Objective Intelligibility Measure Research Articles

Objective intelligibility measurement of reverberant vocoded speech for normal-hearing listeners: Towards facilitating the development of speech enhancement algorithms for cochlear implants.

Cochlear implant (CI) recipients often struggle to understand speech in reverberant environments. Speech enhancement algorithms could restore speech perception for CI listeners by removing reverberant artifacts from the CI stimulation pattern. Listening studies, either with cochlear-implant recipients or normal-hearing (NH) listeners using a CI acoustic model, provide a benchmark for speech intelligibility improvements conferred by the enhancement algorithm but are costly and time consuming. To reduce the associated costs during algorithm development, speech intelligibility could be estimated offline using objective intelligibility measures. Previous evaluations of objective measures that considered CIs primarily assessed the combined impact of noise and reverberation and employed highly accurate enhancement algorithms. To facilitate the development of enhancement algorithms, we evaluate twelve objective measures in reverberant-only conditions characterized by a gradual reduction of reverberant artifacts, simulating the performance of an enhancement algorithm during development. Measures are validated against the performance of NH listeners using a CI acoustic model. To enhance compatibility with reverberant CI-processed signals, measure performance was assessed after modifying the reference signal and spectral filterbank. Measures leveraging the speech-to-reverberant ratio, cepstral distance and, after modifying the reference or filterbank, envelope correlation are strong predictors of intelligibility for reverberant CI-processed speech.

Comparison of intelligibility measures for children with velopharyngeal insufficiency: visual analog scale ratings, interval scales, and orthographic transcription (OT)-based measures

ABSTRACT Regardless of the underlying cause for speech impairment in speakers with cleft palate, a universal consequence of cleft palate is reduced speech intelligibility. Still, there is no standardised approach for measuring intelligibility for speakers with cleft speech. The current study aimed to determine the relationship between orthographic transcription (OT)-based measures, interval-scale ratings, and visual analog scale (VAS) ratings for perceptual judgements of intelligibility in speakers with cleft palate as judged by speech-language pathologists (SLPs). The speaker participants were six speakers with velopharyngeal insufficiency secondary to cleft palate. Four sets of sentences from the Hearing in Noise Test were recorded from each speaker. A total of 14 SLPs provided their intelligibility judgement on these speaker’s recordings by word-by-word orthographic transcriptions, a visual analog scale (0–100), and a 5-point interval rating scale. A Spearman rank correlation test indicated a negative, strong correlation between OT-based measurements and VAS scores (r = −.94; p = 0.01) and between OT-based measurements and interval rating scores (r = −.77, p = 0.01). A strong, positive correlation was found between scores obtained from VAS and interval rating scales (r = .83, p = 0.05). The strong relationship between the objective measure of intelligibility (i.e. OT-based measure) and a subjective measure of intelligibility (i.e. VAS and interval scale) supports using a less time-consuming VAS as a substitute for orthographic transcription in measuring intelligibility in cleft palate speech.

Characterization of the Intelligibility of Vowel–Consonant–Vowel (VCV) Recordings in Five Languages for Application in Speech-in-Noise Screening in Multilingual Settings

The purpose of this study is to characterize the intelligibility of a corpus of Vowel–Consonant–Vowel (VCV) stimuli recorded in five languages (English, French, German, Italian and Portuguese) in order to identify a subset of stimuli for screening individuals of unknown language during speech-in-noise tests. The intelligibility of VCV stimuli was estimated by combining the psychometric functions derived from the Short-Time Objective Intelligibility (STOI) measure with those derived from listening tests. To compensate for the potential increase in speech recognition effort in non-native listeners, stimuli were selected based on three criteria: (i) higher intelligibility; (ii) lower variability of intelligibility; and (iii) shallower psychometric function. The observed intelligibility estimates show that the three criteria for application in multilingual settings were fulfilled by the set of VCVs in English (average intelligibility from 1% to 8% higher; SRT from 4.01 to 2.04 dB SNR lower; average variability up to four times lower; slope from 0.35 to 0.68%/dB SNR lower). Further research is needed to characterize the intelligibility of these stimuli in a large sample of non-native listeners with varying degrees of hearing loss and to determine the possible effects of hearing loss and native language on VCV recognition.

