Adaptive postfiltering methods and systems for decoding speech

• US 20030088405A1
• Filed: 08/09/2002
• Published: 05/08/2003
• Est. Priority Date: 10/03/2001
• Status: Active Grant

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Broadcom Corporation (Broadcom, Inc.)
• Inventors: Chen, Juin-Hwey, Thyssen, Jes, Lee, Chris C.
• Granted Patent: US 8,032,363 B2
• Time in Patent Office: Days
• US Class Current: 704/216
• CPC Class Codes: G10L 19/26 Pre-filtering or post-filtering

Abstract

A method of processing a decoded speech (DS) signal including successive DS frames, each DS frame including DS samples. The method comprises: adaptively filtering the DS signal to produce a filtered signal; gain-scaling the filtered signal with an adaptive gain updated once a DS frame, thereby producing a gain-scaled signal; and performing a smoothing operation to smooth possible waveform discontinuities in the gain-scaled signal.

Claims

1. A method of processing a decoded speech (DS) signal, the DS signal including successive DS frames, each DS frame including DS samples, comprising:

(a) adaptively filtering the DS signal to produce a filtered signal;

(b) gain-scaling the filtered signal with an adaptive gain updated once a DS frame, thereby producing a gain-scaled signal; and

(c) performing a smoothing operation to smooth possible waveform discontinuities in the gain-scaled signal.

2. The method of claim 1, wherein said filtering step (a) comprises at least one of long-term filtering and short-term filtering.

3. The method of claim 2, wherein said filtering step (a) further comprises at least one of

long-term filtering followed by short-term filtering, and

short-term filtering followed by long-term filtering.

4. The method of claim 1, wherein the gain-scaled signal includes successive gain-scaled signal frames corresponding to the successive DS frames, and step (c) comprises performing the smoothing operation to smooth possible waveform discontinuities between adjacent gain-scaled signal frames.

5. The method of claim 4, wherein step (c) further comprises performing an overlap-add operation based on the adjacent gain-scaled signal frames.

6. The method of claim 4, wherein:

step (a) includes filtering a beginning portion of a current DS frame using a current set of filter coefficients, to produce a first filtered frame portion;

step (b) includes gain-scaling the first filtered frame portion using a current gain, to produce a first gain-scaled frame portion; and

the smoothing operation of step (c) includes

filtering and gain-scaling the beginning portion of the current DS frame using a previous set of filter coefficients and a previous gain, respectively, to produce a second gain-scaled portion, and

modifying the first gain-scaled frame portion with the second gain-scaled frame portion so as to smooth a possible waveform discontinuity.

7. The method of claim 1, wherein said filtering step (a) comprises short-term filtering using one of an all-zero filter, an all-pole filter, and a pole-zero filter.

8. The method of claim 1, wherein said filtering step (a) comprises long-term filtering using one of an all-zero filter, an all-pole filter, and a pole-zero filter.

9. The method of claim 1, wherein said filtering step (a) comprises filtering without gain-scaling.

10. An apparatus for processing a decoded speech (DS) signal, the DS signal including successive DS frames, each DS frame including DS samples, comprising:

an adaptive filter configured to adaptively filter the DS signal to produce a filtered signal;

a gain scaler configured to gain-scale the filtered signal with an adaptive gain updated once a DS frame, to produce a gain-scaled signal; and

a module configured to perform a smoothing operation to smooth possible waveform discontinuities in the gain-scaled signal.

11. The apparatus of claim 10, wherein the filter comprises at least one of a long-term filter and a short-term filter.

12. The apparatus of claim 11, wherein the filter further comprises at least one of

a long-term filter followed by short-term filter, and

short-term filter followed by long-term filter.

13. The apparatus of claim 10, wherein the gain-scaled signal includes successive gain-scaled signal frames corresponding to the successive DS frames, and the module is configured to perform the smoothing operation to smooth possible waveform discontinuities between adjacent gain-scaled signal frames.

14. The apparatus of claim 13, wherein the module is configured to perform an overlap-add operation based on the adjacent gain-scaled signal frames.

15. The apparatus of claim 13, wherein:

the filter is configured to filter a beginning portion of a current DS frame using a current set of filter coefficients, to produce a first filtered frame portion;

the gain scaler is configured to gain-scale the first filtered frame portion using a current gain, to produce a first gain-scaled frame portion; and

the module includes means for performing the smoothing operation, the smoothing means including

means for filtering and gain-scaling the beginning portion of the current DS frame using a previous set of filter coefficients and a previous gain, respectively, to produce a second gain-scaled portion, and

means for modifying the first gain-scaled frame portion with the second gain-scaled frame portion so as to smooth a possible waveform discontinuity.

16. The apparatus of claim 10, wherein the filter comprises a short-term filter having one of an all-zero filter response, an all-pole filter response, and a pole-zero filter response.

17. The apparatus of claim 10, wherein the filter comprises a long-term filter having one of an all-zero filter response, an all-pole filter response, and a pole-zero filter response.

18. The apparatus of claim 10, wherein the filter is configured to perform filtering without gain-scaling.