Speaker-phone system change and double-talk detection

Details Speaker-phone system change and double-talk detection Speaker-phone system change and double-talk detection

2000-09-29

2004-06-01

Nguyen, Duc (Department: 2643)

Telephonic communications

Echo cancellation or suppression

C379S406020, C379S406030, C379S406040, C379S406050, C379S406060, C379S406070, C379S406080, C379S406090, C379S406110, C379S406120, C379S406130, C379S406140, C379S406150, C379S406160, C379S420010, C379S388020, C379S406100, C370S286000

Reexamination Certificate

active

06744884

ABSTRACT:

BACKGROUND
1. Technical Field
The invention relates generally to speaker-phones; and, more particularly, it relates to a combined speaker-phone system that is operable to perform system change and double talk detection using an optimal step size in certain embodiments of the invention.
2. Related Art
Conventional speaker-phone systems commonly employ some kind of energy measures to detect system changes in the echo path. These conventional systems commonly use gain measures of the echo canceller and combine them with signal energy measurements to detect system changes. This conventional method to perform system change detection is suitable for large system changes, particular those where the gain of the echo canceller is suddenly negative. However, this conventional method to perform system change detection is unsuitable for detecting smaller system changes, and it is highly unreliable to perform system change detection in general.
For those conventional speaker-phone systems that attempt to perform double-talk detection, certain assumptions concerning the various far end and near end signals commonly exist. For example, a common assumption is that the far end echoes are of lower amplitude that those of the near end signal. Therefore, given that this is a primary design constraint, the speaker-phone double-talk detection method will operate only when this assumption is true. In reality, this assumption is often not the case. Another assumption effectuates the filtering out of the received signal in the speaker-phone. For example, the transmit signal is filtered with the inverse of the band-stop filter in the receive path to perform double-talk detection. In this case, when the energy of the filtered signal in the transmit path increases above a certain threshold, then double-talk is assumed. These several assumptions lead to double-talk detection that is often performed with a substantially low rate of success. For example, the perceptual quality of the operation if the speaker-phone, when double-talk is not detected is substantially reduced. The user friendliness of the speaker-phone is significantly compromised. Other deficiencies exist in speaker-phone systems that employ these many conventional assumptions while attempting to offer multiple functionality.
For example, conventional speaker-phones that attempt to perform both system change detection and double-talk detection typically fail across many fronts. One is the highly unreliable manner in which they seek to perform system change detection, as described above. Another deficiency is inherent assumptions that most designers of double-talk detection speaker-phones rely upon in making critical design decisions, as also mentioned above. The deficiencies of both of these conventional methods is that the total, combined deficiencies typically result in an overall speaker-phone system that simply fails to provide a high degree of user satisfaction. The perceptual qualities of the speaker-phone are radically compromised, in that, unreliable system change detection that can potentially lead to a false system change detection or a failure to perform an accurate system change detection in the event of an actual system change detection will lead to the speaker-phone system to provide significantly reduced perceptual quality. As mentioned above, when the deficiencies of one such method are combined with the deficiencies of another method of function that is sought to be provided within a speaker-phone, then the total deleterious effects are commonly magnified. That is to say, the overall functionality of the speaker-phone suffers even more. The conventional speaker-phone systems do not provide for highly robust and reliable operation of both system change detection and double-talk detection.
Further limitations and disadvantages of conventional and traditional systems will become apparent to one of skill in the art through comparison of such systems with the present invention as set forth in the remainder of the present application with reference to the drawings.
SUMMARY OF THE INVENTION
Various aspects of the present invention can be found in a speaker-phone that is operable to perform system change detection and double-talk detection using an optimal step size. The speaker-phone itself contains, among other things, a main control circuitry, an echo canceller communicatively coupled to the main control circuitry, a transmit circuitry communicatively coupled to the main control circuitry, a receive circuitry communicatively coupled to the main control circuitry, a system change detection circuitry that is operable to perform system change detection, the system change detection circuitry modifies the transmit circuitry and the receive circuitry when a system change is detected, a double-talk detection circuitry that is operable to perform double-talk detection such that the double-talk detection circuitry modifies the transmit circuitry and the receive circuitry when double-talk is detected, a step size control circuitry that is operable to perform step size optimization of a step size. The speaker-phone is operable to differentiate between the system change and the double-talk by comparing a correlation value to a threshold, the correlation value is calculated using an input signal of the echo canceller and an output signal of the echo canceller.
In certain embodiments of the invention, the speaker-phone also contains a microphone having an input signal, and the double-talk detection circuitry calculate a number of autocorrelation coefficients of the output signal of the echo canceller and the input signal of a microphone to enhance double-talk detection. The system change detection circuitry considers a speech detect level and a gain of the echo canceller to perform system change detection. The double-talk detection circuitry uses a time constant to calculate background noise energy. The double-talk detection circuitry calculates a weighted distance to perform double-talk detection. The speaker-phone is a full band digital speaker-phone in one embodiment and a full band analog speaker-phone in another embodiment.
Other aspects of the present invention can be found in a speaker-phone that is operable to perform both system change detection and double-talk detection. The speaker-phone itself contains, among other things, a main control circuitry, a system change detection circuitry, communicatively coupled to the main circuitry, that is operable to perform system change detection, and a double-talk detection circuitry, communicatively coupled to the main circuitry, that is operable to perform double-talk detection. The speaker-phone is operable to differentiate between a system change and a double-talk by comparing a correlation value to a threshold.
In certain embodiments of the invention, the speaker-phone also contains a step size control circuitry that is operable to perform step size optimization of a step size when performing at least one of system change detection and double-talk detection. The step size control circuitry automatically throttles the step size during double-talk. The speaker-phone also contains an acoustic echo canceller having a gain measure, and the speaker-phone is operable to prevent a false double-talk detection by considering the echo canceller gain measure. The speaker-phone also contains a transmit circuitry and a receive circuitry. The speaker-phone modifies the transmit circuitry and the receive circuitry to operate at double-talk when operating at full duplex. The double-talk detection circuitry uses a tuning factor to control a sensitivity used to perform the double-talk detection. The tuning factor is adjusted by a designer of the speaker-phone. When the speaker-phone further contains an acoustic echo canceller, the speaker-phone ceases updating of the acoustic echo canceller when a perturbation is detected.
Other aspects of the present invention can be found in a method to perform system change detection and double-talk detection in a speaker-phone. The method involves, among other things, dete

Affiliated with

Also associated with

Rating

Speaker-phone system change and double-talk detection does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Speaker-phone system change and double-talk detection, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Speaker-phone system change and double-talk detection will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3322372