Arkadiy Prodeus

National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" · Acoustic and Multimedia Electronic Systems

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...

The potential possibilities of the previously proposed method for detecting early reflections, which consists in estimating the excess kurtosis at short intervals of the room impulse response, have been studied. The dependences of detection probability and false alarm probability on segment length were obtained by means of computer simulation. In o...

In this paper, correlation coefficients between the five objective estimates of speech quality, on the one hand, and the Speech Transmission Index as speech intelligibility measure, on the other hand, were estimated. This comparison was performed using binaural room impulse responses corresponded to different points of the three university auditori...

In this paper, five objective measures of the quality of speech signals distorted by reverberation are compared with the Speech Transmission Index (STI). The main aim of the comparison is to further test and explain the reasons for the previously discovered phenomenon of an increase in the speech quality and intelligibility with increasing room siz...

Voice control of an unmanned aerial vehicle has a number of advantages if the operator is indoors. In this case, the distortions of speech commands caused by the influence of noise interference can be significantly reduced. However, the disadvantage of such control is the negative impact of reverberation on speech intelligibility. Therefore, it is...

Estimates of speech quality and intelligibility for three university classrooms of small, medium and large sizes are presented. The quality and intelligibility of speech were assessed by objective methods using binaural room impulse responses, measured at 5-6 points of the premises. The measures of speech quality were log-spectral distortion (LSD),...

The object of this paper was to study the statistical properties of the formant-modulation method for assessing speech intelligibility. Detailed description of the proposed formant-modulation method is presented. Analytical expressions for expectation and variance of estimators of a modulation coefficient and effective signal-to-noise ratio are obt...

The advantage of speech control of unmanned aerial vehicles, which occurs indoors, is the ability to reduce speech distortions caused by noise action. However, the disadvantage of such control is the negative impact of reverberation on speech intelligibility. Therefore, it is advisable to perform a preliminary assessment of speech intelligibility i...

A new method of automatic detection of early reflections in room impulse responses is studied. This method consists in estimating the excess coefficient within the impulse response segments into which they are divided, and in comparing these estimates with a certain threshold value. The threshold value choice is based on a predetermined probability...

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 designi...

In this paper, the fundamental possibility of using the excess coefficient to detect discrete bursts in the binaural room impulse response (BRIR) is investigated. For this purpose, the estimates of IX university auditoriums of small, medium and large size were used as initial data. It is shown that the advantage of estimates of the excess coefficie...

It is shown that fourth standardized moment (kurtosis) and its some functional transformations (inverse value, square root of inverse value) can be objective measures of clipping and quality of speech and music signals. The essential advantages of suggested measures are no need for previous estimation of probability density of analyzed signal as we...

The articulation tests of noised and reverberated speech were carried out under different listening modes: (1) diotic speech presentation through headphones, (2) diotic speech presentation through computer speakers, and (3) dichotic speech presentation through headphones. Developed software toolkit was used for automation of subjective assessment o...

Experimental studies of the use of the developed hardware and software tool “Artificial Head” for two-channel estimation of the intelligibility of the speech distorted by reverberation have been performed in this paper. The peculiarity of this complex is that it contains electro-acoustic equipment of different quality, including household appliance...

Виконано експериментальні дослідження можливості використання розробленого апаратно-програмного комплексу «Штучна голова» для двоканального оцінювання розбірливості мови, спотвореної реверберацією. На першому етапі такого оцінювання здійснюють запис відгуку приміщення на тестовий сигнал у вигляді mls-послідовності. На другому етапі оцінюють імпульс...

In this paper, experimental evaluation of speech intelligibility in small (177 m 3) and medium-sized (270 m 3) classrooms is presented. These studies are based on the assumption that the harmful effect of noise is negligible compared to that for reverberation. Speech intelligibility was computed from pre-measured room impulse response by a modulati...

Reverberation can be considered as a significant interference if the voice control of the unmanned aerial vehicle is performed by an indoor operator. In this paper, two simplified models of early sound reflections in a room are considered. The first model is a single reflection at a time interval of 0-50 ms. The second model is a set of reflections...

Reverberation can be a significant interference if the voice control of the unmanned aerial vehicles is performed by an indoor operator. In this paper, two simplified models of early sound reflections in a room are considered. The first model is a single reflection at a time interval of 0-50 ms. The second model is a set of reflections randomly dis...

This paper compares the results of subjective and objective assessments of the quality of speech and music signals distorted during clipping when large instantaneous signal values are replaced by a certain threshold constant or by values close to it. It was proposed in recent works to use kurtosis and some of its simple functional transforms such a...

