Feasibility of subjective speech intelligibility assessment based on auralization

Subjective speech intelligibility can be assessed by speech recorded in an anechoic chamber and then convolved with room impulse responses that can be created by acoustic simulation. The speech intelligibility (SI) assessment based on auralization was validated in three rooms. The articulation scores obtained from simulated sound field were compared with the ones from measured sound field and from direct listening in rooms. Results show that the speech intelligibility prediction based on auralization technique with simulated binaural room impulse responses (BRIRs) is in agreement with reality and results from measured BRIRs. When this technique is used with simulated and measured monaural room impulse responses (MRIRs), the predicted results underestimate the reality. It has been shown that auralization technique with simulated BRIRs is capable of assessing subjective speech intelligibility of listening positions in the room.     

Chinese speech intelligibility at different speech sound pressure levels and signal-to-noise ratios in simulated classrooms

The speech intelligibility in classroom can be influenced by background-noise levels, speech sound pressure level (SSPL), reverberation time and signal-to-noise ratio (SNR). The relationship between SSPL and subjective Chinese Mandarin speech intelligibility and the effect of different SNRs on Chinese Mandarin speech intelligibility in the simulated classroom were investigated through room acoustical simulation, auralisation technique and subjective evaluation. Chinese speech intelligibility test signals recorded in anechoic chamber were convolved with the simulated binaural room impulse responses, and then reproduced through the headphone by different SSPLs and SNRs. The results show that Chinese Mandarin speech intelligibility scores increase with increasing of SSPLs and SNRs within a certain range in simulated classrooms. Chinese Mandarin speech intelligibility scores have no significant difference with SNRs of no less than 15 dBA under the same reverberation time condition.     

Phonemic restoration effect reversed in a reverberant room

Classic demonstrations of the phonemic restoration effect show increased intelligibility of interrupted speech when the interruptions are caused by a plausible masking sound rather than by silent periods. Previous studies of this effect have been conducted exclusively under anechoic or nearly anechoic listening conditions. This study demonstrates that the effect is reversed when sounds are presented in a realistically simulated reverberant room (broadband T(60) = 1.1 s): intelligibility is greater for silent interruptions than for interruptions by unmodulated noise. Additional results suggest that the reversal is primarily due to filling silent intervals with reverberant energy from the speech signal.     

Localizing nearby sound sources in a classroom: binaural room impulse responses

Binaural room impulse responses (BRIRs) were measured in a classroom for sources at different azimuths and distances (up to 1 m) relative to a manikin located in four positions in a classroom. When the listener is far from all walls, reverberant energy distorts signal magnitude and phase independently at each frequency, altering monaural spectral cues, interaural phase differences, and interaural level differences. For the tested conditions, systematic distortion (comb-filtering) from an early intense reflection is only evident when a listener is very close to a wall, and then only in the ear facing the wall. Especially for a nearby source, interaural cues grow less reliable with increasing source laterality and monaural spectral cues are less reliable in the ear farther from the sound source. Reverberation reduces the magnitude of interaural level differences at all frequencies; however, the direct-sound interaural time difference can still be recovered from the BRIRs measured in these experiments. Results suggest that bias and variability in sound localization behavior may vary systematically with listener location in a room as well as source location relative to the listener, even for nearby sources where there is relatively little reverberant energy.     

Subjective and objective verifications of the inverse functions of binaural room impulse responses

The binaural room impulse responses (BRIRs) can be applied to 3-D sound field reconstruction, virtual reality, noise control, et al. Because the BRIRs are non-minimum phase functions, it is difficult to find the exact inverse functions of the BRIRs, especially when there are two or more sources in a reverberant space. In this work, a method was proposed to find the inverse functions of BRIRs with two sound sources in a reverberant space. The concept of time delays and the method of weighted least squares were used to find the causal, however, approximate inverse functions. The accuracy of the inverse functions was first evaluated objectively by a dummy head system. The result shows that the distortion due to crosstalk and room reverberation can be improved by 16∼18 dB. The inverse functions were also verified subjectively by 20 students. The result of subjective evaluation also shows that the inverse functions can be used successfully to reduce the crosstalk effect and the room reverberation.     

Prediction of the influence of reverberation on binaural speech intelligibility in noise and in quiet

Reverberation usually degrades speech intelligibility for spatially separated speech and noise sources since spatial unmasking is reduced and late reflections decrease the fidelity of the received speech signal. The latter effect could not satisfactorily be predicted by a recently presented binaural speech intelligibility model [Beutelmann et al. (2010). J. Acoust. Soc. Am. 127, 2479-2497]. This study therefore evaluated three extensions of the model to improve its predictions: (1) an extension of the speech intelligibility index based on modulation transfer functions, (2) a correction factor based on the room acoustical quantity "definition," and (3) a separation of the speech signal into useful and detrimental parts. The predictions were compared to results of two experiments in which speech reception thresholds were measured in a reverberant room in quiet and in the presence of a noise source for listeners with normal hearing. All extensions yielded better predictions than the original model when the influence of reverberation was strong, while predictions were similar for conditions with less reverberation. Although model (3) differed substantially in the assumed interaction of binaural processing and early reflections, its predictions were very similar to model (2) that achieved the best fit to the data.     

