Method and communication device for optimizing echo...

Multiplex communications – Duplex – Transmit/receive interaction control

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C370S289000

Reexamination Certificate

active

06678254

ABSTRACT:

BACKGROUND
1. Field
The present invention relates to the field of communications. In particular, this invention relates to a device and method for enhancing the quality of audio transferred over a network without increasing processor cycle time and/or maintaining audio quality with reduced processor cycle time.
2. General Background
During a two-way telephone conference, a caller sometimes can hear his or her own voice as a delayed echo. This phenomenon is referred to as “talker echo”. One cause for talker echo is the presence of signal reflections in a transmission path. Often, signal reflections occur at hybrid circuits. Generally speaking, a hybrid circuit is a transformer circuit that is designed for coupling together different types of transmission lines. It is commonplace for voice-based communication networks to be implemented with different types of transmission lines, and thus, necessities the need for echo cancelling techniques to mitigate or eliminate the echo.
Besides talker echo, acoustic echo is an inherent problem in wireless handsets. The reason is that typical handsets continuously maintain their speaker (or earphone) and microphone in an active state. This creates an acoustic coupling between the speaker (or earphone) and microphone. In addition, when applicable, acoustic echo may be caused by crosstalk in the handset cord. Thus, echo cancellation techniques are also used to mitigate the effects of acoustic coupling.
Typically, an echo canceller is used to reduce or eliminate any type of echo. For example, to reduce echo caused by signal reflections from the hybrid circuit, the echo canceller (i) simulates an estimated echo, and (ii) subtracts the estimate echo from outgoing voice data. The most common method for estimating the amount of echo is through the use of least mean square (LMS) coefficients to model impulse responses for signal reflections from the hybrid circuit. Of course, the same technique can be used to model echo path impulse responses between a receive line (for speaker) and a transmit line (for microphone).
In particular, the LMS coefficients are computed values that are used in connection with samples of audio signals to calculate an estimated echo and minimize residual echo. LMS coefficients are historical in nature; namely, newly computed coefficients are dependent on prior coefficients. Since echo cancellers must be able to handle signal delay and a few milliseconds of signal reflection, a predetermined number of LMS coefficients are stored in memory and correspond to samples of a fixed length delay line. These LMS coefficients are continuously updated every sample time.
One problem with the above-described echo cancellation technique is that a significant number of processing cycles are consumed in order to update the LMS coefficients every sample time. Currently, approximately five million instructions per second (MIPS) are needed to maintain coefficients computed for the most recent 128 audio samples along a single channel. Since the MIPS rate is proportional to the length of the delay line, this MIPS rate has greatly prevented the use of a longer delay line to enhance voice quality as well as the use of higher density (multi-channel) hardware.
Hence, it would be desirable to provide a device and method that reduces the MIPS rate for coefficient updates without quality degradation of the transmitted information.
SUMMARY
In brief, one embodiment of the present invention to a communication device and method for enhancing or maintaining the quality of information communicated over a network with the same or reduced processor cycle time. This is accomplished by separating coefficients into multiple segments, where the coefficients correspond to a memory location of a delay line allocated to contain one of a number of successive audio samples for use in computing echo estimates. After separating the coefficients, certain segments are assigned to a first segment group, where the coefficients of that group are updated more frequently than coefficients assigned to one or more of the remaining groups.
In one embodiment, the grouping is achieved by periodically monitoring all of the segments to determine the cumulative magnitude of the coefficients of each segment and subsequently placing certain segments with the highest cumulative coefficient magnitudes into the first segment group. The remainder of the segments may be placed into a second segment group or multiple groups that are updated less frequently than the first segment group. It is contemplated that the grouping may be based on cumulative energy values in lieu of cumulative magnitudes. In that situation, a greater significance is placed on widely varying coefficients.
This placement of coefficients into segments that are updated at different times provides a variety of benefits. It allows for a reduction in MIPS executed by a processor during echo cancellation while still supporting the same delay line size. Alternatively, it allows for support of longer delay lines without additional consumption of processor cycles, a greater number of channels or enhanced audio features.


REFERENCES:
patent: 4965823 (1990-10-01), Nakagawa et al.
patent: 5016271 (1991-05-01), Ford
patent: 5029167 (1991-07-01), Arnon et al.
patent: 5050160 (1991-09-01), Fuda
patent: 5247512 (1993-09-01), Sugaya et al.
patent: 5278872 (1994-01-01), Greenberg
patent: 5305307 (1994-04-01), Chu
patent: 5313498 (1994-05-01), Sano
patent: 5323458 (1994-06-01), Park et al.
patent: 5323459 (1994-06-01), Hirano
patent: 5343522 (1994-08-01), Yatrou et al.
patent: 5400394 (1995-03-01), Raman et al.
patent: 5530724 (1996-06-01), Abrams et al.
patent: 5594719 (1997-01-01), Oh et al.
patent: 5606550 (1997-02-01), Jangi
patent: 5610909 (1997-03-01), Shaw
patent: 5631899 (1997-05-01), Duttweiler
patent: 5659609 (1997-08-01), Koizumi et al.
patent: 5680451 (1997-10-01), Betts et al.
patent: 5682378 (1997-10-01), Betts et al.
patent: 5748726 (1998-05-01), Unno
patent: 5796819 (1998-08-01), Romesburg
patent: 5905717 (1999-05-01), Hasegawa
patent: 5920548 (1999-07-01), El Malki
patent: 5987098 (1999-11-01), Wintour
patent: 6064873 (2000-05-01), Eriksson et al.
patent: 6160886 (2000-12-01), Romesburg et al.
patent: 6185300 (2001-02-01), Romesburg
patent: 6201866 (2001-03-01), Ariyama et al.
patent: 6236725 (2001-05-01), Takada et al.
□Selective coefficient update of gradient-based adaptive algorithms Aboulnasr, T.; Mayyas, K, Speech, and Signal Processing, 1997. ICASSP-97., 1997 IEEE Conference on, vol.: 3 , 1997 Page(s): 1929-1932 vol. 3.*
Kuo, Sen Multiple-Microphone Acoustic Echo Cancellation System with the Partial Adaptive Process, Digital Signal Processing 3,(1993).

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Method and communication device for optimizing echo... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Method and communication device for optimizing echo..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and communication device for optimizing echo... will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-3254534

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.