(54) METHOD OF DETECTING CLIPPING OF SPEECH AND MUSIC SIGNALS (57) Annotation: The method of detecting the clipping of speech and music signals includes comparing the parameter extracted from the signal. The cumulative coefficient of the statistical distribution of the instantaneous values of the acoustic signal or the functional transformation of...

In this paper, the technology of correction of the frequency response of the measuring path of the hardwaresoftware complex "Artificial head", intended for acoustic examination of the rooms, has been developed. This correction isnecessary because the amplitude-frequency response of the loudspeaker and the microphone are not perfectly uniform in the...

Розроблено технологію коригування частотної характеристики вимірювального тракту апаратно-програмного комплексу «Штучна голова», призначеного для акустичної експертизи приміщень. Показано, що таке коригування може бути виконано шляхом контрольованого ділення частотної характеристики системи «гучномовець-приміщення-мікрофон» на попередньо отриману о...

In this paper, an assessment of the combined effects of the noise, early and late reflections on speech (monosyllables of consonance-vowel-consonance type) intelligibility for binaural listening mode was made. Test signals were synthesized in two ways. In the first case, the noisy monosyllables (signal-to-noise ratio was varied from ?15 dB to +5 dB...

Specifics of acoustic signal processing prior partial signal-to-noise ratios estimating when measuring speech intelligibility under noise dominance was studied. A pre-processing algorithm was proposed and tested using real signals. The speech intelligibility was evaluated using three versions of the formant method as well as the speech transmission...

It is shown that the kurtosis and the normalized variance can be used as a measures of the clipping value of speech signals. The use of the proposed measures makes it possible to significantly simplify and speed up the clipping value calculations compare to the methods where preliminarily estimation of the probability density function of the analyz...

A detailed description of the speech intelligibility prediction algorithm using analytical modeling is presented. The efficiency of the proposed algorithm is tested for four types of noise interference: white, pink, brown and typical for classrooms. The consistency of the results with known similar results indicates the correctness of the proposed...

When transmitting music signals via infocommunication channels, there is a risk of distortion of these signals due to clipping. The consequence of these distortions is a decrease in subjective assessments of the quality of musical signals, so clipping detection is an important issue. In this paper, the kurtosis and the inverse value of kurtosis are...

In this paper, the possibility of constructing a simple, in technical implementation, classification system based on the use of spectral features is shown. The efficiency of this system is demonstrated by an example of an analysis of the behaviour a bee colonies during honey collection using characteristics of noise inside the hive is explored. Ana...

In this paper, it is shown that kurtosis and some of its functional transformations are expedient to use as a measures of the clipping value of speech signals. In addition to the kurtosis, two more measures of the clipping value are considered: they are the reciprocal of the kurtosis, as well as the square root of the reciprocal of the kurtosis. Th...

It is shown the principle possibility of using the fourth order moment, as well as the simple functional transformation of the fourth order moment, as measures of the quality of clipped speech signals. Averaged dependences of subjective and objective assessments of the quality of speech signals was constructed, and maps of the correspondence of the...

The possibilities of two variants of the classification system designed to recognize the state of a bee family, namely the state of a honey gathering and the state of complete cessation of the honey gathering, on the characteristics of its noise inside the hive have been experimentally investigated. It is shown that the power spectrum density estim...

The possibilities of two variants of the classification system designed to recognize the state of a bee family, namely the state of a honey gathering and the state of complete cessation of the honey gathering, on the characteristics of its noise inside the hive have been experimentally investigated. It is shown that the power spectrum density estim...

A room is a special kind of filter that affects speech intelligibility in two ways. Late reverberation, like noise, reduces speech intelligibility, while early reflections increase speech intelligibility. The influence of the listening mode on the intelligibility of noised speech is also known. Binaural listening is preferable to monaural one becau...

Clipping of speech leads to the appearance of higher orders harmonics and, as a result, to reducing of accuracy of automatic speech recognition systems used as kind of artificial intellect of aircraft control systems and flight control systems for unmanned aerial vehicles. In this paper, the subjective and objective estimates of clipped speech qual...

Comparison of the behavior of subjective assessments of quality and intelligibility of speech distorted by white, pink and brown noise is performed. The results obtained indicate the similarity of this behavior in the range of signal-to-noise (SNR) values 0<SNR<5 dB, with the best quality and intelligibility for brown noise, and the worst for pink...

Clipping of speech leads to the appearance of higher orders harmonics and, as a result, to reducing of accuracy of automatic speech recognition systems used as kind of artificial intellect of aircraft control systems and flight control systems for unmanned aerial vehicles. In this paper, the subjective and objective estimates of clipped speech qual...

