Speech Transmission Index Measurements Research Articles

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.

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.

Multicomponent Signal for Comparing Direct and Indirect Methods of Speech Transmission Index Measurement

When evaluating the intelligibility of speech distorted by noise and reverberation, direct or indirect methods of measuring the speech transmission index are used. However, it remains insufficiently studied how significantly differ the results of measurements obtained by direct and indirect methods. To find an answer to this question, the use of a multicomponent test signal consisting of four "elementary" signals separated by pauses is proposed in this paper. As "elementary" signals, it is proposed to use a maximum-length sequence, a speech shaped maximum-length sequence, a speech shaped stationary noise, and a speech shaped amplitude-modulated noise. Use of amplitude-modulated noise allows estimating speech transmission index by a direct method. Other "elementary" signals make it possible to estimate speech transmission index by two variants of indirect method. The proposed algorithms and corresponding computer programs were tested on trial signal models, while the consistency of the obtained results with the results of previous studies was revealed. The results of the signal models studies show that both considered variants of the indirect speech transmission index measurement method lead to underestimated results compared to the direct method. For one of the variants of the indirect method, the value of the estimate bias is 0.03–0.04, regardless of the interfering conditions. For another variant of the indirect method, the estimate bias varies from 0.01 to 0.18, depending on the interference conditions.

Rapid quantitative measure of speech intelligibility through personal protective equipment facemasks using two head & torso simulators

A test method has been developed that provides a rapid and quantitative measure of speech intelligibility via speech transmission index (STI) through personal protective facemasks using two Brüel & Kjær Head and Torso Simulators (HATS) under a variety of acoustical environmental conditions. STI measurements were made using this new Test Method (TM) on a collection of 3M prototype and production facemasks and respirators, a competitor facemask, and a “no mask” baseline. This new TM enables rapid differentiation among facemask designs. Comparisons among the set of face coverings shows a clear tradeoff of filtration level for STI. New 3M prototypes with 90% filtration achieved an Excellent STI of 0.85, only 0.03 below a perfect no mask baseline. There is a dramatic step from maximum STI with no filtration at all to Excellent STI with 90% filtration. Most of the 95% filtration masks had STIs clustered near the lower end of the Excellent band at 0.78. P100 filtration significantly compromised speech intelligibility in the 3M 6300 half facepiece reusable respirator with a Fair quality rating on the STI scale at 0.60.

Experimental comparison of speech transmission index measurement in natural sound rooms and auditoria

Speech intelligibility is an important factor for natural sound quality evaluation in rooms and auditoria. As an objective metric for predicting speech intelligibility, the application of the speech transmission index (STI) has increased significantly in recent years. In this study, three measurement systems were selected to measure the STI and speech transmission index for public address systems (STIPA) of 44 listening conditions at 11 receiver positions in 4 rooms. Thereafter, to explore the substantial differences among the three systems, three new measurement systems were developed, and the STI and STIPA under the same 44 listening conditions were measured again. The results reveal the following: (1) An obvious difference existed among the measurement results of the three selected systems, mainly caused by the influence of the sound source directivity, as well as the possible influences of the test signal sound pressure level calibration and the calculation procedure. The directivity of the microphone as well as the presence of a human body near the microphone influence the STI measurement results, although these affects are relatively small. (2) A small difference existed between the STIPA and STI, mainly owing to the room frequency response characteristics. (3) There was no essential difference between the measurement results of the indirect and direct methods; STI can be considered to expand the application scope of the indirect method. The findings in this study are significant for the application of the STI in the field of architectural acoustics.

Acoustic comfort for hypermarket cashiers: Problems and solutions

An acoustic investigation carried out in a large hypermarket pointed out that several acoustic “zones” could be identified. Among them, the most critical was the checkout area, where cashiers were exposed to many noise sources like customers passing with shopping trolleys, “beeps” due to barcode readers, people voices, packaging and paging messages, and, above all, background music. In fact, as the checkout position usually divides the hypermarket from the shopping arcade, workers in this area are further exposed to music (and noise) from the nearby cafes, restaurants, and stores. As a result background noise LA90 is about 66 dB, with peaks (LA10) of about 72 dB. Subjective analysis also reported the highest rate of complaints in this area. Starting from a detailed analysis of the different noise sources and of the room acoustic conditions of the area (including reverberation time and speech transmission index measurements), a set of mitigation actions is analyzed and discussed.

Towards measuring the Speech Transmission Index in fluctuating noise: Accuracy and limitations.