On DCT-based MMSE estimation of short time spectral amplitude for single-channel speech enhancement

This paper proposes Discrete Cosine Transform (DCT) based speech enhancement algorithms. These algorithms utilize minimum mean square error (MMSE) estimator of clean short-time spectral amplitude, which respectively uses Gaussian, Laplace and Gamma probability density functions (PDF) as speech priors. We consider the noise process is additive and Gaussian. The proposed estimators are closed-form solutions, whereas the conventional Discrete Fourier Transform (DFT) based estimators derived under super-Gaussian speech priors have no closed-form solutions. We also examine the estimators with the Speech Presence Uncertainty (SPU) that addresses the speech or silence problem with probability. Compared to the alternative approaches, such as the Ephraim and Malah or the Erkelens et.al MMSE-STSA estimators, the proposed methods demonstrate superior performance in terms of Segmental SNR (SegSNR), Perceptual Evaluation of Speech Quality (PESQ), short-time objective intelligibility measure (STOI), and mean subjective preference score, while exhibiting an equal or lower complexity.

Performance analysis of low complexity fully connected neural networks for monaural speech enhancement

We compare the run-time complexity of recent deep neural network (DNN) and non-DNN based monaural speech enhancement algorithms. Specifically, we consider fully connected, convolutional, and genetic-algorithm based DNNs and compare their performance to the image analysis technique, which is non-DNN based. It is demonstrated that for the same speech enhancement performance, a simple fully connected DNN has the lowest run-time computational complexity in terms of floating-point operations and execution time on a standard laptop. The objective indices used for the evaluation of the speech enhancement performance are the perceptual evaluation of speech quality and short-time objective intelligibility measures. In addition, the subjective intelligibility measures involved in the experiment are the modified rhyme test and the mean opinion score. Both stationary and non-stationary noise in addition to interfering speech is considered.

Ultrasonic Doppler Based Silent Speech Interface Using Perceptual Distance

Moderate performance in terms of intelligibility and naturalness can be obtained using previously established silent speech interface (SSI) methods. Nevertheless, a common problem associated with SSI has involved deficiencies in estimating the spectrum details, which results in synthesized speech signals that are rough, harsh, and unclear. In this study, harmonic enhancement (HE), was used during postprocessing to alleviate this problem by emphasizing the spectral fine structure of speech signals. To improve the subjective quality of synthesized speech, the difference between synthesized and actual speech was established by calculating the distance in the perceptual domains instead of using the conventional mean square error (MSE). Two deep neural networks (DNNs) were employed to separately estimate the speech spectra and the filter coefficients of HE, connected in a cascading manner. The DNNs were trained to incrementally and iteratively minimize both the MSE and the perceptual distance (PD). A feasibility test showed that the perceptual evaluation of speech quality (PESQ) and the short-time objective intelligibility measure (STOI) were improved by 17.8 and 2.9%, respectively, compared with previous methods. Subjective listening tests revealed that the proposed method yielded perceptually preferred results compared with that of the conventional MSE-based method.

94 GHz Asymmetric Antenna Radar for Speech Signal Detection and Enhancement via Variational Mode Decomposition and Improved Threshold Strategy

To further improve the detection distance and sensitivity of bio-radar, a 94 GHz asymmetric antenna radar sensor is employed to detect speech signal. However, the radar speech is often mixed with various noise, which will seriously affect the quality and intelligibility of the speech signal. Therefore, a novel method based on variational mode decomposition (VMD) and improved threshold strategy (ITS) is proposed in this paper for improving the quality and intelligibility of the radar speech. VMD is a novel adaptive decomposition method, which overcomes the problem of mode aliasing and end effect in empirical mode decomposition (EMD). ITS can overcome the limitation of traditional wavelet threshold and achieve the best compromise between speech intelligibility and noise reduction. Firstly, EMD is applied to determine the number of decomposition level, and then radar speech is decomposed into several limited bandwidth intrinsic mode functions by VMD. Secondly, ITS is employed to remove noise from useful modes which are determined by Pearson correlation coefficient (PCC). The performance of the proposed method is evaluated by perceptual evaluation of speech quality (PESQ), short-time objective intelligibility (STOI) and composite measures (CMs). The experimental results show that the radar sensor can detect long distance speech signal and the proposed method can effectively improve the quality and intelligibility of the radar speech signal. Due to the good performance, the proposed method will provide a promising alternative for various applications related to radar speech and traditional microphone speech signal enhancement.