On the importance of early reflections for speech in rooms

This paper presents the results of new studies based on speech intelligibility tests in simulated sound fields and analyses of impulse response measurements in rooms used for speech communication. The speech intelligibility test results confirm the importance of early reflections for achieving good conditions for speech in rooms. The addition of early reflections increased the effective signal-to-noise ratio and related speech intelligibility scores for both impaired and nonimpaired listeners. The new results also show that for common conditions where the direct sound is reduced, it is only possible to understand speech because of the presence of early reflections. Analyses of measured impulse responses in rooms intended for speech show that early reflections can increase the effective signal-to-noise ratio by up to 9 dB. A room acoustics computer model is used to demonstrate that the relative importance of early reflections can be influenced by the room acoustics design.     

Blind estimation of reverberation time

The reverberation time (RT) is an important parameter for characterizing the quality of an auditory space. Sounds in reverberant environments are subject to coloration. This affects speech intelligibility and sound localization. Many state-of-the-art audio signal processing algorithms, for example in hearing-aids and telephony, are expected to have the ability to characterize the listening environment, and turn on an appropriate processing strategy accordingly. Thus, a method for characterization of room RT based on passively received microphone signals represents an important enabling technology. Current RT estimators, such as Schroeder's method, depend on a controlled sound source, and thus cannot produce an online, blind RT estimate. Here, a method for estimating RT without prior knowledge of sound sources or room geometry is presented. The diffusive tail of reverberation was modeled as an exponentially damped Gaussian white noise process. The time-constant of the decay, which provided a measure of the RT, was estimated using a maximum-likelihood procedure. The estimates were obtained continuously, and an order-statistics filter was used to extract the most likely RT from the accumulated estimates. The procedure was illustrated for connected speech. Results obtained for simulated and real room data are in good agreement with the real RT values.     

Speech transmission performance and the effect of acoustical remedies in a dome

For the purpose of improving speech transmission performance in a dome space, the acoustical properties in a dome having a diameter of 20 m were examined. The acoustical properties measured evenly on the floor of the dome were evaluated both objectively and subjectively and the interrelationship of the objective measures and subjective measures were also examined. Then, on the basis of the results of the study, simplified acoustical remedies were applied to the dome to improve speech intelligibility and the effect of the remedies was also examined. The following findings were obtained from this investigation.(1) The speech transmission performance in the dome space without treatment by absorptive materials varies greatly with the locations of sound sources and observation points: a range of 0.17-0.59 for RASTI value and a range of 30-97% for speech intelligibility test results. (2) There are peculiar observation points at which speech transmission quality is very high due to a considerable sum of the energy arriving in the first 0.06 s after the arrival of the direct sound. (3) Of all the measured acoustical parameters, RASTI, EDT in 1 kHz band, early-to-late arriving sound energy ratio, and Ts corresponded well to the speech intelligibility test scores. (4) Rubber tiles, cotton canvas 12 m in length, and glass wool board, are effective in improving speech intelligibility remarkably due to increased sound absorption and the diffusion effect.     

A system for simulating room acoustical environments for one’s own voice

The real-time simulation of room acoustical environments for one’s own voice using generic software has been difficult until very recently due to the computational load involved: requiring real-time convolution of a person’s voice with a potentially large number of long room impulse responses. This paper describes a software-based solution that accomplishes real-time convolution with head-tracking to simulate the effect of room acoustical environments on the sound of one’s own voice, using binaural technology. Actual rooms are characterized by measuring the room impulse response from the mouth to ears of the same head (oral binaural room impulse response, OBRIR). By repeating this process at 2° yaw increments for a given head position, the rooms are binaurally scanned around a given position to obtain a collection of OBRIRs, which is then used by the software-based simulation system. In the simulated rooms, a person equipped with a near-mouth microphone and near-ear loudspeakers can speak or sing and hear their voice, as it would sound in the recorded rooms, while physically being in an anechoic room. By continually updating the person’s head orientation using head-tracking, the corresponding OBRIR is chosen for convolution with their voice. The system described in this paper achieves the low latency that is required to simulate nearby reflections, and it can perform convolution with long room impulse responses.     