In the field of room acoustics, the modulation transfer function (MTF) can be used to predict speech intelligibility in stationary noise and reverberation and can be expressed in one single value: the Speech Transmission Index (STI). One drawback of the classical STI measurement method is that it is not validated for fluctuating background noise. As opposed to the classical measurement method, the MTF due to reverberation can also be calculated using an impulse response measurement. This indirect method presents an opportunity for STI measurements in fluctuating noise, and a first prerequisite is a reliable impulse response measurement. The conditions under which the impulse response can be measured with sufficient precision were investigated in the current study. Impulse response measurements were conducted using a sweep stimulus. Two experiments are discussed with variable absorption, different levels of stationary and fluctuating background noise, and different sweep levels. Additionally, simulations with different types of fluctuating noise were conducted in an attempt to extrapolate the experimental findings to other acoustical conditions. The experiments and simulations showed that a minimum impulse-to-noise ratio of +25 dB in fluctuating noise was needed.

Negative Effect of Acoustic Panels on Listening Effort in a Classroom Environment.

Acoustic panels are used to lessen the pervasive effects of noise and reverberation on speech understanding in a classroom environment. These panels, however, predominately absorb high-frequency energy important to speech understanding. Therefore, a classroom environment treated with acoustic panels might negatively influence the transmission of the target signal, resulting in an increase in listening effort exerted by the listener. Acoustic panels were installed in a public school environment that did not meet the ANSI-recommended guidelines for classroom design. We assessed the modifications to the acoustic climate by quantifying the effect of (1) acoustic panel (i.e., without, with) on the transmission of a standardized target signal at different seat positions (i.e., A-D) using the Speech Transmission Index (STI) and (2) acoustic panel and seat position on listening-effort performance in a group of third-grade students having normal-hearing sensitivity using a dual-task paradigm. STI measurements are described qualitatively. We used a repeated-measures randomized design to assess listening-effort performance of monosyllabic words in a primary task and digit recall in a secondary task for the independent variables of acoustic panel and seat position. Twenty-seven, third-grade students (12 males, 15 females), ranging in age from 8.3 to 9.4 yr (mean = 8.7 yr, standard deviation = 0.7), participated in this study. Qualitatively, we performed STI measurements under both testing conditions (i.e., panel and seat location). For the primary task of the dual-task paradigm, participants heard a ten-item list of monosyllabic words (i.e., ten words per list) recorded through a manikin in the classroom environment without and with acoustic panels and at different seat positions. Participants were asked to repeat each word exactly as it was heard. During the secondary task, participants were shown a single, random string of five digits before the presentation of the monosyllabic words. After each list in the primary task was completed, participants were asked to recall the string of five digits verbatim. Word-recognition and digit-recall performance decreased with the presence of acoustic panels and as the distance from the target signal to a given seat location increased. The results were validated using the STI, as indicated by a decrease in the transmission of the target signal in the presence of acoustic panel and as the distance to a given seat location increased. The inclusion of acoustic panels reduced the negative effects of noise and reverberation in a classroom environment, resulting in an acoustic climate that complied with the ANSI-recommended guidelines for classroom design. Results, however, revealed that participants required an increased amount of mental effort when the classroom was modified with acoustic treatment compared to no acoustic treatment. Independent of acoustic treatment, mental effort was greatest at seat locations beyond the critical distance (CD). With the addition of acoustic panels, mental effort was found to increase significantly at seat locations beyond the CD compared to the unmodified room condition. Overall, results indicate that increasing the distance between the teacher and child has a detrimental impact on mental effort and, ultimately, academic performance.

Speech intelligibility assessment of protective facemasks and air-purifying respirators

ABSTRACTSpeech Intelligibility (SI) is the perceived quality of sound transmission. In healthcare settings, the ability to communicate clearly with coworkers, patients, etc., is crucial to quality patient care and safety. The objectives of this study were to: (1) assess the suitability of the Speech Transmission Index (STI) methods for testing reusable and disposable facial and respiratory personal protective equipment (protective facemasks [PF], N95 filtering facepiece respirators [N95 FFR], and elastomeric half-mask air-purifying respirators [EAPR]) commonly worn by healthcare workers; (2) quantify STI levels of these devices; and (3) contribute to the scientific body of knowledge in the area of SI. SI was assessed using the STI under two experimental conditions: (1) a modified version of the National Fire Protection Association 1981 Supplementary Voice Communications System Performance Test at a Signal to Noise Ratio (SNR) of −15 (66 dBA) and (2) STI measurements utilizing a range of modified pink noise levels (52.5 dBA (−2 SNR) – 72.5 dBA (+7 SNR)) in 5.0 dBA increments. The PF models (Kimberly Clark 49214 and 3 M 1818) had the least effect on SI interference, typically deviating from the STI baseline (no-mask condition) by 3% and 4% STI, respectively. The N95FFR (3 M 1870, 3 M 1860) had more effect on SI interference, typically differing from baseline by 13% and 17%, respectively, for models tested. The EAPR models (Scott Xcel and North 5500) had the most significant impact on SI, differing from baseline by 42% for models tested. This data offers insight into the performance of these apparatus with respect to STI and may serve as a reference point for future respirator design considerations, standards development, testing and certification activities.

