Quality assessment of speech coding and speech synthesis systems

Kitawaki, Nobuhiko; Nagabuchi, Hiromi

doi:10.1109/35.7665

Cited by 24 publications

(7 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The equation states that 2 is the sum of unexplained variance in the estimation model, MOS estimation error due to limited number of listeners, and bias error between subjective and objective MOSs. In comparing estimation algorithms using the same databases, σ 2 MOS is an irreducible noise term affecting all the algorithms equally. Tables 6 and 7 show that PESQ produces large values on databases Exp1D, Exp3A, and Exp3D, even though R is quite high for databases Exp1D and Exp3D.…”

Section: Database Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Objective Speech Quality Measurement Using Statistical Data Mining

Zha

Chan

2005

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

Measuring speech quality by machines overcomes two major drawbacks of subjective listening tests, their low speed and high cost. Real-time, accurate, and economical objective measurement of speech quality opens up a wide range of applications that cannot be supported with subjective listening tests. In this paper, we propose a statistical data mining approach to design objective speech quality measurement algorithms. A large pool of perceptual distortion features is extracted from the speech signal. We examine using classification and regression trees (CART) and multivariate adaptive regression splines (MARS), separately and jointly, to select the most salient features from the pool, and to construct good estimators of subjective listening quality based on the selected features. We show designs that use perceptually significant features and outperform the state-of-the-art objective measurement algorithm. The designed algorithms are computationally simple, making them suitable for real-time implementation. The proposed design method is scalable with the amount of learning data; thus, performance can be improved with more offline or online training

show abstract

Section: Database Resultsmentioning

confidence: 99%

“…Online objective measurement is the only viable means of measuring voice quality, for the purpose of real-time call monitoring and control, on a networkwide scale. Other applications of voice quality measurement include evaluation of disordered speech [1] and synthesized speech [2].…”

Section: Introductionmentioning

confidence: 99%

Objective Speech Quality Measurement Using Statistical Data Mining

Zha

Chan

2005

EURASIP J. Adv. Signal Process.

View full text Add to dashboard Cite

show abstract

“…Indeed, people can forget their responses to preserve their self-image or they may say what is considered more socially acceptable (Crosby et al 1980;DeMaio 1984;Edwards 1957;Maass et al 2000;Marlow and Crowne 1961;Paulhus 1984;Roccato and Zogmaister 2010). Moreover, subjects' answers can be distorted unintentionally due to the lack of introspection abilities (Banaji 2001;Boca 1996;Greenwald and Banaji 1995;Kitawaki and Nagabuchi 1998) and the difficulty of verbalizing emotions (Penn 2006). For these reasons, direct techniques are suitable to measure emotions that are definite and conscious and that the subject intends to report truthfully to the researcher (Jacoby et al 1992;Greenwald and Banaji 1995).…”

Section: Neuro and Psychophysiological Signalsmentioning

confidence: 99%

“The Theater of the Mind”: The Effect of Radio Exposure on TV Advertising

et al. 2020

View full text Add to dashboard Cite

Contemporary society requires communication strategies that integrate different media channels in order to improve advertising performance. Currently, there are not many scientific research studies of the various mass media, comparing the results of audiovisual advertising to purely audio or visual messages aimed at detecting the best combination of media, especially from a neurophysiological perspective. This study aims to investigate the effects of previous exposure to an advertisement via radio on the consumers’ response to the same advertisement shown on television (TV) or as a banner on a website. A total of seventy participants in a between-subjects experiment watched several television commercials during the advertising break of a documentary or saw some banners during a web surfing task. Half were first exposed to the same advertisements via radio. The results have shown that participants who previously listened to the radio advertisements spent a longer time looking at the brand and had a higher engagement when watching the same advertisements on television. Moreover, they had a different kind of visual attention to the website banners. This pattern of results indicates the effect of mere exposure—that is, the exposure to a radio advertisement enhances the effectiveness of the same advertisement via television or web, offering useful insights for media planning campaigns. Even if mere exposure has been extensively studied, cross-media research is scarcely explored, whereas this study detected the effects of mere exposure in a cross-media communication strategy, showing that it can be measured through psychophysiological methods.

show abstract

“…In recent years, speech quality assessment has gained increasing importance in multi-media and telecommunication applications, as well as performance evaluation of speech coders [1], text-to-speech synthesis [2] and automatic speech recognition (ASR) [3]. Speech quality can be defined as the subjective opinion of a given speech signal, as determined by the listener [4].…”

Section: Introductionmentioning

confidence: 99%

Non-Intrusive POLQA Estimation of Speech Quality using Recurrent Neural Networks

Sharma

Hogg

Wang

et al. 2019

2019 27th European Signal Processing Conference (EUSIPCO)

View full text Add to dashboard Cite

Estimating the quality of speech without the use of a clean reference signal is a challenging problem, in part due to the time and expense required to collect sufficient training data for modern machine learning algorithms. We present a novel, non-intrusive estimator that exploits recurrent neural network architectures to predict the intrusive POLQA score of a speech signal in a short time context. The predictor is based on a novel compressed representation of modulation domain features, used in conjunction with static MFCC features. We show that the proposed method can reliably predict POLQA with a 300 ms context, achieving a mean absolute error of 0.21 on unseen data. The proposed method is trained using English speech and is shown to generalize well across unseen languages. The neural network also jointly estimates the mean voice activity detection (VAD) with an F1 accuracy score of 0.9, removing the need for an external VAD.

show abstract

Quality assessment of speech coding and speech synthesis systems

Cited by 24 publications

References 9 publications

Objective Speech Quality Measurement Using Statistical Data Mining

Objective Speech Quality Measurement Using Statistical Data Mining

“The Theater of the Mind”: The Effect of Radio Exposure on TV Advertising

Non-Intrusive POLQA Estimation of Speech Quality using Recurrent Neural Networks

Contact Info

Product

Resources

About