Data processing: speech signal processing – linguistics – language – Speech signal processing – For storage or transmission
Reexamination Certificate
1999-12-29
2004-03-30
Dorvil, Richemond (Department: 2654)
Data processing: speech signal processing, linguistics, language
Speech signal processing
For storage or transmission
C704S223000, C341S094000, C340S870030
Reexamination Certificate
active
06714908
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a speech decoder and speech decoding method used in speech CODECs.
BACKGROUND ART
CELP (Code Excited Linear Prediction) has received some attention as a speech coding method.
This CELP is a type of predictive coding wherein samples of the present speech signal are predicted from past decoded results, and the prediction errors which are the differences between the predicted values for the speech signal samples and the actual sample values are coded and sent. Then, in this CELP, a vector quantization is carried out on the prediction errors using with noise sequences and the quantized results are transmitted.
In the decoders for receiving decoding the coded speech signals obtained by a predictive coding method such as CELP and for decoding the received signals, excited signals are generated from the coded speech signals and the internal state of the decoder every standard frame period. Speech signals are decoded from these excited signals, and the internal state can be renewed by these excited signals. Here, when code errors are detected from the frames of the coded speech signals, a concealment process of attenuating the excited signals is performed. This concealment process is explained, for example, in R. Salam et al., “Design and Description of CS-ACELP: A Toll Quality p8 kb/s Speed Coder”, IEEE Trans. on Speech and Audio Processing, vol. 6, no. 2, March 1998.
This concealment process prevents the generation of unpleasant distortions due to frame errors.
However, when coding errors occur in the received frames for a plurality of frames in succession and this concealment process is repeated, the internal state of the decoder will become approximately equal to zero. Thus, subsequently, even if the successive frame errors end, a long time is required in order for correctly received coded speech signals to be decoded at the correct power.
DISCLOSURE OF THE INVENTION
The present invention has been accomplished in consideration of the above situation, and has the object of offering a speech decoder and a speech decoding method in which even when the coded speech signal has successive errors, it is possible to return to a state in which decoding is possible at the originally intended power within an early time upon the coded speech signals being input normally once again.
In order to achieve this object, the present invention provides a speech decoder which receives frames of predictively coded speech signals, generates excited signals from said coded speech signals and an internal state, generates decoded speech signals based on these excited signals while renewing the internal state, and performs a concealment process for attenuating the excited signals when code errors are detected from the frames; the speech decoder comprising amplification control means for, when successive frame errors exceeding a predetermined standard frame error number occur, outputting amplification instruction signals during a predetermined time after the successive frame errors have disappeared, and amplification means for amplifying the excited signals while the amplification instruction signals are being outputted.
According to the present invention, when continuous frame errors occur, and the frame errors subsequently disappear, the internal state of the decoder can rapidly be returned to the state prior to the frame errors to allow decoding at the originally intended power, thus allowing degradations in the subjective speech quality after restoration from the frame errors to be reduced.
REFERENCES:
patent: 4178549 (1979-12-01), Ledenbach et al.
patent: 5664055 (1997-09-01), Kroon
patent: 5699485 (1997-12-01), Shoham
patent: 5732389 (1998-03-01), Kroon et al.
patent: 6085158 (2000-07-01), Naka et al.
patent: 05-113798 (1993-05-01), None
patent: 09-101800 (1997-04-01), None
patent: 09-134198 (1997-05-01), None
IEEE Transactions on Speech and Audio Processing, vol. 6, No. 2, Mar. 1998, Red Salami et al., “Design and Description of CS-ACELP: A Toll Quality 8 kb/s Speech Coder”, pp. 116-130.
Brinks Hofer Gilson & Lione
Dorvil Richemond
Han Qi
NTT Mobile Communications Network Inc.
LandOfFree
Modified concealing device and method for a speech decoder does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Modified concealing device and method for a speech decoder, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Modified concealing device and method for a speech decoder will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3240708