The calculation of most objective speech intelligibility assessment metrics requires clean speech as a reference. Such a requirement may limit the applicability of these metrics in real-world scenarios. To overcome this limitation, we propose a deep learning-based non-intrusive speech intelligibility assessment model, namely STOI-Net. The input and output of STOI-Net are speech spectral features and predicted STOI scores, respectively. The model is formed by the combination of a convolutional neural network and bidirectional long short-term memory (CNN-BLSTM) architecture with a multiplicative attention mechanism. Experimental results show that the STOI score estimated by STOI-Net has a good correlation with the actual STOI score when tested with noisy and enhanced speech utterances. The correlation values are 0.97 and 0.83, respectively, for the seen test condition (the test speakers and noise types are involved in the training set) and the unseen test condition (the test speakers and noise types are not involved in the training set). The results confirm the capability of STOI-Net to accurately predict the STOI scores without referring to clean speech.

The first part of the paper describes the Short Time Objective Intelligibility - STOI algorithm, which makes an objective evaluation of intelligibility, as well as an algorithm for estimating the optimal parameters of the STOI algorithm, N and . Second part of the paper, described an experiment that examines the intelligibility of the sentences formed from the Serbian Matrix Sentence Test - SMST base using: a). subjective MOS (Mean Opinion Score) test and b). objective test STOI algorithm. Subsequently, a comparative analysis of the results were made and it was determinate an analytical formula which connecting optimal pairs (N, ). Using the results of the MOS test as a reference, a mean absolute error of estimation of the STOI algorithm was determined.

In this paper we propose a Deep Neural Network (DNN) based Speech Enhancement (SE) system that is designed to maximize an approximation of the Short-Time Objective Intelligibility (STOI) measure. We formalize an approximate-STOI cost function and derive analytical expressions for the gradients required for DNN training and show that these gradients have desirable properties when used together with gradient based optimization techniques. We show through simulation experiments that the proposed SE system achieves large improvements in estimated speech intelligibility, when tested on matched and unmatched natural noise types, at multiple signal-to-noise ratios. Furthermore, we show that the SE system, when trained using an approximate-STOI cost function performs on par with a system trained with a mean square error cost applied to short-time temporal envelopes. Finally, we show that the proposed SE system performs on par with a traditional DNN based Short-Time Spectral Amplitude (STSA) SE system in terms of estimated speech intelligibility. These results are important because they suggest that traditional DNN based STSA SE systems might be optimal in terms of estimated speech intelligibility.

Speech intelligibility prediction methods have recently gained popularity in the speech processing community as supplements to time consuming and costly listening experiments. Such methods can be used to objectively quantify and compare the advantage of different speech enhancement algorithms, in a way that correlates well with actual speech intelligibility. One such method is the short-time objective intelligibility (STOI) measure. In a recent publication, we proposed a binaural version of the STOI measure, based on a modified version of the equalization cancellation (EC) model. This measure was shown to retain many of the advantageous properties of the STOI measure, while at the same time being able to predict intelligibility correctly in conditions involving both binaural advantage and non-linear signal processing. The biggest prediction errors were found for conditions involving multiple spatially distributed interferers. In this paper, we report results for a new listening experiment including different mixtures of isotropic and point source noise. This exposes that the binaural STOI measure has a tendency to overestimate the intelligibility in conditions with spatially distributed interferes at low signal to noise ratios (SNRs). This condition-dependent error can make it difficult to compare intelligibility across different acoustical conditions. We investigate the cause of this upward bias, and propose a correction which alleviates the problem. The modified method is evaluated with five datasets of measured intelligibility, spanning a wide range of realistic acoustic conditions. Within the tested conditions, the modified method yields very accurate predictions, and entirely alleviates the aforementioned tendency to overestimate intelligibility in conditions with spatially distributed interferers.

Speech enhancement concerns the processes required to remove unwanted background sounds from the target speech to improve its quality and intelligibility. In this paper, a novel approach for single-channel speech enhancement is presented using colored spectrograms. We propose the use of a deep neural network (DNN) architecture adapted from the pix2pix generative adversarial network (GAN) and train it over colored spectrograms of speech to denoise them. After denoising, the colors of spectrograms are translated to magnitudes of short-time Fourier transform (STFT) using a shallow regression neural network. These estimated STFT magnitudes are later combined with the noisy phases to obtain an enhanced speech. The results show an improvement of almost 0.84 points in the perceptual evaluation of speech quality (PESQ) and 1 % in the short-term objective intelligibility (STOI) over the unprocessed noisy data. The gain in quality and intelligibility over the unprocessed signal is almost equal to the gain achieved by the baseline methods used for comparison with the proposed model, but at a much reduced computational cost. The proposed solution offers a comparative PESQ score at almost 10 times reduced computational cost than a similar baseline model that has generated the highest PESQ score trained on grayscaled spectrograms, while it provides only a 1 % deficit in STOI at 28 times reduced computational cost when compared to another baseline system based on convolutional neural network-GAN (CNN-GAN) that produces the most intelligible speech.

