be as accurate as the original expressions, and the estimated uncertainties are comparable to those reported by Yadav et al. (2019). If a unique value of uncertainty is to be retained for each SNQ, this study suggests 0.6 dB for D2S and LpAS4m, 7% for rC, and 10% for rD.

Chinese speech intelligibility and speech intelligibility index for the elderly

The speech intelligibility index (SII) and speech transmission index (STI) are widely accepted objective metrics for assessing speech intelligibility. In previous work, the relationship between STI and Chinese speech intelligibility (CSI) scores was studied. In this paper, the relationship between SII and CSI scores in rooms for the elderly aged 60–69 and over 70 is investigated by using auralization method under different background noise levels (40dBA and 55dBA) and different reverberation times. The results show that SII has good correlation with CSI score of the elderly. To get the same CSI score as the young adults, the elderly need a larger SII value, and the value increases with the increase of the age for the elderly. Since hearing loss of the elderly is considered in the calculation of SII, the difference in the required SII between the elderly and young is less than that of the required STI under the same CSI score condition. This indicates that SII is a more consistent evaluation criterion for different ages.

A context-based model to predict the intelligibility of sentences in non-stationary noises.

The context-based Extended Speech Transmission Index (cESTI) (van Schoonhoven et al., 2022, J. Acoust. Soc. Am. 151, 1404-1415) was successfully applied to predict the intelligibility of monosyllabic words with different degrees of context in interrupted noise. The current study aimed to use the same model for the prediction of sentence intelligibility in different types of non-stationary noise. The necessary context factors and transfer functions were based on values found in existing literature. The cESTI performed similar to or better than the original ESTI when noise had speech-like characteristics. We hypothesize that the remaining inaccuracies in model predictions can be attributed to the limits of the modelling approach with regard to mechanisms, such as modulation masking and informational masking.

Experimental comparisons of speech transmission index prediction methods

Good speech intelligibility is the primary goal of acoustic environment control in lecture rooms and auditoria. The speech transmission index (STI) is the primary objective evaluation metric for speech intelligibility. Therefore, accurate prediction of the STI during the design stage is important for acoustic environment control in lecture rooms and auditoria. Two types of STI prediction methods based on simulated impulse responses and statistical are recommended by the IEC standard. In this study, the prediction accuracy and influencing factors of these two types of prediction methods were systematically analysed using STI measurements collected at 25 receiver positions in six rooms. The results reveal the following. (1) When the influence of the signal-to-noise ratio (SNR) is not considered, for the STI prediction method based on simulated impulse responses, the prediction accuracy is in an acceptable range with the average predicted STI discrepancy of the 25 receiver positions being −0.004 and the maximum discrepancy being 0.046. For the STI prediction method based on statistical, the average predicted STI discrepancy of the 25 receiver positions was −0.015, and the maximum discrepancy was −0.091. STI prediction based on statistical exhibited a larger prediction error at the receiver positions close to the sound source and smaller prediction error at the receiver positions farther away from the sound source. (2) When the influence of the SNR was considered, the errors of both prediction methods increased significantly. For the STI prediction method based on simulated impulse responses, the average predicted STI discrepancy of the 25 receiver positions was −0.035 and the maximum discrepancy was −0.088. For the STI prediction method based on statistical, the average predicted STI discrepancy of the 25 receiver positions was −0.030 and the maximum discrepancy was −0.112. (3) Using a sound source that is different from the onsite measurement sound source can induce noticeable errors in both methods. (4) For the STI prediction method based on statistical using the actual Q value of the sound source, the prediction accuracy of receiver positions close to the sound source can be significantly improved and the effect on receiver positions farther away from the sound source is relatively small. (5) Both methods require significant technical expertise from users, rendering it difficult to obtain accurate prediction results during the design stage.

Classroom architectural design evolution: Acoustic evaluation of public-school classrooms in Malaysia

Education transformation greatly emphasises curriculum modification to produce impactful future generations and yet often disregards the impact of classroom design in achieving desired education outcomes. The prioritisation of optimal acoustic quality in classroom design is crucial due to the inherent reliance on auditory abilities in the process of teaching and learning. Therefore, a comprehensive acoustic standard guideline for classrooms, ANSI Standard 12.60, was launched in 2002 and adopted for school classrooms in the USA. However, the scenario might be different in Malaysia, as there is no acoustic standard guideline established. Therefore, this study seeks to identify the actual acoustic conditions of classrooms that were constructed in the post-independence era. On-site acoustic measurements were performed to evaluate significant acoustic parameters, including reverberation time (RT), background noise level (BNL), speech transmission index (STI), and sound pressure level (SPL). The findings revealed that the RT for both classrooms was within the recommended value, while the BNL and STI of both classrooms failed to comply with the established recommended guidelines. In a similar vein, these findings translate the degree of awareness among the education institutions and construction sector of the importance of classroom acoustics in providing a better learning experience among students.

