Perceptual echo gate approach and design for improved echo control to support higher audio and conversational quality

Patent metadata

Field	Value
Publication number	US-9503815-B2
Application number	US-201414289437-A
Country	US
Kind code	B2
Filing date	May 28, 2014
Priority date	May 28, 2014
Publication date	Nov 22, 2016
Grant date	Nov 22, 2016

Abstract

To address issues with present echo gate control, a method and apparatus for more intelligently operating an echo gate is described herein. In particular, the decision of whether to mute an uplink signal, or not, is formulated herein as primarily a perceptual decision based on an appropriate analysis of the perceptual interaction of the current residual echo and the current near-end signal(s). By doing so, the application of muting through an echo gate may be minimized and/or more appropriately engaged. This will lead to fewer dropouts and muting of speech onsets and offsets 1) during periods such as double-talk or 2) during periods of downlink playback in the presence of low near-end signal levels, two cases of particular importance.

First claim

What is claimed is:

1. A method for controlling an echo gate in a communications system, comprising:
detecting sound at a near-end location by a set of microphones to produce a primary signal, wherein the sound includes one or more of 1) desired sounds to be transmitted to a far-end location, 2) echo introduced by a loudspeaker playing sound received from the far-end location, and 3) noise introduced at the near-end location;
processing the primary signal to reduce the echo;
estimating an amount of residual echo in the processed primary signal;
estimating an amount of the desired sounds in the processed primary signal;
performing a perceptual analysis on the processed primary signal using the estimations of the amount of residual echo and the amount of the desired sounds, to determine masking properties of residual echo in the processed primary signal in relation to the desired sounds in the processed primary signal; and
toggling an echo gate between open and closed states based on the perceptual analysis.

2. The method of claim 1 , wherein toggling the echo gate is performed between un-mute, partial mute and full mute states.

3. The method of claim 1
wherein processing the primary signal to reduce the echo comprises using one or more of linear echo cancellers, a non-linear residual echo suppressor, and a noise suppressor.

4. The method of claim 3 , further comprising:
adjusting the estimation of the amount of residual echo based on suppression factors introduced by the non-linear residual echo suppressor and the noise suppressor.

5. The method of claim 1 , wherein estimating the amount of the desired sounds in the processed primary signal comprises:
subtracting the estimation of the amount of residual echo from the processed primary signal.

6. The method of claim 5 , further comprising:
biasing the estimation of the amount of residual echo prior to subtracting from the processed primary signal.

7. The method of claim 1 , further comprising:
comparing the estimation of the amount of residual echo and the estimation of the amount of desired sounds against a set of thresholds, wherein the perceptual analysis is performed in response to the estimation of the amount of residual echo and the estimation of the amount of desired sounds falling within an intermediate region between the set of thresholds.

8. The method of claim 1 , further comprising:
toggling the echo gate open upon determining by the perceptual analysis that that the residual echo fails to mask the desired sounds by a predefined level.

9. The method of claim 1 , wherein performing the perceptual analysis comprises:
estimating a masking threshold based on signals representing past and present desired sounds in the processed primary signal, by spreading power of said signals over frequency by cochlear filters, calculating a tonality measure that determines stationarity of a masker as a function of frequency, and mapping from tonality to a masking threshold offset;
computing a supra-threshold loudness metric based on the masking threshold; and
comparing the supra-threshold loudness metric with a set of thresholds to determine whether the residual echo masks the desired sounds by a predefined level.

10. The method of claim 9 , further comprising:
normalizing the supra-threshold loudness metric based on the estimation of the amount of the desired sounds in the processed primary signal.

11. A system for controlling an echo gate in a communications system, comprising:
a set of microphones to detect sound at a near-end location to produce a primary signal, wherein the sound includes one or more of 1) desired sounds to be transmitted to a far-end location, 2) echo introduced by a loudspeaker playing sound received from the far-end location, and 3) noise introduced at the near-end location;
an echo canceller to process the primary signal to reduce the echo;
an estimation unit to estimate an amount of residual echo in the processed primary signal and estimate an amount of the desired sounds in the primary signal;
an echo gate to 1) determine masking properties of residual echo in the processed primary signal in relation to the desired sounds in the processed primary signal and 2) toggle uplink transmission between mute and un-mute states based on the masking properties.

12. The system of claim 11 , wherein the echo canceller comprises:
one or more of linear echo cancellers, a non-linear residual echo suppressor, and a noise suppressor.

13. The system of claim 12 , wherein the estimation unit further adjusts the estimate of the amount of residual echo based on suppression factors introduced by the non-linear residual echo suppressor and the noise suppressor.

14. The system of claim 11 , wherein the echo gate compares the estimation of the amount of residual echo and the estimation of the amount of desired sounds against a set of thresholds, wherein the masking properties are determined in response to the estimation of the amount of residual echo and the estimation of the amount of desired sounds falling within an intermediate region between the set of thresholds.

15. The method of claim 11 , wherein the echo gate is opened upon determining that that the residual echo fails to mask the desired sounds by a predefined level.

16. The system of claim 11 , wherein the echo gate is to further:
estimate a masking threshold based on signals representing past and present desired sounds in the processed primary signal, by spreading power of such signals over frequency by cochlear filters, calculating a tonality measure that determines stationarity of a masker as a function of frequency, and a mapping from tonality to a masking threshold offset;
compute a supra-threshold loudness metric based on the masking threshold; and
compare the supra-threshold loudness metric with a set of thresholds to determine whether the residual echo masks the desired sounds by a predefined level.

17. An article of manufacture for controlling an echo gate in a communications system, comprising:
a non-transitory machine-readable storage medium that stores instructions which, when executed by a processor in a computing device,
detect sound at a near-end location by a set of microphones to produce a primary signal, wherein the sound includes one or more of 1) desired sounds to be transmitted to a far-end location, 2) echo introduced by a loudspeaker playing sound received from the far-end location, and 3) noise introduced at the near-end location;
process the primary signal to reduce the echo;
estimate an amount of residual echo in the processed primary signal;
estimate an amount of the desired sounds in the processed primary signal;
perform a perceptual analysis on the processed primary signal using the estimations of the amounts of residual echo and the desired sounds, to determine masking properties of residual echo in the processed primary signal in relation to the desired sounds in the processed primary signal; and
toggle uplink transmission between mute and un-mute states based on the masking properties that is determined by the perceptual analysis.

18. The article of manufacture of claim 17 , wherein the non-transitory machine-readable storage medium stores further instructions which when executed by the processor:
process the primary signal to reduce the echo using one or more of linear echo cancellers, a non-linear residual echo suppressor, and a noise suppressor.

19. The article of manufacture of claim 18 , wherein the non-transitory machine-readable storage medium s

Assignees

• Apple Inc

Inventors

• Ramprashad Sean Anthony
• Yang Qing

Classifications

• H04M19/08

  with current supply sources at the substations (generating ringing current H04M19/04) · CPC title

• H04R3/02 Primary

  for preventing acoustic reaction {, i.e. acoustic oscillatory feedback (specially adapted for hearing aids H04R25/453)} · CPC title

• H04M9/08

  Two-way loud-speaking telephone systems with means for conditioning the signal, e.g. for suppressing echoes for one or both directions of traffic · CPC title