Speech processing coder, decoder and command recognizer

Details Speech processing coder, decoder and command recognizer Speech processing coder, decoder and command recognizer

1997-03-11

1999-06-01

Hudspeth, David R.

Data processing: speech signal processing, linguistics, language

Speech signal processing

For storage or transmission

G10L 302

Patent

active

059096624

DESCRIPTION:

BRIEF SUMMARY
FIELD OF THE INVENTION

The present invention generally relates to a speech processing device which is equipped with both a speech coding/decoding function and a speech recognition function The speech processing device of the present invention can be used as a speech coding/decoding device (CODEC: COder-DECoder) in a digital mobile phone, for example, and can incorporate a speech recognition function into equipment having such a speech coding/decoding device.


BACKGROUND OF THE INVENTION

FIG. 1 is a block diagram of a speech coding/decoding device of the prior art which encodes and decodes speech information by using a codebook.
In a speech coding device, speech input via a speech inputting unit 1 is analyzed by a speech analysis unit 2. Results of the analysis by the speech analysis unit 2 are then quantized based on a codebook 5 by a quantizing unit 3 to generate quantization parameters The quantizing unit 3 further generates more than one type of quantization codes (e.g., indexes of the codebook) which indicate the quantized values (quantized parameters), and supplies these quantization codes of more than one type to a coding unit 4. The coding unit 4 multiplexes the quantization codes to generate encoded codes. Here, the codebook 5 is stored in a ROM.
In a speech decoding device, received encoded codes are separated by a decoding unit 6 into more than one type of quantization codes The separated quantization codes are then subjected to an inverse quantizing process based on the codebook 5 by an inverse-quantizing unit 7 to generate quantization parameters. A speech synthesizing unit 8 synthesizes the speech by using the quantization parameters, so that a speech outputting unit 9 outputs the speech.
Parameters in the codebook 5 used for the quantization process may vary in types, and different processes may be carried out, depending on the types of the parameters, by the speech analysis unit 2, the quantizing unit 3, the inverse-quantizing unit 7, and the speech synthesizing unit 8. For different types of parameters, the speech coding/decoding device may have different configurations as shown in FIG. 2.
FIG. 2 is a table chart showing two different configurations of the vice coding/decoding device of the prior art.
In FIG. 2, the type-1 speech coding/decoding device uses speech waveforms as parameters of the codebook 5. In the speech coding unit, an input speech signal is divided (windowed) into speech signals of a predetermined time length. The quantizing unit then searches in the codebook for speech waveforms closest to the windowed speech signals, and obtains quantization codes of these speech waveforms. In the speech decoding unit, speech waveforms are successively extracted from the codebook by using received quantization codes. The speech waveforms are then interpolated and connected by the speech synthesizing unit to output a speech signal.
The type-2 speech coding/decoding device of type 2 is a device based on a CELP (code excited linear prediction) method, for example, and uses speech-source signals and LPC coefficients as parameters of the codebook. In the type-2 speech coding device, a speech signal is divided (windowed) into speech signals of a predetermined time length, and an LPC analysis is applied. The quantizing unit searches in the codebook for quantized LPC coefficients (quantization parameters) closest to the results of the analysis and for quantization codes indicating the quantization parameters, and, also, searches in the codebook for the most appropriate speech source. In the speech decoding unit, LPC coefficients and speech-source signals are extracted from the codebook by using received quantization codes. The synthesizing unit then synthesizes speech by using the LPC coefficients and the speech-source signals.
In the following, an example of a configuration when the CELP method is used in a speech coding/decoding device will be described.
FIG. 3 is a block diagram of an example of a speech coding device which employs the CELP method. In FIG. 3, the same reference numerals as

REFERENCES:
patent: 4799261 (1989-01-01), Lin et al.
patent: 5305421 (1994-04-01), Li
patent: 5333236 (1994-07-01), Bahl et al.
patent: 5627939 (1997-05-01), Huang et al.
patent: 5633984 (1997-05-01), Aso et al.
patent: 5680506 (1997-10-01), Kroon et al.

Affiliated with

Also associated with

Rating

Speech processing coder, decoder and command recognizer does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Speech processing coder, decoder and command recognizer, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Speech processing coder, decoder and command recognizer will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-962387