Представлены результаты оценивания, с применением объективных и субъективных мер, качества музыкальных сигналов. Субъективное оценивание осуществлялось 23 слушателями средним возрастом 22 года, без недостатков слуха. Для объективного оценивания использованы 4 меры качества, среди которых сегментное отношение сигнал-шум, лог-спектральные искажения,...

In this paper, the results of automated subjective assessment of Ukrainian speech intelligibility are presented. Speech monosyllables of the consonant-vowel-consonant (CVC) type were listened in two modes: through headphones and through acoustic monitors. The assessment was carried out with the help of specially developed software that allowed auto...

In this paper, a prototype of an automated system for subjective evaluation of intelligibility of the Ukrainian speech in communication channels is proposed. The structure of this system includes nine articulation tables of speech monosyllables of consonant-vowel-consonant (CVC) type and a set of computer programs that allow automating most of the...

In this paper, the results of speech intelligibility subjective assessment of Ukrainian speech monosyllabic sound combinations against background noise and reverberation through articulation tests are presented. The evaluation was carried out with the help of specially developed software that allowed automating and thus greatly facilitating and acc...

In this paper, the results of experimental studies of elements of the automated system of subjective evaluation of speech intelligibility are presented. Special set of proposed computer programs and diagnostic articulation tables of Ukrainian speech monosyllables of the CVC type is a core of the automated system. The set of computer programs allows...

В даній роботі наведено результати оцінювання впливу стаціонарних та нестаціонарних синтезованих шумів на якість та розбірливість мовних сигналів. Для випадку стаціонарних шумів показано, що при малих відношеннях сигнал-шум білий шум поступається за маскувальною здатністю рожевому й коричневому шумам. Досліджено два простих, з обчислювальної точки...

In this paper, the results of experimental studies of the reasons for the low reliability of the log-spectral distortion (LSD) measure for estimating the quality of speech signals limited in the frequency band are presented. It is shown that the LSD measure has an increased sensitivity not only to the speech formants, but also to the shape of frequ...

The features of simple, from a computational point of view, quality measures of speech and music signals, such as segment signal-to-noise ratio and log-spectral distortions are investigated. The proposals for increasing the reliability of assessments of these measures have been clarified.

In this paper, the results of quality and intelligibility assessment of speech masked by stationary and nonstationary noise have been proposed. Subjective speech quality assessment technique has been used to show that white noise masking ability is lower than one for pink and even for brown noise when SNR is less than 0 dB. Two algorithms of nonsta...

Development of automatic speech recognition (ASR) systems robust to late reverberation action is urgent task. It is well known that a late reverberation reduction algorithm used as ASR pre-processor demands prior estimation of reverberation time. Blind reverberation time measurements are less accurate than ones for known room impulse response (RIR)...

In this paper, subjective and objective estimators of the quality of speech and music signals subjected to phase distortion are compared, and mapping between objective and subjective quality estimates is realized. It was found that the phase distortion of speech signals is perceived stronger than ones for musical signals. Two types of phase distort...

В монографии рассмотрены субъективные и объективные (инструментальные) методы оценивания качества речевых и музыкальных сигналов в помещениях и каналах связи. При этом значительное внимание уделено проблемам надежности и автоматизации оценивания объективных мер качества и разборчивости речи. Вопросы коррекции речевых сигналов, искаженных шумом и ре...

Сопоставлены шесть алгоритмов шумоподавления с использованием объективных показателей качества речевого сигнала, а также с использованием сквозного показателя качества системы автоматического распознавания речи в виде точности распознавания речи. Показано, что алгоритмы радикального шумоподавления уступают традиционным алгоритмам шумоподавления как...

Recent preliminary studies have shown that phase distortion of music signals are acceptable for human auditory system when the maximum difference of group delay time in high and low frequencies is less than 70 ms. This value is less than 50 ms for speech signals. In this paper, we obtain a more accurate subjective speech quality assessment and its...

A form of classrooms acoustic passport handy both school administrators and architects involved in the reconstruction of the rooms was developed. Produced acoustic certification of seven classrooms NTU "KPI" revealed the shortcomings of these audiences and demonstrated the convenience of the developed form. Ref. 16, figure 1, table 1.

Evaluation of speech quality for signals passed through low-pass filter showed that log-spectral distortion (LSD) does not decrease monotonically when filter bandwidth is broadening. Experiments have shown that increasing of the analysis time can not compensate for this deficiency. It was suggested that a possible reason for the observed phenomenon...

