SPEECH CODER AND SPEECH DECODER

• US 20100256975A1
• Filed: 05/17/2010
• Published: 10/07/2010
• Est. Priority Date: 11/07/1996
• Status: Active Grant

• Current Assignee: Godo Kaisha IP Bridge 1 (IP Bridge, Inc.)
• Original Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Inventors: EHARA, Hiroyuki, YASUNAGA, Kazutoshi, MORII, Toshiyuki
• Granted Patent: US 8,036,887 B2
• Time in Patent Office: Days
• US Class Current: 704/219
• CPC Class Codes: G10L 19/12 the excitation function being a code excitation, e.g. in code excited linear prediction [CELP] vocoders
• CPC Class Codes: G10L 19/135 Vector sum excited linear prediction [VSELP]
• CPC Class Codes: G10L 2019/0007 Codebook element generation
• CPC Class Codes: G10L 2019/0013 Codebook search algorithms

Abstract

A CELP speech decoder includes an adaptive codebook that generates an adaptive code vector and a random codebook that generates a random code vector. The random codebook includes an input vector provider that provides an input vector including at least one pulse, each pulse having a position and a polarity, a fixed waveform storage that stores at least one fixed waveform, and a selector that selects at least one of a first process and a second process based on a value of an adaptive codebook gain. The random codebook further includes a convolution section that generates the random code vector by convoluting the at least one fixed waveform with the input vector when the first process is selected. A synthesis filter outputs synthesized speech by performing linear prediction coefficient synthesis on a signal based on the adaptive code vector and the random code vector.

Claims

1. A CELP speech decoder, comprising:

an adaptive codebook configured to generate an adaptive code vector;

a random codebook configured to generate a random code vector, the random codebook comprising;

an input vector provider configured to provide an input vector including at least one pulse, each pulse having a position and a polarity,

a fixed waveform storage configured to store at least one fixed waveform,

a selector configured to select at least one of a first process and a second process based on a value of an adaptive codebook gain, and

a convolution section configured to generate the random code vector by convoluting the at least one fixed waveform with the input vector when the first process is selected; and

a synthesis filter configured to output synthesized speech by performing linear prediction coefficient synthesis on a signal based on the adaptive code vector and the random code vector.

2. The CELP speech decoder according to claim 1,

wherein the selector is configured to select the input vector as the random code vector when the second process is selected.

3. The CELP speech decoder according to claim 1,

wherein the convolution section is configured to generate the random code vector by linear convoluting.

4. The CELP speech decoder according to claim 1, wherein the first process includes a modification process of the input vector, and the second process includes a no-modification process of the input vector.

5. The CELP speech decoder according to claim 1, wherein the selector is configured to select the first process if the value of the adaptive codebook gain is low, and configured to select the second process if the value of the adaptive codebook gain is high.

6. A CELP speech decoder, comprising:

an adaptive codebook configured to generate an adaptive code vector;

a random codebook configured to generate a random code vector, the random codebook comprising;

an input vector provider configured to provide an input vector including at least one pulse, each pulse having a position and a polarity,

a fixed waveform storage configured to store a first fixed waveform and a second fixed waveform,

a selector configured to select at least one of the first fixed waveform and the second fixed waveform if a value of an adaptive codebook gain is low, and

a convolution section configured to generate the random code vector by convoluting at least one of the selected first fixed waveform and the selected second fixed waveform with the input vector if the value of the adaptive codebook gain is low; and

a synthesis filter configured to output synthesized speech by performing linear prediction coefficient synthesis on a signal based on the adaptive code vector and the random code vector.

7. The CELP speech decoder according to claim 6,

wherein the selector is configured to select the input vector as the random code vector if the value of the adaptive codebook gain is high.

8. The CELP speech decoder according to claim 6,

wherein the convolution section is configured to generate the random code vector by linear convoluting.