Relationship between objective measures and listener intelligibility of speech processed by source-separation algorithms

A major determinant of success for separating a speech signal from a noisy environment is the intelligibility of the extracted speech signal. Intelligibility is best measured as the fraction of words correctly recognized by listeners, but computational measures are often preferred because they are less labor-intensive than collecting listener judgements. We compare listener intelligibility data to three acoustically derived measures: (a) SNRs estimated from the processed mixtures, (b) coherence and (c) speech-based Speech Transmission Index (sSTI). Sentences were recorded against restaurant babble, white Gaussian noise, and nonstationary noise by four microphones at different SNRs ranging from +4 dB to −8 dB. Processing conditions included (1) the original mixture; (2) the mixture processed by a critically determined 4-channel blind source separation (BSS) algorithm; (3) the mixture processed by a 2-channel, underdetermined, BSS algorithm; (4) the residual after subtracting noise estimates determined using a least-mean squared (LMS) algorithm; (5) an estimate of speech extracted using an LMS algorithm to remove the two noises and then 2-channel BSS to separate the sentences from the babble; and (6) pristine speech recorded with no noise. The computational measures are compared to gold-standard intelligibility results from listening tests.

An audibility model of the headband trial with a bone conduction device in single-sided deaf subjects

Objective Modelling the head-shadow effect compensation and speech recognition outcomes, we aimed to study the benefits of a bone conduction device (BCD) during the headband trial for single-sided deafened (SSD) subjects. Design This study is based on a database of individual patient measurements, fitting parameters, and acoustic BCD properties retrospectively measured on a skull simulator or from existing literature. The sensation levels of the Bone-Conduction and Air-Conduction sound paths were compared, modelling three spatial conditions with speech in quiet. We calculated the phoneme score using the Speech Intelligibility Index for the three conditions in quiet and seven in noise. Study sample Eighty-five SSD adults fitted with BCD during headband trial. Results According to our model, most subjects did not achieve a full head-shadow effect compensation with the signal at the BCD side and in front. The modelled speech recognition in the quiet conditions did not improve with the BCD on the headband. In noise, we found a slight improvement in some specific conditions and minimal worsening in others. Conclusions Based on an audibility model, this study challenges the fundamentals of a BCD headband trial in SSD subjects. Patients should be counselled regarding the potential outcome and alternative approaches.

Syllable-rate-adjusted-modulation (SRAM) predicts clear and conversational speech intelligibility.

Objectively predicting speech intelligibility is important in both telecommunication and human-machine interaction systems. The classic method relies on signal-to-noise ratios (SNR) to successfully predict speech intelligibility. One exception is clear speech, in which a talker intentionally articulates as if speaking to someone who has hearing loss or is from a different language background. As a result, at the same SNR, clear speech produces higher intelligibility than conversational speech. Despite numerous efforts, no objective metric can successfully predict the clear speech benefit at the sentence level. We proposed a Syllable-Rate-Adjusted-Modulation (SRAM) index to predict the intelligibility of clear and conversational speech. The SRAM used as short as 1 s speech and estimated its modulation power above the syllable rate. We compared SRAM with three reference metrics: envelope-regression-based speech transmission index (ER-STI), hearing-aid speech perception index version 2 (HASPI-v2) and short-time objective intelligibility (STOI), and five automatic speech recognition systems: Amazon Transcribe, Microsoft Azure Speech-To-Text, Google Speech-To-Text, wav2vec2 and Whisper. SRAM outperformed the three reference metrics (ER-STI, HASPI-v2 and STOI) and the five automatic speech recognition systems. Additionally, we demonstrated the important role of syllable rate in predicting speech intelligibility by comparing SRAM with the total modulation power (TMP) that was not adjusted by the syllable rate. SRAM can potentially help understand the characteristics of clear speech, screen speech materials with high intelligibility, and convert conversational speech into clear speech.