The coefficients of correlation and matching maps of the results of objective and subjective assess-ment of the speech quality for signals distorted due to the bandwidth limitations had been evaluated. It is shown that matching maps are much more useful in comparison with correlation coefficients because, firstly, they perform the function of the c...

In this paper, the results of intelligibility assessment of speech masked by synthetic noises have been proposed. Articulation technique has been used for the assessment. It is shown that white noise masking ability is lower than one for pink and even for brown noise when SNR is less than 0 dB. Two algorithms of speech-like noise have been proposed...

There are compared six noise suppression algorithms with application of objective factors of the speech signal quality, and also with application of through quality factor of the system of automated speech recognition in form of speech recognition accuracy. It is shown that radical noise suppression algorithms are worse than traditional noise suppr...

In this paper, two techniques of automatic speech recognition (ASR) system training on noised speech are compared with technique of training on clean speech. The comparing has been made by means of speech recognition accuracy measure, with usage of fourteen kinds of noise. These were noises of household appliances and computers, street and transpor...

In this paper, two techniques of automatic speech recognition system training on noised speech are compared with technique of training on clean speech. The comparing has been made by means of speech recognition accuracy measure, with usage of fourteen kinds of noise. These were noises of household appliances and computers, street and transport, tea...

004.934 И.В. Котвицкий , А.Н. Продеус, д.-р.техн.наук Национальный технический университет Украины «Киевский политехнический институт», ул. Политехническая, 16, корпус 12, г. Киев, 03056, Украина. Объективное и субъективное оценивание качества речевых и музыкальных сигналов, подвергнутых фазовым искажениям Недавние предварительные исследования пока...

534.6 Ю.С. Костючок, Л.С. Мартинович, Д.Е. Моторнюк, В.А. Нечитайло, А.В. Храпачевский, А.Н. Продеус, д.-р.техн.наук Национальный технический университет Украины «Киевский политехнический институт», ул. Политехническая, 16, корпус 12, г. Киев, 03056, Украина Акустическая паспортизация учебных помещений Разработана форма акустического паспорта, удоб...

Six noise reduction algorithms were compared by means of nine objective speech quality measures and speech recognition accuracy (Acc%). Negative consequences of excessive noise reduction for automatic speech recognition was demonstrated. Study of quality measures matching showed that only log-likelihood ratio and signal composite index were in good...

Noise and late reverberation reduction algorithms were studied using objective speech quality and speech recognition accuracy (Acc%) measures. Negative consequences of excessive noise and late reverberation reduction for automatic speech recognition had been demonstrated. Study of speech quality measures showed that only few of them were in good ag...

Путем вычисления коэффициентов корреляции и построения карт соответствия, сопоставлены результаты объективного и субъективного оценивания качества речевых сигналов, искаженных из-за ограничения полосы частот. Показано, что карты соответствия, дополненные информацией о способе их построения, более полезны в прикладном плане, нежели коэффициенты корр...

Effect of "decision-directed " , maximum likelihood and "rough " a priori signal-to-noise ratio (SNR) assessment methods and their parameters on the noise reduction algorithms quality had been considered. It was shown that "rough " assessment method which doesn't contain averaging procedure is optimal in terms of recognition accuracy (Acc%) for SNR...

Noise and late reverberation reduction algorithms were compared by means of objective speech quality and speech recognition accuracy (Acc%) measures. Negative effects of excessive noise reduction for automatic speech recognition (ASR) had been shown. It was found possibility of improvement the noise suppression algorithms quality, in terms of Acc%,...

Pattern transformations or operations aimed at increasing system robustness, e.g. against channel noise or different working conditions

С целью анализа влияния способа и параметров оценки априорного отношения сигнал-шум (SNR) на качество алгоритмов шумоподавления, сопоставлены три способа оценивания априорного SNR. Показано, что оценка методом «управление решением» позволяет обеспечить максимально высокое качество звучания речевого сигнала. «Грубая» оценка при интегральном SNR>15 д...

С целью анализа влияния способа и параметров оценки априорного отношения сигнал-шум (SNR) на качество алгоритмов шумоподавления, сопоставлены три способа оценивания априорного SNR. Показано, что оценка методом «управление решением» позволяет обеспечить максимально высокое качество звучания речевого сигнала. «Грубая» оценка при интегральном SNR>15 д...

Сопоставлены три метода оценивания априорного SNR: «управление решением», максимального правдоподобия и «грубая» оценка. Выработаны рекомендации по выбору параметров усреднения при использовании данных методов. Показано, что при шумоподавлении в системе автоматического распознавания речи метод «управление решением» позволяет обеспечить наилучшие ре...

