Method and apparatus for encoding an audio signal

• US 9,129,600 B2
• Filed: 09/26/2012
• Issued: 09/08/2015
• Est. Priority Date: 09/26/2012
• Status: Active Grant

• Current Assignee: Google Technology Holdings LLC (Alphabet Inc.)
• Original Assignee: Google Technology Holdings LLC (Alphabet Inc.)
• Inventors: Gibbs, Jonathan A., Francois, Holly L.
• Primary Examiner(s): Desir, Pierre-Louis
• Assistant Examiner(s): THOMAS-HOMESCU, ANNE L
• Application Number:
  US13/626,923
• Publication Number:
  US 20140088973A1
• Time in Patent Office: 1,077 Days
• Field of Search: 704200-500, 381/107
• US Class Current: 1/1
• CPC Class Codes: G10L 19/20 using sound class specific coding, hybrid encoders or object based coding
• CPC Class Codes: G10L 25/81 for discriminating voice from music

Abstract

A hybrid speech encoder detects changes from music-like sounds to speech-like sounds. When the encoder detects music-like sounds (e.g., music), it operates in a first mode, in which it employs a frequency domain coder. When the encoder detects speech-like sounds (e.g., human speech), it operates in a second mode, and employs a time domain or waveform coder. When a switch occurs, the encoder backfills a gap in the signal with a portion of the signal occurring after the gap.

Claims

1. A method of encoding an audio signal the method comprising:

processing the audio signal in a first encoder mode;

switching from the first encoder mode to a second encoder mode at a first time;

processing the audio signal in the second encoder mode, wherein a processing delay of the second mode creates a gap in the audio signal having a time span that begins at or after the first time and ends at a second time;

copying a portion of the processed audio signal wherein the copied portion occurs at or after the second time; and

inserting a signal into the gap, wherein the inserted signal is based on the copied portion, wherein the copied portion comprises a time-reversed sine window portion and a cosine window portion, wherein inserting the copied portion comprises combining the time-reversed sine window portion with the cosine window portion, and inserting at least part of the combined sine and cosine window portions into the gap portion.

2. The method of claim 1, wherein the time span of the copied portion is longer than the time span of the gap, the method further comprising combining an overlapping part of the copied portion with at least part of the processed audio signal that occurs after the second time.

3. The method of claim 1, wherein switching the encoder from a first mode to a second mode comprises switching the encoder from a music mode to a speech mode.

4. The method of claim 1, wherein the steps are performed on a first communication device, the method further comprising:

following the inserting step, transmitting the encoded speech signal to a second device.

5. The method of claim 1, further comprising:

if the audio signal is determined to be a music signal encoding the audio signal in the first mode;

determining that the audio signal has switched from the music signal to a speech signal;

if it is determined that the audio signal has switched to be a speech signal encoding the audio signal in the second mode.

6. The method of claim 5, wherein the first mode is a music coding mode and the second mode is a speech coding mode.

7. The method of claim 1, further comprising using a frequency domain coder in the first mode and using a CELP coder in the second mode.

8. An apparatus for encoding an audio signal the apparatus comprising:

an encoder having a processor configured to act as

a first coder;

a second coder;

a speech-music detector, wherein when the speech-music detector determines that an audio signal has changed from music to speech, the audio signal ceases to be processed by the first coder and is processed by the second coder;

wherein a processing delay of the second coder creates a gap in the audio signal having a time span that begins at or after the first time and ends at a second time; and

a missing signal generator that copies a portion of the processed audio signal wherein the copied portion occurs at or after the second time and inserts a signal based on the copied portion into the gap,

wherein the copied portion comprises a time-reversed sine window portion and a cosine window portion, wherein inserting the copied portion comprises combining the time-reversed sine windowed portion with the cosine windowed portion, and inserting at least part of the combined sine and cosine windowed portions into the gap portion.

9. The apparatus of claim 8, wherein the signal output by the missing signal generator is a gap-filled bandwidth extension target signal the apparatus further comprising a gain computer that uses the gap-filled bandwidth extension target signal to determine ideal gains for at least part of the audio signal.

10. The apparatus of claim 8, wherein the time span of the copied portion is longer than the time span of the gap, the method further comprising combining an overlapping part of the copied portion with at least part of the processed audio signal that occurs after the second time.

11. The apparatus of claim 8, wherein the signal output by the missing signal generator is a gap-filled bandwidth extension target signal the apparatus further comprising a linear predictive coding analyzer that determines the spectrum of the gap-filled bandwidth extension target signal and, based on the determined spectrum, outputs linear predictive coding coefficients.

12. The apparatus of claim 8, wherein the first coder is a frequency domain coder and the second coder is a CELP coder.