Data processing: speech signal processing – linguistics – language – Speech signal processing – For storage or transmission
Patent
1995-12-14
1999-03-30
Hudspeth, David R.
Data processing: speech signal processing, linguistics, language
Speech signal processing
For storage or transmission
G10L 302, G10L 900
Patent
active
058901043
DESCRIPTION:
BRIEF SUMMARY
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to a method and apparatus for testing telecommunications apparatus.
2. Related Art
In testing telecommunications apparatus (for example, a telephone line, a telephone network, or communications apparatus such as a codec) a test signal is introduced to the input of the telecommunications apparatus, and some test is applied to the resulting output of the apparatus. It is known to derive "objective" test measurements, such as the signal to noise ratio, which can be calculated by automatic processing apparatus. It is also known to apply "subjective" tests, in which a human listener listens to the output of the telecommunications apparatus, and gives an opinion as to the quality of the output.
Some elements of telecommunications systems are linear. Accordingly, it is possible to apply simple artificial test signals, such as discrete frequency sine waves, swept sine signals or chirp signals, random or pseudo random noise signals, or impulses. The output signal can then be analyzed using, for example, Fast Fourier Transform (FFT) or some other spectral analysis technique. One or more such simple test signals are sufficient to characterise the behaviour of a linear system.
On the other hand, modern telecommunications systems include an increasing number of elements which are nonlinear and/or time variant. For example, modern low bit-rate digital speech codecs, forming part of mobile telephone systems, have a nonlinear response and automatic gain controls (AGCs), voice activity detectors (VADs) and associated voice switches, and burst errors contribute time variations to telecommunications systems of which they form part. Accordingly, it is increasingly less possible to use simple test methods developed for linear systems to derive objective measure of the distortion or acceptability of telecommunications apparatus.
Recently in the paper "Measuring the Quality of Audio Devices" by John G. Beerends and Jan A. Stemerdink, presented at the 90th AES Convention, 1991 February 19-22, Paris, printed in AES Preprints as Preprint 3070 (L-8) by the Audio Engineering Society, it has been proposed to measure the quality of a speech codec for digital mobile radio by using, as test signals; a database of real recorded speech and analyzing the corresponding output of the codec using a perceptual analysis method designed to correspond in some aspects to the processes which are thought to occur in the human ear.
It has also been proposed (for example in "Objective Measurement Method for Estimating Speech Quality of Low Bit Rate Speech Coding", Irii, Kurashima, Kitawaki and Itoh, NTT Review, Vol 3. No. 5 September 1991) to use an artificial voice signal (i.e. a signal which is similar in a spectral sense to the human voice, but which does not convey any intelligence) in conjunction with a conventional distortion analysis measure such as the cepstral distance (CD) measure, to measure the performance of telecommunications apparatus.
It would appear obvious, when testing apparatus such as a codec which is designed to encode human speech, and when employing an analysis method based on the human ear, to use real human speech samples as was proposed in the above paper by Beerends and Stemerdink. In fact, however, the performance of such test systems is not particularly good.
One artificial voice test signal is disclosed in CCITT Recommendation P50 (Recommendation on Artificial Voices, Vol. Rec. P50, Melbourne 1988, published by CCITT). In the P50 test signal, there is a randomly selected sequence of 16 predetermined spectral patterns, provided in segments of predetermined length with smooth transitions between the segments. The P50 signal has a long term and short term spectral similarity to speech when averaged over about ten seconds. However, we have found that some differences between the P50 test signal and real speech are significant in testing telecommunications systems.
Another type of artificial test signal which has been proposed utilises spherically in
REFERENCES:
patent: 4352182 (1982-09-01), Billi et al.
patent: 4860360 (1989-08-01), Boggs
patent: 4972484 (1990-11-01), Theile et al.
patent: 5195124 (1993-03-01), Ishioka
patent: 5369644 (1994-11-01), Oliver
patent: 5490234 (1996-02-01), Narayan
British Telecommunications public limited company
Hudspeth David R.
Sax Robert Louis
LandOfFree
Method and apparatus for testing telecommunications equipment us 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 testing telecommunications equipment us, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for testing telecommunications equipment us will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-1225170