Acoustic assessment of a novel visor concept with aerodynamic sealing for medical care

This paper presents the experimental acoustic assessment and optimization of a novel solid visor with aerodynamic sealing for medical care. This Personal Protective Equipment (PPE) ensures an effective aerodynamic sealing of the breathing zone thus reducing the risk of inhaling droplets and aerosols carrying a potential infectious agent. The PPE relies on an air curtain system that may generate considerable noise unsuitable for prolonged usage, so it is essential to investigate solutions to reduce the noise level. The main investigated sound descriptors were the A-weighted Noise Equivalent Level, the Third-Octave Frequency Distribution, and the Speech Intelligibility Index (SII), considering different solutions for air flow rate, tube diameter, thickness and insulation, configurations of the tube splitter and plenum, and the incorporation of physical barriers and a sound muffler. Through several optimizations, noise equivalent levels were reduced by 20 to 30 dBA for air flow rates of 15 to 30 l/min. The final prototype generates a maximum noise level of 45.2 dBA. The SII was improved from poor to excellent, with values close to 1 for all air flow rates considered. The impact on the speech was investigated and the results were also compared to those obtained with a surgical mask and a FFP2 respirator. This study shows the potential of employing different strategies to enhance the acoustic performance of personalized air-curtain protective devices.

Determining cochlear implant candidacy in children with residual hearing: A scoping review

ObjectivesThe purpose of this review was to summarize what is known about criteria, assessments, and recommendations for evaluating cochlear implant (CI) candidacy in children with residual hearing. MethodsPeer-reviewed studies were identified through a systematic search in five electronic databases. Articles were screened and assessed for eligibility. From the eligible studies, data were extracted to summarize and present a narrative synthesis of the findings. ResultsA total of seven articles (two reviews and five primary studies) were included in the final analysis. Hearing levels better than the moderately severe to severe range (65–90 dB HL) tend to be supported as audiological candidacy criteria for pediatric CI. Recommendations for candidacy consideration based on audiologic thresholds range from 65 to 80 dB Hl pure-tone average as the lower boundary. Our review did not identify any specific assessment protocols. However, additional decision-making considerations related to borderline hearing loss configurations and assessment tools (the Speech Intelligibility Index and the Pediatric Minimum Speech Test Battery) were identified. Supplementary assessment considerations were also reported. ConclusionThere is limited information regarding specific assessment protocols for children with residual hearing. The literature is primarily focused on guidelines related to audiologic criteria, although it is widely recommended that other areas of functioning should also be considered. Most recommendations appear to be based on expert opinion, clinical expertise, and evidence from overall pediatric CI outcomes rather than empirical evidence targeting children with residual hearing. There is an ongoing need for research to further develop protocols and tools that can assist clinicians and families in making cochlear implantation decisions for children with residual hearing.

Peculiarities of a Rarely Used Method of Measuring the Speech Transmission Index in Premises

Evaluation of speech transmission index (STI) in premises allows for determining the speech intelligibility, and therefore the suitability of premises for speech communication. STI measurements using the speech transmission index for telecommunication systems (STITEL) method are rarely performed in rooms, possibly due to insufficient information on the accuracy of this method. In this paper, computer simulations were used to estimate the STI estimation errors by the STITEL method under conditions of noise and reverberation. The pink noise model and the room impulse response estimate of a real room with a reverberation time T60=0.8 s were used for the research. The duration of the test signals varied between 4, 8, 16, 32, and 64 seconds, and the signal-to-noise ratio varied from minus 28 dB to plus 28 dB. The dependences of the bias, standard deviation, and total error of the STI estimate on the duration of the test signal and the signal-to-noise ratio are obtained. It is shown that the total error of the STI estimation is close to 0.03 when the duration of the test signal is 8 s. Under conditions of noise action, this error decreases with a further increase in the duration of the test signal. Under conditions of joint action of noise and reverberation, such a decrease was not observed, while the total error is within 0.03-0.04.

Difference in SII provided by initial fit and NAL-NL2 and its relation to self-reported hearing aid outcomes

Objective The speech intelligibility index (SII) is used to quantify the audibility of the speech. This study examined the relationship between self-reported hearing aid (HA) outcomes and the difference in aided SII (SIIA) calculated from the initial fit (IF) gain and that prescribed as per the second generation of National Acoustic Laboratory Non-Linear (NAL-NL2). Design A prospective observational study. Study sample The study included 718 first-time and 253 experienced HA users. All users had a valid real-ear measurement (REM) at three input levels (55, 65 and 80 dB SPL). Results The gain provided by IF was lower than NAL-NL2 at 55 and 65 dB SPL. IF gain exhibited reduced compression than NAL-NL2 as input levels increased from 55 to 80 dB SPL. On average, the SIIA provided by IF was significantly lower than that for NAL-NL2 at all input levels. The difference in SIIA between IF and NAL-NL2 at 80 dB SPL input level with 0 dB signal-to-noise ratio (SNR) turned out to be a predictor for self-reported outcome for first-time HA users. Conclusion The study suggests that an SIIA close to that provided by NAL-NL2 at high input levels would be preferred to obtain a better self-reported outcome.