To date, the most commonly used outcome measure for assessing ideal binary mask estimation algorithms is based on the difference between the hit rate and the false alarm rate (H-FA). Recently, the error distribution has been shown to substantially affect intelligibility. However, H-FA treats each mask unit independently and does not take into account how errors are distributed. Alternatively, algorithms can be evaluated with the short-time objective intelligibility (STOI) metric using the reconstructed speech. This study investigates the ability of H-FA and STOI to predict intelligibility for binary-masked speech using masks with different error distributions. The results demonstrate the inability of H-FA to predict the behavioral intelligibility and also illustrate the limitations of STOI. Since every estimation algorithm will make errors that are distributed in different ways, performance evaluations should not be made solely on the basis of these metrics.

Although cochlear implants work well for people with hearing impairment in quiet conditions, it is well-known that they are not as effective in noisy environments. Noise reduction algorithms based on machine learning allied with appropriate speech features can be used to address this problem. The purpose of this study is to investigate the importance of acoustic features in such algorithms. Acoustic features are extracted from speech and noise mixtures and used in conjunction with the ideal binary mask to train a deep neural network to estimate masks for speech synthesis to produce enhanced speech. The intelligibility of this speech is objectively measured using metrics such as Short-time Objective Intelligibility (STOI), Hit Rate minus False Alarm Rate (HIT-FA) and Normalized Covariance Measure (NCM) for both simulated normal-hearing and hearing-impaired scenarios. A wide range of existing features is experimentally evaluated, including features that have not been traditionally applied in this application. The results demonstrate that frequency domain features perform best. In particular, Gammatone features performed best for normal hearing over a range of signal-to-noise ratios and noise types (STOI = 0.7826). Mel spectrogram features exhibited the best overall performance for hearing impairment (NCM = 0.7314). There is a stronger correlation between STOI and NCM than HIT-FA and NCM, suggesting that the former is a better predictor of intelligibility for hearing-impaired listeners. The results of this study may be useful in the design of adaptive intelligibility enhancement systems for cochlear implants based on both the noise level and the nature of the noise (stationary or non-stationary).

Underwater acoustic target classification (UATC) has traditionally relied on time–frequency (TF) analysis, with the mel-spectrogram being widely used due to its 2D image-like format and size efficiency. However, many previous UATC approaches have not adequately acknowledged the distinction between natural images and mel-spectrograms. In this paper, we propose a novel approach to transform mel-spectrograms into 3D data and introduce an efficient 3D Spectrogram Network (3DSNet) that treats the time and frequency dimensions separately. Our 3DSNet consists of three key components: the Time–Frequency Separate Convolution (TFSConv) module, Asymmetric Pooling (AsyPool) module, and Channel-Time Attention (CTA) module. We introduce the TFSConv module, which utilizes two convolution operators for the time and frequency dimensions to extract 3D T-F features. This module serves as a lightweight approximation of 3D convolution, effectively reducing approximately 85 percent of parameters. To maintain the inherent distinction between the frequency and time dimensions, we propose an AsyPool module that employs two downsampling strategies. Additionally, we introduce a CTA module to capture more informative and meaningful T-F features. We evaluate our 3DSNet on two publicly available underwater acoustic datasets, and the results demonstrate that our method achieves the optimal balance between performance and model parameters compared to other mainstream methods.

Speech enhancement concerns the processes required to remove unwanted background sounds from the target speech to improve its quality and intelligibility. In this paper, a novel approach for single-channel speech enhancement is presented, using colored spectrograms. We propose the use of a deep neural network (DNN) architecture adapted from the pix2pix generative adversarial network (GAN) and train it over colored spectrograms of speech to denoise them. After denoising, the colors of spectrograms are translated to magnitudes of short-time Fourier transform (STFT) using a shallow regression neural network. These estimated STFT magnitudes are later combined with the noisy phases to obtain an enhanced speech. The results show an improvement of almost 0.84 points in the perceptual evaluation of speech quality (PESQ) and 1% in the short-term objective intelligibility (STOI) over the unprocessed noisy data. The gain in quality and intelligibility over the unprocessed signal is almost equal to the gain achieved by the baseline methods used for comparison with the proposed model, but at a much reduced computational cost. The proposed solution offers a comparative PESQ score at almost 10 times reduced computational cost than a similar baseline model that has generated the highest PESQ score trained on grayscaled spectrograms, while it provides only a 1% deficit in STOI at 28 times reduced computational cost when compared to another baseline system based on convolutional neural network-GAN (CNN-GAN) that produces the most intelligible speech.

Intelligibility listening tests are necessary during development and evaluation of speech processing algorithms, despite the fact that they are expensive and time consuming. In this paper, we propose a monaural intelligibility prediction algorithm, which has the potential of replacing some of these listening tests. The proposed algorithm shows similarities to the short-time objective intelligibility (STOI) algorithm, but works for a larger range of input signals. In contrast to STOI, extended STOI (ESTOI) does not assume mutual independence between frequency bands. ESTOI also incorporates spectral correlation by comparing complete 400ms length spectrograms of the noisy/processed speech and the clean speech signals. As a consequence, ESTOI is also able to accurately predict the intelligibility of speech contaminated by temporally highly modulated noise sources in addition to noisy signals processed with time-frequency weighting. We show that ESTOI can be interpreted in terms of an orthogonal decomposition of short-time spectrograms into intelligibility subspaces, i.e., a ranking of spectrogram features according to their importance to intelligibility. A free MATLAB implementation of the algorithm is available for noncommercial use at http://kom.aau.dk/~jje/.