Source characterization using recordings made in a reverberant underwater channel

The ability to accurately characterize an underwater sound source is an important prerequisite for many applications including detection, classification, monitoring and mitigation. Unfortunately, anechoic underwater recording environments required to make ideal recordings are generally not available. This paper presents a practical approach to source characterization when working in an imperfect recording environment; the source spectrum is obtained by equalizing the recording with the inverse of the channel’s impulse response (IR). An experiment was conducted in a diving well (depth of 5.18 m) using a logarithmic chirp to obtain the IR. IR length is estimated using methods borrowed from room acoustics and inversion of non-minimum phase IR is accomplished separately in the time and frequency domain to allow for a direct comparison. Results indicate that the energy of controlled sources can be recovered with root-mean-square error of −70 dB (10–70 kHz band). Two equations, one coherent and the other incoherent, are presented to calculate source spectral levels of an unknown source in a reverberant environment. This paper introduces a practical procedure outlining steps to obtain an anechoic estimate of an unknown source using equipment generally available in an acoustic laboratory.     

Effect of source spectrum on sound localization in an everyday reverberant room

Two experiments explored how frequency content impacts sound localization for sounds containing reverberant energy. Virtual sound sources from thirteen lateral angles and four distances were simulated in the frontal horizontal plane using binaural room impulse responses measured in an everyday office. Experiment 1 compared localization judgments for one-octave-wide noise centered at either 750 Hz (low) or 6000 Hz (high). For both band-limited noises, perceived lateral angle varied monotonically with source angle. For frontal sources, perceived locations were similar for low- and high-frequency noise; however, for lateral sources, localization was less accurate for low-frequency noise than for high-frequency noise. With increasing source distance, judgments of both noises became more biased toward the median plane, an effect that was greater for low-frequency noise than for high-frequency noise. In Experiment 2, simultaneous presentation of low- and high-frequency noises yielded performance that was less accurate than that for high-frequency noise, but equal to or better than for low-frequency noise. Results suggest that listeners perceptually weight low-frequency information heavily, even in reverberant conditions where high-frequency stimuli are localized more accurately. These findings show that listeners do not always optimally adjust how localization cues are integrated over frequency in reverberant settings.     

Relationship between listening difficulty and acoustical objective measures in reverberant sound fields

The previous work [Morimoto et al., J. Acoust. Soc. Am. 116, 1607-1613] showed that listening difficulty ratings can be used to evaluate speech transmission performance more exactly and sensitively than intelligibility. Meanwhile, speech transmission performance is usually evaluated using acoustical objective measures, which are directly associated with physical parameters of room acoustic design. However, the relationship between listening difficulty ratings and acoustical objective measures was not minutely investigated. In the present study, a total of 96 impulse responses were used to investigate the relationship between listening difficulty ratings and several objective measures in unidirectional sound fields. The result of the listening test showed that (1) the correlation between listening difficulty ratings and speech transmission index (STI) is the strongest of all tested objective measures, and (2) A-weighted D(50), C(50), and center time, which are obtained from the impulse responses passed through an A-weighted filter, also strongly correlate with listening difficulty ratings, and their correlations with listening difficulty ratings are not statistically different from the correlation between listening difficulty ratings and STI.     

Prediction of speech intelligibility in spatial noise and reverberation for normal-hearing and hearing-impaired listeners

Binaural speech intelligibility of individual listeners under realistic conditions was predicted using a model consisting of a gammatone filter bank, an independent equalization-cancellation (EC) process in each frequency band, a gammatone resynthesis, and the speech intelligibility index (SII). Hearing loss was simulated by adding uncorrelated masking noises (according to the pure-tone audiogram) to the ear channels. Speech intelligibility measurements were carried out with 8 normal-hearing and 15 hearing-impaired listeners, collecting speech reception threshold (SRT) data for three different room acoustic conditions (anechoic, office room, cafeteria hall) and eight directions of a single noise source (speech in front). Artificial EC processing errors derived from binaural masking level difference data using pure tones were incorporated into the model. Except for an adjustment of the SII-to-intelligibility mapping function, no model parameter was fitted to the SRT data of this study. The overall correlation coefficient between predicted and observed SRTs was 0.95. The dependence of the SRT of an individual listener on the noise direction and on room acoustics was predicted with a median correlation coefficient of 0.91. The effect of individual hearing impairment was predicted with a median correlation coefficient of 0.95. However, for mild hearing losses the release from masking was overestimated.     

Effects of a single reflection with varied horizontal angle and time delay on speech intelligibility.

