REW parametric vector quantization and dual-predictive SEW vector quantization for waveform interpolative coding
View Patent ↗An enhanced analysis-by-synthesis waveform interpolative speech coder able to operate at 2.8 kbps. Novel features include dual-predictive analysis-by-synthesis quantization of the slowly-evolving waveform, efficient parametrization of the rapidly-evolving waveform magnitude, and analysis-by-synthesis vector quantization of the rapidly evolving waveform parameter. Subjective quality tests indicate that it exceeds G.723.1 at 5.3 kbps, and of G.723.1 at 6.3 kbps.
1. A computerized system for coding a digital input signal representative of an audio waveform, comprising:
a vector quantizer module comprising an input accepting a vector or rapidly evolving waveform spectra determined form the digital input signal;
a computer readable codebook comprising a plurality of arrays of quantized parametric vectors;
computer readable program instructions for searching said parameter codebook for one of said plurality of arrays of quantized parametric vectors which minimizes distortion; and
computer program instructions for outputting a computer readable indicator that points to the one of said plurality of arrays of quantized parameter vectors which minimizes distortion.
2. The system of claim 1 wherein the indicator determines a quantized vector of the rapidly evolving waveform.
3. The system of claim 2 further comprising a slowly evolving waveform vector quantizer, said slowly evolving waveform quantizer in communication with the vector quantizer for inputting the quantized vector of the rapidly evolving waveform, said slowly evolving waveform quantizer comprising instructions for determining an implied slowly evolving waveform as a function of the quantized rapidly evolving waveform.
4. The system of claim 1 further comprising a synthesis filter coupled to the output of the codebook wherein said synthesis filter outputs a synthesized parameter vector.
5. The system of claim 1 further comprising computer readable instructions for spectrally weighting a plurality of coefficient vectors mapped from synthesized parameter vectors.
6. The system of claim 1 wherein the audio waveform comprises speech.
7. The system of claim 6 , wherein said speech comprises mixed voiced and unvoiced speech segments.
8. The system of claim 1 further comprising a converter that samples an analog input signal to create the digital input signal.
9. The system of claim 1 , wherein the system outputs a coded representation of the digital input signal and the computer readable indicator is embedded in that coded representation.
10. A computer implemented method for coding a digital input signal representative of an audio waveform, comprising:
determining a vector of rapidly evolving waveform spectra from the digital input signal;
searching a computer readable codebook comprising a plurality of arrays of quantized parametric vector for one of said plurality of arrays of quantized parameter vectors which minimizes distortion; and
outputting a computer readable indicator that points to the one of said plurality of arrays of quantized parameter vectors which minimizes said distortion.
11. The method of claim 10 wherein the indicator determines a quantized vector of the rapidly evolving waveform.
12. The method of claim 10 further comprising synthesis filtering the output of the codebook and outputting a synthesized parameter vector.
13. The method of claim 10 further comprising spectrally weighting a plurality of coefficient vectors mapped from synthesized parameter vectors.
14. The method of claim 10 further comprising determining an implied slowly evolving waveform as a function of the quantized rapidly evolving waveform.
15. The method of claim 10 wherein the digital input signal comprises speech.
16. The method of claim 15 , wherein said speech comprises mixed voiced and unvoiced speech segments.
17. The method of claim 10 further comprising sampling an analog input signal to create said digital input signal.
18. The method of claim 10 , wherein the method outputs a coded representation of the digital input signal and the computer readable indicator is embedded in that coded representation.