Influence of Test Room Acoustics on Non-Native Listeners’ Standardized Test Performance

Understanding the impact of room acoustics on non-native listeners is crucial, particularly in standardized English as a foreign language (EFL) proficiency testing environments. This study aims to elucidate how acoustics influence test scores, considering variables overlooked in prior research such as seat position and baseline language proficiency. In this experiment, 42 Japanese university students’ performance on standardized EFL listening tests was assessed in two rooms with distinct acoustic qualities, as determined by the speech transmission index (STI) and reverberation time (RT). The rooms differed significantly in their STI values and RT measurements, with one exhibiting high speech intelligibility qualities of ≥0.66 STI and RT0.5–2kHz < 0.7 s and the other falling below these benchmarks. The findings revealed that listening test scores were consistently higher in the acoustically favorable room across all participants. Notably, the negative effect of poor acoustics was more pronounced for students with lower baseline language proficiency. No significant score differences were observed between front- and rear-seat positions, suggesting that overall room acoustics may be more influential than individual seating locations. The study concludes that acoustics play a significant role in the standardized EFL test performance, particularly for lower-proficiency learners. This highlights the necessity of standardized testing environments to be more carefully selected in order to ensure the fair and reliable assessment of language proficiency.

Evaluating the effect of masks on the vocal health and speech intelligibility of healthcare workers in various hospital unit

The COVID-19 pandemic has enforced mandatory mask usage, particularly in hospitals, and its impact on the vocal problems and speech intelligibility of healthcare workers was examined in this study. A total of 76 healthcare workers from various units, including emergency, inpatient, outpatient, and intensive care, participated in an online questionnaire survey to assess vocal issues experienced while wearing masks. The participants' self-assessed vocal handicaps were quantified using the Vocal Handicap Index (VHI) score. The findings revealed that healthcare workers in emergency units faced the most severe vocal problems, followed by those in outpatient units. The most reported vocal symptoms were clearing the throat and throat dryness. Additionally, noise levels were measured continuously for four hours in each unit, and sources of noise were identified. The measured sound pressure levels (SPL) in emergency, inpatient, outpatient, and intensive care units were 68.64, 63, and 72 dBA, respectively. These values exceeded the World Health Organization's recommended limit of 30-40 dBA, indicating significant discomfort. Lastly, to delve deeper into the increased vocal effort required when wearing masks, the Speech Transmission Index (STI) parameter was analyzed using the recorded SPL as background noise and audio recordings of healthcare workers in different mask-wearing conditions. The study discovered that using a mask considerably reduced speech intelligibility compared to not wearing a mask. Keyword: healthcare workers, vocal problems, noise level, pandemic

Relation between sound absorption and speech intelligibility in room, part 1: Theoretical investigation

In Japan, requirements and design guidelines on sound absorption of a variety of ordinary spaces are currently under consideration. In general, sound absorption works for reverberation suppression and noise reduction, both of which contributes to improve speech intelligibility performance. However, its combined effect is not fully clarified, especially on the relationship among of reverberation time, average absorption coefficient and speech intelligibility indices. In this paper, two theoretical case studies are performed based on the formulations of the clarity C50 and the useful-to-detrimental ratio U50 in the conditions without and with background noise. For rectangular rooms with an aspect ratio of 1:2:4, those indices are calculated at three source-to-receiver distances, changing the room volume widely under different noise levels. The results demonstrated that the absorption coefficient required for a target C50 increases with increasing the room volume, but it decreases beyond a certain volume. Moreover, it was shown that under high background noise, U50 is not effectively improved by increasing the absorption coefficient if the communication distance is much longer than the critical distance.