From Speech Acoustics to Communicative Participation in Dysarthria: Toward a Causal Framework.

We proposed and tested a causal instantiation of the World Health Organization's International Classification of Functioning, Disability and Health (ICF) framework, linking acoustics, intelligibility, and communicative participation in the context of dysarthria. Speech samples and communicative participation scores were collected from individuals with dysarthria (n = 32). Speech was analyzed for two acoustic metrics (i.e., articulatory precision and speech rate), and an objective measure of intelligibility was generated from listener transcripts. Mediation analysis was used to evaluate pathways of effect between acoustics, intelligibility, and communicative participation. We observed a strong relationship between articulatory precision and intelligibility and a moderate relationship between intelligibility and communicative participation. Collectively, data supported a significant relationship between articulatory precision and communicative participation, which was almost entirely mediated through intelligibility. These relationships were not significant when speech rate was specified as the acoustic variable of interest. The statistical corroboration of our causal instantiation of the ICF framework with articulatory acoustics affords important support toward the development of a comprehensive causal framework to understand and, ultimately, address restricted communicative participation in dysarthria.

How hospital noise impacts intelligibility of medically related sentences

Hospital noise often exceeds recommended sound levels set by health organizations, potentially leading to communication breakdowns between healthcare providers and patients. Despite this concern, there is a paucity of literature on speech perception in healthcare settings. Studies investigating how hospital noise impacts intelligibility have been limited by not using medical terminology or not including objective intelligibility measures with listeners. To address these gaps, a new corpus of medically related sentences, with quantified word frequency, word familiarity, and sentence predictability, was employed in an intelligibility task. Recordings of the medically related sentences with different familiarity/frequency types (low/low, high/low, high/high) plus standard speech perception sentences were presented to listeners in quiet, speech-shaped noise, or hospital noise for transcription. Results showed main effects of sentence type and listening condition and an interaction between the two. Although listeners were less accurate on all sentence types in noise, sentences with low frequency and/or familiarity words showed the greatest intelligibility decreases in noise. Thus, miscommunications between healthcare providers and patients could occur when medical information is provided in noisy settings like hospitals, situations that could lead to lack of comprehension of discharge instructions or other medical communication errors. [Work supported by the IU Institute for Advanced Study.]

Adjustable deterministic pseudonymization of speech

While public speech resources become increasingly available, there is a growing interest to preserve the privacy of the speakers, through methods that anonymize the speaker information from speech while preserving the spoken linguistic content. In this paper, a method for pseudonymization (reversible anonymization) of speech is presented, that allows to obfuscate the speaker identity in untranscribed running speech. The approach manipulates the spectro-temporal structure of the speech to simulate a different length and structure of the vocal tract by modifying the formant locations, as well as by altering the pitch and speaking rate. The method is deterministic and partially reversible, and the changes are adjustable on a continuous scale. The method has been evaluated in terms of (i) ABX listening experiments, and (ii) automatic speaker verification and speech recognition. ABX experimental results indicate that the speaker identifiability among forced choice pairs reduced from over 90% to less than 70% through pseudonymization, and that de-pseudonymization was partially effective. An evaluation on the VoicePrivacy 2020 challenge data showed that the proposed approach performs better than the signal processing based baseline method that uses McAdams coefficient and performs slightly worse than the neural source filtering based baseline method. Further analysis showed that the proposed approach: (i) is comparable to the neural source filtering baseline based method in terms of phone posterior feature based objective intelligibility measure, (ii) preserves formant tracks better than the McAdams based method, and (iii) preserves paralinguistic aspects such as dysarthria in several speakers.