Linking design decisions to speech intelligibility

What is the largest classroom that will support unamplified communication? What design attributes and acoustical conditions contribute to speech intelligibility, and how much should each design attribute be weighted when predicting speech transmission index (STI)? This line of inquiry aims to correlate room design and mechanical equipment (HVAC) design to speech intelligibility (STI) in university lecture rooms. It seeks to link the STI of classrooms at Virginia Tech to ambient noise levels, reverberation time, impulse response characteristics, room volume, number of seats, distance to nearest HVAC terminal device, receiver location, and speaker source power. Ambient noise measurements and STI will be mapped on a 1 m x 1 m grid for a subset of the measured rooms to visualize the noise on the listener plane and better estimate the haptic component of signal-to-noise ratio in the spaces. The research also will explore quirks in the STI measurement itself, including differences between the direct and indirect measurement methods, differences permitted for test speech level, and differences from seat to seat within a single room.

Comparison of source stimulus input method on measured speech transmission index values of sound reinforcement systems

One purpose of a sound reinforcement system is to increase the talker’s speech intelligibility. A common metric for speech intelligibility evaluation is the Speech Transmission Index (STI) defined by IEC-60268-16 Revision 4. The STI of a sound reinforcement system can be measured by inputting a stimulus signal into the sound system, which is modified by the system electronics, and radiated by the sound system loudspeakers to the audience seats. The stimulus signal can be input via a line level connection to the sound system or by playing the stimulus signal through a small loudspeaker that is picked-up by a sound system microphone. This latter approach factors the entire sound system signal chain from microphone input to loudspeaker output. STI measurements were performed on two sound systems, one in a reverberant room and the other in relatively non-reverberant room. Measurement results compare both signal input techniques using omnidirectional and hypercardioid sound system microphones and three loudspeakers claimed to be designed to have directivity characteristics similar to the human voice.

Uncertainties in speech transmission index measurements

A detailed study has been carried out into the uncertainties associated with Speech Transmission Index (STI) measurements. The uncertainties and measurement errors are shown to be either systematic or random in nature. Systematic errors were found to include limitations of the technique itself as well as uncertainties related to measurement hardware and software implementations. Systematic errors were found to be caused by a range of issues including measurement microphone properties, test signal generation and replay errors, talker loudspeaker directivity and frequency response variations, and measurement system algorithms. Some forms of digital signal processing are also shown to affect the measured result and were found to be highly dependent upon the nature of the test signal and its processing. Measurement uncertainties due to random errors were found to include out of band, high sound pressure level, low frequency modulations or overloading of the microphone preamplifier stages, the pseudo random nature of the STI test signal itself, and the sensitivity of some test signals, such as maximal length sequences, to short term environmental and acoustic changes. Changes and fluctuations in the background noise level during measurements were also found to be a significant cause of error and uncertainly.

Some effects of reflections and delayed sound arrivals on the perception of speech and corresponding measurements of the speech transmission index

Although the effects of reflections and later arriving sound repetitions (echoes) have been well researched and published over the past 60 years—ranging from Haas, Wallach, and more recently to Bradley & Sato and Toole, their effect on Speech Transmission Index measurements and assessments has only be cursorily studied. Over the past 20 years, the speech Transmission Index (STI) has become the most widely employed measure of potential speech intelligibility for both natural speech and more importantly of Public Address and emergency sound systems and Voice Alarms. There is a common perception that STI can fully account for echoes and late, discrete sound arrivals and reflections. The paper shows this not to be the case but that sound systems achieving high STI ratings can exhibit poor and unacceptable speech intelligibility due to the presence of late sound arrivals and echoes. The finding is based on the results of a series of listening tests and extensive sound system modeling, simulations and measurements. The results of the word score experiments were found to be highly dependent upon the nature of the test material and presentation.

Logatom corpus for the assessment of the intelligibility in Spanish speaking environments and its relation with STI measurements

A 1000 consonant–vowel–consonant structure logatoms corpus (CVC-structure), grouped in 10 phonetically equally balanced lists of 100 words each, was developed to satisfy the need of subjective assessment of speech intelligibility in American Spanish speaking environments. This corpus was tested and correlated with the Speech Transmission Index (STI) measurements to compare its articulation intelligibility score with other lists’ scores.Through the development of this work it was determined that in two different acoustically poor rooms that have the same STI (with STI<0.50), the intelligibility score is lower when the articulation test is performed in a quiet room with high reverberation time than when it is performed in a very noisy room with low reverberation time. The final correlation curve of the American Spanish CVC-structure corpus was around 10% points higher than the CVCEQB curve obtained by Steeneken and Houtgast in 2002.