In this paper, six noise reduction algorithms had been compared with the use of a set of indicators. Among them are popular noise reduction algorithms such as spectral subtraction, Wiener filtering, MMSE and logMMSE, and two less well-known Wiener-TSNR and Wiener-HRNR algorithms. It is shown that when the noise reduction system is used as preproces...

Показано, что для слуховой системы человека приемлемыми являются фазовые искажения речевых и музыкальных сигналов, если максимальная разница групповых времен задержки в области высоких и низких частот не превышает 50-70 мс

A new method of classification of a speaker's gender based on cumulant coefficients is proposed. The effect of an additive noise and measurement error of classification signs on accuracy of classification is analyzed. The expediency of construction of an adaptive system of classification operating with considering of masking of a speech signal by n...

In this paper, quantitative assessment of influence of the time-alignment error on segmental signal-to-noise ratio (SSNR) estimation is made.It is shown that an effective way to reduce sensitivity of SSNR estimator to time-alignment error is to increase the sample rate of the compared signals in 2...4 times by means of their interpolation. It was f...

In this paper, traditional noise reduction algorithms such as spectral subtraction, Wiener, MMSE and logMMSE filtering algorithms, and two less known Wiener-TSNR and Wiener-HRNR filtering algorithms had been compared with the use of a set of quality measures. It is found that excessive noise reduction leads to insignificant degradation of the speec...

Correction of speech signals distorted by reverberation is topical in building communications systems, automatic speech recognition systems, and hearing aids. The late reverberation suppression by the spectral subtraction method or the frequency correction method involves the need of estimating the late reverberation spectrum. Though the procedure...

In this paper, five noise reduction algorithms with usage of nine quality measures were studied and potential possibilities of the measures were studied at the same time.

Refined recommendations for choosing optimal, in the sense of automatic speech recognition (ASR) accuracy maximum, parameters of the late reverberation suppression technique, have been proposed in this paper. It was shown that best value of boundary between early reflections and late reverberation approximates to 100 ms for ASR systems. It was show...

Optimal, in the sense of automatic speech recognition (ASR) accuracy maximum, parameters of the late reverberation suppression technique have been proposed in this paper. It was shown that the value 50 ms as boundary between early reflections and late reverberation, which usually is used when problems of speech quality and intelligibility is studie...

In this paper, subjective and objective estimators of the quality of speech and music signals subjected to phase distortion are compared, and mapping between objective and subjective quality estimates is realized. It was found that the phase distortion of speech signals is perceived stronger than ones for musical signals. Two types of phase distort...

Dependence of objective quality evaluation of speech band-limited signals is experimentally obtained. As part of this task, a comparison of the considered indicators of the speech quality had been made. It is shown that computationally simple indicators, such as segmental SNR (SSNR) and log-spectral distortion (LSD), may not adequately respond to c...

Boundary values between early reflections and late reverberation, optimal in sense of such criteria as speech recognition accuracy and speech quality, had been found. When optimal boundary value is chosen, usage of logMMSE method for late reverberation suppression makes it possible to increase recognition accuracy from 22 ... 30% to 56…74% and spee...

При построении систем автоматического распознавания речи актуальной является задача коррекции речевых сигналов, искаженных реверберацией, для решения которой необходимо предварительно измерить время реверберации. Слепые измерения времени реверберации менее эффективны, в смысле точности распознавания речи, по сравнению с прямыми измерениями, однако...

Enhancement of speech distorted by reverberation is issue of the day. Before suppression of late reverberation by spectral subtraction or frequency correction techniques, it is necessary to estimate the spectrum of the late reverberation. Recommendations for optimizing of late reverberation spectrum estimation had been proposed in the paper.

Dependence of dereverberation systems quality indicators on speech signal distortion degree. Using the experimental methods, a comparative analysis of cross-sectional and intermediate indicators of the quality of the systems for reducing the reverberation noise, which are used as preprocessors in automatic speech recognition systems, was performed...

Enhancement of speech distorted by reverberation is issue of the day. The problem has been actively studied in the last decade. However, it is still extremely difficult to find clear recommendations on choice of boundary value between early reflections and late reverberation, optimal in sense of such criteria as speech recognition accuracy and spee...

Efficiency of protection constructions is usually estimated by criterion of “signal-to-noise ratio” at the reception point. In this paper it is proposed to use criterion of speech intelligibility, which is more suitable from viewpoint of finite user. Represented results of computer simulation show constructiveness of proposed approach.