Assessment of speech intelligibility in university lecture rooms of different sizes using objective and subjective methods

The scores of speech intelligibility, obtained using objective and subjective methods for three university lecture rooms of the small, medium, and large sizes with different degrees of filling, were presented. The problem of achieving high speech intelligibility is relevant for both students and university administration, and for architects designing or reconstructing lecture rooms. Speech intelligibility was assessed using binaural room impulse responses which applied an artificial head and non-professional quality audio equipment for measuring. The Speech Transmission Index was an objective measure of speech intelligibility, while the subjective evaluation of speech intelligibility was carried out using the articulation method. Comparative analysis of the effectiveness of parameters of impulse response as a measure of speech intelligibility showed that Early Decay Time exceeded the score of the T30 reverberation time but was ineffective in a small lecture room. The C50 clarity index for all the considered lecture rooms was the most informative. Several patterns determined by the influence of early sound reflections on speech intelligibility were detected. Specifically, it was shown that an increase in the ratio of the energy of early reflections to the energy of direct sound leads to a decrease in speech intelligibility. The exceptions are small, up to 30‒40 cm, distances from the back wall of the room, where speech intelligibility is usually slightly higher than in the middle of the room. At a distance of 0.7–1.7 m from the side walls of the room, speech intelligibility is usually worse for the ear, which is closer to the wall. The usefulness of the obtained results lies in refining the quantitative characteristics of the influence of early reflections of sound on speech intelligibility at different points of lecture rooms.

Noise reduction optimization of sound sensor based on a Conditional Generation Adversarial Network

To address the problems in the traditional speech signal noise elimination methods, such as the residual noise, poor real-time performance and narrow applications a new method is proposed to eliminate network voice noise based on deep learning of conditional generation adversarial network. In terms of the perceptual evaluation of speech quality (PESQ) and shorttime objective intelligibility measure (STOI) functions used as the loss function in the neural network, which were used as the loss function in the neural network, the flexibility of the whole network was optimized, and the training process of the model simplified. The experimental results indicate that, under the noisy environment, especially in a restaurant, the proposed noise reduction scheme improves the STOI score by 26.23% and PESQ score by 17.18%, respectively, compared with the traditional Wiener noise reduction algorithm. Therefore, the sound sensor’s noise reduction scheme through our approach has achieved a remarkable noise reduction effect, more useful information transmission, and stronger practicability.

Glimpse-based estimation of speech intelligibility from speech-in-noise using artificial neural networks

While human listeners can, to some extent, understand the information conveyed by the speech signal when it is mixed with noise, traditional objective intelligibility measures usually fail to operate without a priori knowledge of the clean speech signal. This hence limits the usability of those measures in situations where the clean speech signal is inaccessible. In this paper a glimpse-based method is extended to make speech intelligibility predictions directly from speech-plus-noise mixtures. Using a neural network, the proposed method estimates the time-frequency regions with a local speech-to-noise ratio above a given threshold – known as glimpses – from the mixture signal, instead of separately comparing the speech signal against the noise signal. The number and locations of the glimpses can then be used to produce an intelligibility score. In Experiment I where listener intelligibilities were measured in one stationary and nine fluctuating noise maskers, the predictions produced by the proposed method were highly correlated with the subjective data, with correlation coefficients above 0.90. In Experiment II, with the same neural network trained on normal natural speech as in Experiment I, the proposed method was used to predict the intelligibility of speech signals modified by intelligibility-enhancement algorithms and synthetic speech. The method can still maintain its predictive power by demonstrating a similar performance to its intrusive counterpart with an overall correlation coefficient of 0.81, which is superior to many modern traditional measures evaluated under the same conditions. Therefore, the proposed method can be used to estimate speech intelligibility in place of traditional measures in conditions where their capacity falls short.

Supervised binaural source separation using auditory attention detection in realistic scenarios

Speech separation in crowded environments is challenging for hearing-impaired listeners. Unfortunately, hearing aids cannot segregate the main speaker from other interferences without knowing to which speaker the listener attends. In this paper, a new robust binaural speech separation system based on a supervised deep neural network (DNN) is introduced to separate attended speaker in different simulated room conditions (i.e., noisy and reverberant room with different azimuthal locations of the target speaker). For each speaker in the mixture of sound sources, a set of DNNs yields estimated ratio masks where one of them is selected by the auditory attention detection (AAD) procedure to separate the attended speech. The detection of the attended speaker among others is achieved by a Bi-directional long short-term memory (Bi-LSTM) network using phase-locking values (PLVs) extracted from the electroencephalography (EEG) signals. Two experiments are conducted to assess the performance of the proposed AAD-based binaural separation system vs. a state-of-the-art monaural segregation system from the literature, used as a baseline. The results of evaluating the AAD processing unit show high detection accuracy as compared with the detection method used in the baseline system. Also, systematic evaluations of the proposed binaural separation system against the baselines confirm its superiority in terms of various objective intelligibility and quality measures. The new AAD-based binaural separation system is advantageous in several aspects. The AAD method determines the direction of attention from single-trial EEG signals without access to audio signals from the speakers. The model presented in this research can be considered as an important processing tool in the structure of neuro-steered hearing aid devices employed in cocktail party scenarios.

