Data processing: speech signal processing – linguistics – language – Speech signal processing – For storage or transmission
Reexamination Certificate
1999-12-16
2003-08-19
Abebe, Daniel (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
For storage or transmission
C704S228000, C455S424000
Reexamination Certificate
active
06609092
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to speech processing systems, and more particularly to techniques for determining speech quality in such systems.
BACKGROUND OF THE INVENTION
The most accurate known techniques for evaluating the performance of speech coding systems are subjective speech quality assessment tests such as the well-known mean opinion score (MOS) test. However, these subjective tests are generally costly and time-consuming, and also difficult to reproduce. It is therefore desirable to replace the subjective tests with an objective test for evaluating speech coding performance.
As a result, considerable effort has been devoted to attempting to find a suitable objective distortion measure that will correlate well with subjective MOS measurements. One such objective distortion measure is known as the perceptual speech-quality measure (PSQM), and is described in J. G. Beerends and J. A. Stemerdink, “A perceptual speech-quality measure based on psychoacoustic sound representation,” J. Audio Eng. Soc., Vol. 42, pp. 115-123, March 1994, which is incorporated by reference herein. The PSQM measure has been adopted as the ITU-T standard recommendation P.861 for telephone band speech. See ITU-T Recommendation P.861, Objective Quality Measurement of Telephone-Band (300-3400 Hz) Speech Codecs, Geneva, 1996, which is incorporated by reference herein.
Nonetheless, a number of significant problems remain with PSQM and other conventional objective distortion measures. For example, it has not been determined whether or how such measures can be mapped onto the subjective MOS scale in a database independent manner. In addition, conventional objective measures are in some cases unable to accurately assess the quality of processed speech when the source has been corrupted by environmental noise.
A need therefore exists for improved techniques for predicting the quality of speech and other audio signals, such that a subjective MOS measure or other type of subjective quality measure can be determined accurately and efficiently from a corresponding objective distortion measure, in a manner that is robust in the presence of environmental noise.
SUMMARY OF THE INVENTION
The invention provides methods and apparatus for estimating subjective measures of audio signal quality using objective distortion measures. In accordance with the invention, a mapping function is generated between subjective measures of audio signal quality, e.g., mean opinion score (MOS) measures, degradation MOS (DMOS) measures or other measures, and corresponding objective distortion measures, e.g., auditory speech quality measures (ASQMs), perceptual speech quality measures (PSQMs) or other objective distortion measures, for known audio signals. The audio signals may be speech signals or any other type of audio signals.
The subjective measures and corresponding objective distortion measures are determined in accordance with, e.g., modulated noise reference unit (MNRU) conditions or other suitable distortion conditions placed on the audio signals, and a regression analysis is applied to the results to generate the mapping function. The mapping function may then be utilized, e.g., to evaluate speech quality of additional source speech from a particular speech coding system. In this case, the objective distortion measure is generated using the additional source speech, and the resulting objective measure is applied as an input to the mapping function to generate an estimate of the value of the subjective measure.
Advantageously, the invention allows an objective distortion measure to be mapped in a database-independent manner to a subjective measure, e.g., a MOS or DMOS scale. The mapping function is database independent in that it can be used to generate accurate estimates of subjective measures of speech quality for speech databases unrelated to those used in generating the mapping function. In addition, the objective distortion to subjective quality measure mapping in an illustrative embodiment of the invention provides more accurate prediction than conventional techniques in the presence of environmental noise. The invention may be implemented in numerous and diverse speech and audio signal processing applications, and considerably improves the accuracy of quality prediction in such applications. These and other features and advantages of the present invention will become more apparent from the accompanying drawings and the following detailed description.
REFERENCES:
patent: 4905285 (1990-02-01), Allen et al.
patent: 5621854 (1997-04-01), Hollier
patent: 5794188 (1998-08-01), Hollier
patent: 5987320 (1999-11-01), Bobick
patent: 6205421 (2001-03-01), Morii
ITU-T Recommendation P.861, Objective Quality Measurement of Telephone-Band (300-3400 Hz) Speech Codecs, Geneva, 43 pages, Feb. 1998.
ITU-T Recommendation P.810, Modulated Noise Reference Unit (MNRU), 13 pages, Feb. 1996.
O. Ghitza, “Auditory Nerve Representation as a Basis for Speech Processing,” Advances in Speech Signal Processing, S. Furui and M. M. Sondhi, eds., New York: Marcel Dekker, pp. 453-485, 1992.
D.S. Kim et al., “Auditory Processing of Speech Signals for Robust Speech Recognition in Real-World Noisy Environments,” IEEE Trans. on Speech and Audio Processing, pp. 1-38, Mar. 1998.
O. Ghitza, “Auditory Models and Human Performance in Tasks Related to Speech Coding and Speech Recognition,” IEEE Trans. on Speech and Audio Processing, vol. 2, No. 1, Part II, pp. 115-132, Jan. 1994.
Ghitza Oded
Kim Doh-suk
Kroon Peter
Abebe Daniel
Ryan & Mason & Lewis, LLP
LandOfFree
Method and apparatus for estimating subjective audio signal... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for estimating subjective audio signal..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for estimating subjective audio signal... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3113471