Comparison of the articulation test and speech transmission index values measured in two different acoustical poor conditions

In this work, a comparison is made between articulation tests and speech transmission index (STI) values, measured in the same place, in two different acoustically poor room conditions, both with STI values of less than 0.4. In the first room condition, a 200 m3 reverberation chamber with background noise of less than 32 dBA was used. In the second room condition, pink noise was added to a small room with reverberation time of less than 0.6 s, till values of STI lower than 0.4 were reached. The articulation test corpus consisted in a 1000 phonetically Spanish combination of a consonant, vowel, and consonant (CVC logatoms). The logatoms were recorded in an anechoic chamber. In the articulation test and STI measurement, both signals were emitted in the rooms using a NTI Talkbox with a sound power equivalent to a normal human voice. The STI and articulation was measured at the listener's seats which were located at different distances from the source but within STI values less than 0.4. The articulation test results of both acoustical conditions are correlated separately with the measured STI. The results of the measurements indicate that for the same STI value, the subjective response statistically differs.

Speech Intelligibility Measurements in an Office Building

This study investigates how changes to an emergency voice communication system (EVCS) or the building environment affect speech intelligibility in four spaces of an office building: lobby, stairwell, dining area, and open-floor-plan office space. Effects of additional or alternate speaker locations, floor and wall coverings, and background noise on intelligibility are measured via objective means (i.e., meters that evaluated the speech transmission index (STI) of a test signal broadcast via the EVCS). Current practice recommends measurement spacing that appears to appropriately characterize the single-floor spaces, although changes to the environment can reduce STI measurements by nearly 20%. Several suggestions address concerns of when and where to perform speech intelligibility testing.

Is the Speech Transmission Index (STI) a robust measure of sound system speech intelligibility performance?

Although RaSTI is a good indicator of the speech intelligibility capability of auditoria and similar spaces, during the past 2–3 years it has been shown that RaSTI is not a robust predictor of sound system intelligibility performance. Instead, it is now recommended, within both national and international codes and standards, that full STI measurement and analysis be employed. However, new research is reported, that indicates that STI is not as flawless, nor robust as many believe. The paper highlights a number of potential error mechanisms. It is shown that the measurement technique and signal excitation stimulus can have a significant effect on the overall result and accuracy, particularly where DSP-based equipment is employed. It is also shown that in its current state of development, STI is not capable of appropriately accounting for a number of fundamental speech and system attributes, including typical sound system frequency response variations and anomalies. This is particularly shown to be the case when a system is operating under reverberant conditions. Comparisons between actual system measurements and corresponding word score data are reported where errors of up to 50 implications for VA and PA system performance verification will be discussed.

Improving the intelligibility of aircraft PA systems

The speech intelligibility of commercial aircraft audio systems is of great importance. The authors have noted however, that subjectively, the intelligibility of such systems can vary considerably. Measurements made on a number of commercial flights indicate that signal to noise ratios are often around 0 dBA and frequently negative. As aircraft loudspeakers are located relatively close to the passenger, the collapsing narrower dispersion and beaming of the high frequencies from conventional cone loudspeakers can cause loss of intelligibility for passengers not on the axis of the loudspeaker and produce uneven sound distribution within the cabin. Distributed mode loudspeakers (DML) have demonstrated intelligibility superior to conventional loudspeakers in other applications such as public address systems and the opportunity was taken to evaluate their performance within an aircraft cabin environment. DMLs are very light in weight and can even be built into the internal trim, resulting in a near zero weight loudspeaker. Their essentially flat configuration also requires less volume. This paper presents a comparative evaluation of the speech intelligibility of current aircraft loudspeakers and DMLs as determined by both conventional word score tests and Speech Transmission Index measurements using both full STI and RaSTI. The tests were performed under both quiet and representative in-flight cabin noise conditions.

Factors influencing perceived listening quality in university classrooms

The speech transmission index (STI) can be used to describe acoustical quality in classrooms. The objective of the work reported here was to establish the relationship between STI and user subjective perception of the listening environment by studying thirty university classrooms. STI measurements were made in these rooms when unoccupied. A questionnaire was developed to create a summary measure of perceived listening quality (PLQ) and to enquire about user-related factors that could affect PLQ. The questionnaire was administered to over 5700 students in 107 classes in the 30 rooms. Preliminary results suggest that the main factors contributing to increased PLQ are the perceived instructor speech level, the perceived quality of the classroom air and lighting, the proportion of time that the instructor speaks toward the audience, and the ‘‘unoccupied’’ STI. Use of a speech-reinforcement system reduced PLQ. [See NOISE-CON Proceedings for full paper.]