Improving Patient Voice Intelligibility by using a Euclidian Distance-Based Approach to Improve Voice Assistant Accuracy

Voice assistance (VA) is gaining domestic consumer attention in a variety of products, such as Amazon Alexa, Google Home, Apple’s Siri, and Microsoft’s Cortana. Furthermore, VA has recently shown its usefulness and ability to improve inpatient experience in hospitals and clinics. Nevertheless, none of the VA products has an accuracy rate greater than 90%. The accuracy decreases even more in noisy or public environments. Hence, improving VA accuracy in noisy environments requires a speech signal algorithm with good quality and intelligibility. There is great interest in developing an objective intelligibility measure that shows maximum correlation with subjective speech intelligibility and that can measure the effect of speech enhancement algorithms on the processing of noisy speech signals. In this paper, Euclidian distance-based speech intelligibility prediction is proposed to measure the correlation with subjective intelligibility in different noisy environments. This paper also presents a comparative analysis and general background research in speech intelligibility improvement. The results show that no single algorithm is effective in improving the intelligibility of speech signals.

The effects of intensive speech treatment on intelligibility in Parkinson's disease: A randomised controlled trial.

BackgroundMore than 6,000,000 individuals worldwide are diagnosed with Parkinson's disease (PD). Nearly 90% develop speech signs that may substantially impair their speech intelligibility, resulting in losses in their communication and quality of life. Benefits of intensive speech treatment have been documented for a range of speech signs. However, the critical question of whether speech is more intelligible after treatment has not been investigated in a randomised controlled trial (RCT). We hypothesised that intensive speech treatment would improve speech intelligibility in PD.MethodSixty-four patients with hypokinetic dysarthria secondary to PD participated in this single-centre, parallel arm, statistically-powered RCT. Reporting follows CONSORT guidelines for non-pharmacological treatment. Patients were recruited from US clinics and randomised using a statistician-derived minimisation algorithm, to intensive speech treatment (16 1-hour sessions/1 month) targeting voice (voice group) or targeting articulation (articulation group) or to an untreated group (no treatment group). Speech treatments were delivered by speech clinicians who specialised in treating patients with PD. Trial design minimised bias and supported equipoise. For intelligibility assessment, blinded listeners (n = 117) orthographically transcribed 57 patients’ recorded, self-generated narrative speech samples, randomly presented in multi-talker babble noise. Listeners were American-English speakers, ages 18–35 years, with normal hearing. The primary outcome was baseline (pre-treatment) to post-treatment change in transcription accuracy (TA), recognised as the most objective measure of intelligibility. TA was defined as the percentage of words transcribed correctly. Listeners, data collectors, and data managers were blinded to treatment conditions and groups. Reliability was evaluated using intraclass correlation coefficients and differences amongst groups were evaluated by mixed-effects models, in accordance with the intention-to-treat approach.This trial was registered with ClinicalTrials.gov Identifier: NCT00123084.FindingsBetween June 23, 2016 and August 14, 2017, blinded listeners transcribed baseline and post-treatment speech samples for intelligibility assessment of 57 patients in the voice (n = 19), articulation (n = 19) and no treatment (n = 19) groups. Between-group differences (d) in changes from baseline to post-treatment in TA indicated significantly greater increases following treatment targeting voice than treatment targeting articulation (d = 26·2%, 95% CI 1·5 − 51·0; p = 0·04; ES=1·0). Differences between TA changes in the treatment targeting voice and in the no treatment group were significant (d = 42·8%, 95% CI 22·4 − 63·2; p = 0·0002; ES=1·8). Differences between TA changes in the treatment targeting articulation and in the no treatment group were not significant (d = 16·5%, 95% CI -6·1 − 39·2; p = 0·147; ES=0·9).InterpretationThese findings provide the first RCT evidence that intensive speech treatment targeting voice improves speech intelligibility in PD. Thus, this evidence-based treatment may positively impact health-related quality of life for patients with PD globally when it is included in patient management.Funding10.13039/100000002NIH—NIDCD R01 DC01150 and LSVT Global, Inc.