Finite element analysis of sound fields in the corridors of the hospital ward: Investigation on installing patterns of sound absorbing materials

In this paper, using the time-domain finite element method (TDFEM), an effective installation pattern of sound absorbing materials in corridors of the hospital ward is investigated. First, several models with different installation patterns of sound absorbing materials are analyzed by TDFEM. Next, spatial distributions of the squared sound pressure, the sound pressure level and the speech intelligibility index are calculated in each model. As a result, the amount of decrease of the sound pressure level by installing sound absorbing materials on the ceiling and part of the wall in case that room average (α¯) equals 0.18 is larger than those by installing sound absorbing materials on the ceiling in case that α¯ = 0.20. Finally, several models with different installation patterns of sound absorbing materials, of which is the same, are analyzed by TDFEM, and compared among the amount of decrease of sound pressure levels in each condition.

Relation between sound absorption and speech intelligibility in room, part 2: Case study based on geometric simulation

Sound absorption works for reverberation suppression and noise reduction, both of which contribute to improving speech transmission performance. In the previous paper, its combined effect was theoretically investigated to clarify the relationship among of reverberation time, average absorption coefficient and speech intelligibility indices. In this paper, a case study is first conducted to examine the applicability of the theoretical estimation of clarity C50 based on geometric simulation. In rectangular rooms with different volumes, aspect ratios and absorption distributions, impulse responses are simulated for various source-to-receiver positions, and then the distributions of the simulated C50 are compared with the theoretical values. Furthermore, a subjective experiment on speech intelligibility is conducted by using a 6-channel sound field reproduction system, where a speech signal is convolved with the simulated impulse responses under background noise conditions. The results show the correlation between the listening difficulty rating and the estimated useful-to-detrimental ratio U50.

One, Two, Three! Can Everybody Hear Me? Acoustics of Roman Contiones. Case Studies of the Capitoline Hill and the Temple of Bellona in Rome

Abstract Rhetoric was one of the cornerstones of Roman education and public speaking, the essence of being a Roman politician. The speakers attempted to captivate the audience with their style and convince them of their arguments. Studying the audience is therefore just as important as investigating the speakers and their speeches. The aim of this article is to estimate the number of people who could intelligibly hear a speaker delivering a speech from two speaking platforms located in the city of Rome: the podium of the Temple of Bellona in the Campus Martius (in the Late Republican and Late Augustan periods) and the Capitoline Temple. To do this, we built virtual reconstructions of both venues according to the current state of knowledge about them, taking into account the geometry of the space as well as the materials from which they were built. On the models thus prepared, we carried out acoustic simulations for three different levels of background noise (36, 49, and 55 dBA), resulting in Speech Transmission Index maps. The results became the basis upon which we estimated the size of the maximum potential crowds that could hear speech intelligibly, using two methods based on the behaviour of contemporary crowds. We further compared our results with those of previous studies that concern other speaking platforms in Rome.

Relationship Between Chinese Speech Intelligibility of Elderly and Speech Transmission Index

In this paper, the relationship between Chinese speech intelligibility (CSI) scores of the elderly aged 60–69 and over 70 years old, and speech transmission index (STI) were investigated through the auralization method under different reverberation time and background noise levels (BNL, 40 dBA and 55 dBA). The results show that the CSI scores of the elderly are significantly worse than those of young adults. For the elderly over 70, the CSI scores become much lower than those of young adults. To be able to achieve the same CSI, the elderly, especially those over 70, need much higher STI and greater SNR than the young. The elderly aged 60–69 and over 70 need to improve their STI by 0.419 and 0.058 respectively under BNL 40 dBA, as well as 0.282 and 0.072 respectively under BNL 55 dBA, so as to obtain the same CSI scores as the young adults.

A comparison of automated verification using paediatric hearing aids

Objective Best-practice guidelines recommend the use of hearing aid verification in children; however, this is not always performed. Automated hearing aid verification has been reported to be more accurate and efficient than manual verification in adults, but it is not known if this transfers to the paediatric population. Design A within-group design compared manual and automated hearing aid verification on four measures; fitting accuracy, prescription targets, completion time, and the speech intelligibility index. Sample Twenty paediatric patient hearing aid profiles (M = 8.25 years) with unilateral or bilateral hearing aids. Results A Wilcoxon-signed rank test indicated manual verification achieved a significantly closer match to target at 0.5 kHz, by an average of 1 dB. There were no significant differences at any other frequency. Across 80 comparisons (four frequencies measured in 20 listeners), 82.5% of automated verifications were identical to, or within 1 dB of, manual verifications. A paired-samples t-test confirmed automated verification to be an average of 91.9 seconds faster than manual verification. Conclusion Automated verification was able to provide an accurate match to target within recommended tolerances for hearing aid fittings and was significantly quicker than manual verification. These data suggest that automated verification of hearing aids could play a role in paediatric audiological management.