Previously, almost all physical measures for estimating speech intelligibility in a room have been derived from only temporal-monaural criteria. This paper shows that speech intelligibility for a sound field with a single reflection depends not only on the temporal-monaural factor but also on the spatial-binaural factor of the sound field. Articulation tests for sound fields simulated with a single reflection of delay time delta t1 after the direct sound were conducted changing the horizontal incident angle xi of the reflection. Remarkable findings are as followings: (1) speech intelligibility (SI) decreases with increasing delay time delta t1, (2) SI increases when xi approaches 90 degrees; the horizontal angle of the reflection causes a significant effect on SI, and (3) the analysis of variance for articulation test scores clearly demonstrated that the effects of both delta t1 and xi on SI are fully independent. Concerning result (2), if listeners get a spatial separation of signals at the two ears, then the listener's capability for speech perception is assumed to be improved due to "adding" further information to the temporal pattern recognition.     

The effect of overlap-masking on binaural reverberant word intelligibility

Reverberation interferes with the ability to understand speech in rooms. Overlap-masking explains this degradation by assuming reverberant phonemes endure in time and mask subsequent reverberant phonemes. Most listeners benefit from binaural listening when reverberation exists, indicating that the listener's binaural system processes the two channels to reduce the reverberation. This paper investigates the hypothesis that the binaural word intelligibility advantage found in reverberation is a result of binaural overlap-masking release with the reverberation acting as masking noise. The tests utilize phonetically balanced word lists (ANSI-S3.2 1989), that are presented diotically and binaurally with recorded reverberation and reverberation-like noise. A small room, 62 m3, reverberates the words. These are recorded using two microphones without additional noise sources. The reverberation-like noise is a modified form of these recordings and has a similar spectral content. It does not contain binaural localization cues due to a phase randomization procedure. Listening to the reverberant words binaurally improves the intelligibility by 6.0% over diotic listening. The binaural intelligibility advantage for reverberation-like noise is only 2.6%. This indicates that binaural overlap-masking release is insufficient to explain the entire binaural word intelligibility advantage in reverberation.     

Relationship between listening difficulty rating and objective measures in reverberant and noisy sound fields for young adults and elderly persons

Listening difficulty ratings [Morimoto et al., J. Acoust. Soc. Am. 116, 1607-1613 (2004)] were obtained for 20 young adult listeners and 34 elderly listeners in reverberant and noisy sound fields simulated in an anechoic room. The listening difficulty ratings were compared with acoustical objective measures. The results and analyses showed the following: (i) The correlation between listening difficulty ratings and the revised speech transmission index (STI(r)), and that for the useful-detrimental ratio (U(50)) were high, regardless of the age of the listeners. (ii) STI(r) and U(50) need to be increased by 0.12 and 4.2 dB, respectively, to equalize the listening difficulty ratings for the elderly listeners with those for the young listeners. (iii) The estimation accuracies for STI(r) and U(50) can be improved by calculating them with the L(eq) of background noise linearly increased by 4 to 10 dB, which depends on the age of the listeners and the objective measures. However, the improvement was not statistically significant for the elderly listeners.     

Binaural prediction of speech intelligibility in reverberant rooms with multiple noise sources

When speech is in competition with interfering sources in rooms, monaural indicators of intelligibility fail to take account of the listener's abilities to separate target speech from interfering sounds using the binaural system. In order to incorporate these segregation abilities and their susceptibility to reverberation, Lavandier and Culling [J. Acoust. Soc. Am. 127, 387-399 (2010)] proposed a model which combines effects of better-ear listening and binaural unmasking. A computationally efficient version of this model is evaluated here under more realistic conditions that include head shadow, multiple stationary noise sources, and real-room acoustics. Three experiments are presented in which speech reception thresholds were measured in the presence of one to three interferers using real-room listening over headphones, simulated by convolving anechoic stimuli with binaural room impulse-responses measured with dummy-head transducers in five rooms. Without fitting any parameter of the model, there was close correspondence between measured and predicted differences in threshold across all tested conditions. The model's components of better-ear listening and binaural unmasking were validated both in isolation and in combination. The computational efficiency of this prediction method allows the generation of complex "intelligibility maps" from room designs.     

Intelligibility of reverberant noisy speech with ideal binary masking

For a mixture of target speech and noise in anechoic conditions, the ideal binary mask is defined as follows: It selects the time-frequency units where target energy exceeds noise energy by a certain local threshold and cancels the other units. In this study, the definition of the ideal binary mask is extended to reverberant conditions. Given the division between early and late reflections in terms of speech intelligibility, three ideal binary masks can be defined: an ideal binary mask that uses the direct path of the target as the desired signal, an ideal binary mask that uses the direct path and early reflections of the target as the desired signal, and an ideal binary mask that uses the reverberant target as the desired signal. The effects of these ideal binary mask definitions on speech intelligibility are compared across two types of interference: speech shaped noise and concurrent female speech. As suggested by psychoacoustical studies, the ideal binary mask based on the direct path and early reflections of target speech outperforms the other masks as reverberation time increases and produces substantial reductions in terms of speech reception threshold for normal hearing listeners.