GEDI: Gammachirp envelope distortion index for predicting intelligibility of enhanced speech

In this study, we propose a new concept, the gammachirp envelope distortion index (GEDI), based on the signal-to-distortion ratio in the auditory envelope, SDRenv, to predict the intelligibility of speech enhanced by nonlinear algorithms. The objective of GEDI is to calculate the distortion between enhanced and clean-speech representations in the domain of a temporal envelope extracted by the gammachirp auditory filterbank and modulation filterbank. We also extend GEDI with multi-resolution analysis (mr-GEDI) to predict the speech intelligibility of sounds under non-stationary noise conditions. We evaluate GEDI in terms of the speech intelligibility predictions of speech sounds enhanced by a classic spectral subtraction and a Wiener filtering method. The predictions are compared with human results for various signal-to-noise ratio conditions with additive pink and babble noises. The results showed that mr-GEDI predicted the intelligibility curves better than short-time objective intelligibility (STOI) measure, extended-STOI (ESTOI) measure, and hearing-aid speech perception index (HASPI) under pink-noise conditions, and better than HASPI under babble-noise conditions. The mr-GEDI method does not present an overestimation tendency and is considered a more conservative approach than STOI and ESTOI. Therefore, the evaluation with mr-GEDI may provide additional information in the development of speech enhancement algorithms.

A 1.5 mW Programmable Acoustic Signal Processor for Hearing Assistive Devices With Speech Intelligibility Enhancement

This paper presents a low-power, programmable acoustic signal processor for hearing assistive devices with speech intelligibility enhancement. The reprogrammable design provides considerable flexibility for the devices to deal with personal conditions of hearing loss. A spectral-change enhancement (SCE) algorithm is implemented to improve speech intelligibility. The power consumption is minimized by adding dedicated hardware accelerators. The short-time objective intelligibility (STOI) measure is utilized for optimizing the datapath architecture. Optimization on the critical MAC operations results in 34% power and area reductions when compared to the direct-mapped design. A 50% reduction in SRAM storage is also achieved owing to the reduced memory storage for the associated MAC operations. With the aid of the optimized MAC unit and data buffer, the overall execution time is reduced by 99.2%. Designed in a 40-nm CMOS technology, the processor integrates 431k gates in area of 0.3 mm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> . The design dissipates 1.5 mW at a clock frequency of 10.5 MHz from a 0.7V supply, with a processing latency of 1.05 ms.

End-to-End Post-Filter for Speech Separation With Deep Attention Fusion Features

In this article, we propose an end-to-end post-filter method with deep attention fusion features for monaural speaker-independent speech separation. At first, a time-frequency domain speech separation method is applied as the pre-separation stage. The aim of pre-separation stage is to separate the mixture preliminarily. Although this stage can separate the mixture, it still contains the residual interference. In order to enhance the pre-separated speech and improve the separation performance further, the end-to-end post-filter (E2EPF) with deep attention fusion features is proposed. The E2EPF can make full use of the prior knowledge of the pre-separated speech, which contributes to speech separation. It is a fully convolutional speech separation network and uses the waveform as the input features. Firstly, the 1-D convolutional layer is utilized to extract the deep representation features for the mixture and pre-separated signals in the time domain. Secondly, to pay more attention to the outputs of the pre-separation stage, an attention module is applied to acquire deep attention fusion features, which are extracted by computing the similarity between the mixture and the pre-separated speech. These deep attention fusion features are conducive to reduce the interference and enhance the pre-separated speech. Finally, these features are sent to the post-filter to estimate each target signals. Experimental results on the WSJ0-2mix dataset show that the proposed method outperforms the state-of-the-art speech separation method. Compared with the pre-separation method, our proposed method can acquire 64.1%, 60.2%, 25.6% and 7.5% relative improvements in scale-invariant source-to-noise ratio (SI-SNR), the signal-to-distortion ratio (SDR), the perceptual evaluation of speech quality (PESQ) and the short-time objective intelligibility (STOI) measures, respectively.

Listening difficulty estimation model using short-time objective intelligibility measure for outdoor public address systems

Listening difficulty estimation model using short-time objective intelligibility measure for outdoor public address systems