Electrical audio signal processing systems and devices – Dereverberators
Reexamination Certificate
1998-12-14
2004-08-17
Nguyen, Duc (Department: 2643)
Electrical audio signal processing systems and devices
Dereverberators
C379S406010
Reexamination Certificate
active
06778671
ABSTRACT:
BACKGROUND
The present invention relates to an echo canceler provided in a data processing system.
Echo cancelers are known per se. In many audio applications, a first audio signal is broadcast from an audio output device such as a speaker. The first audio signal may be captured by an audio input device such as a microphone and returned to the signal's source. The audio input device also may capture other audio signals. The first audio signal, often called the “echo,” may interfere with the intelligibility or perceived quality of the other audio signals. Echo cancelers are used to remove the echo signal but permit the other audio signals to remain.
FIG. 1
illustrates an audio sub-system of a known data processing system in which echo cancellation may be used. The audio sub-system receives and buffers audio data, called the “reference channel,” to be output from the system. The audio sub-system also captures and buffers ambient audio in a second channel, called the “echo channel.” It delivers the echo channel to the data processing system.
The audio subsystem includes an echo canceling system
100
, an output device
200
and an input device
300
. The echo canceling system
100
includes an audio rendering filter
110
and an echo canceler
120
. The audio rendering filter
110
receives and formats as necessary the reference channel. It outputs audio data to the output device
200
.
The output device
200
includes a buffer
210
(“output buffer”) and a speaker
220
or other announcement device (“speaker”). As is known, the output device
200
also may include a controller and device driver equipment (not shown). The reference channel is received from the echo cancellation system at the output buffer
210
. The output buffer
210
stores the reference channel for a predetermined period of time before delivering it to the speaker
220
. The reference channel is output from the output device
200
through the speaker
220
. As the reference channel is output from the output buffer
210
to the speaker
220
, a copy of the reference channel is delivered to the input device
300
over line
230
.
The input device
300
includes its own buffer
310
(“input buffer”) and a microphone
320
or other audio capture device (“microphone”). The microphone
320
captures audio data and generates an echo channel signal therefrom. It delivers the echo channel signal to the input buffer
310
. The input buffer
310
stores the echo channel for a predetermined period of time, then delivers it to the echo cancellation system
100
.
The input buffer
320
also receives a copy of the reference channel from the output device
200
. It buffers the reference channel for a predetermined period of time then delivers it to the echo canceler
120
.
FIG. 1
figuratively illustrates the echo phenomenon that may occur in a speakerphone type of implementation of the audio sub-system. Assume that the audio sub-system is provided in a 10′×10′ room. When the reference channel is output by the speaker
220
, sound energy may travel to the microphone
320
through a variety of different paths, P
1
-P
4
. The shortest path is provided on a direct path P
1
from the speaker
220
to the microphone
320
. It is associated with a first delay. Other paths P
2
-P
4
may occur because of reflections from nearby objects such as wall surfaces, people and things. Each path is associated with its own delay. Of course, people and things (including the speaker and microphone) may move which can cause the delay and/or attenuation of the various paths to change with time. The echo channel captured by the microphone
320
is, in fact, an aggregate of a variety of reflections that may occur in a particular environment. The echo channel may include an echo signal that represents the multiple echos of paths P
1
-P
4
and also other signals representing audio generated at the microphone
320
such as input speech.
The echo canceler
120
receives the reference signal having been delayed by the input and output buffers
210
-
310
and the captured echo channel from the input device
300
. For the echo canceler
120
to operate effectively, it must receive an onset of the echo signal within a predetermined positive offset from its receipt of the corresponding signal in the reference channel. The echo canceler
120
typically stores only a predetermined portion of the reference channel (e.g. 60-200 msec). If the acoustical echo signal corresponding to the buffered portion is not received within the 60-200 msec window, the echo canceler
120
will have discarded the portion of the reference channel when the echo signal is received. The echo canceler
120
would neither recognize the echo signal as echo nor cancel it in this case.
Traditional computer audio systems include “correlated audio” device drivers that can measure the delay imposed by the input and output buffers
210
,
310
. The same buffering is applied to both the reference channel and the echo channel. Accordingly, only the delay of paths P
1
-P
4
contribute to any skew between the reference channel and an echo signal contained in the echo channel. Adaptive filtering at the echo canceler
120
approximates the delay of sound paths P
1
-P
4
to correct for the skew.
Modern computer systems may discontinue use of the correlated audio device drivers that makes the precise delay tuning possible. Thus, there is a need in the art for an audio system that maintains time alignment between a reference channel and an echo channel even in the absence of correlated audio device drivers.
SUMMARY
An embodiment of the present invention provides an echo cancellation system in which an echo canceler has an echo channel signal input, a reference channel signal input and a control output, the two inputs being uncorrelated, and a delay buffer has an input for a reference channel signal, an output in communication with reference channel signal input and a control input in communication with the control output from the echo canceler.
REFERENCES:
patent: 4582963 (1986-04-01), Danstrom
patent: 5263019 (1993-11-01), Chu
patent: 5664011 (1997-09-01), Crochiere et al.
patent: 5721782 (1998-02-01), Piket et al.
patent: 6085072 (2000-07-01), Komiya
patent: 6108720 (2000-08-01), Tal et al.
patent: 6181753 (2001-01-01), Takada et al.
patent: 6272173 (2001-08-01), Hatamian
patent: 6320958 (2001-11-01), Sekine
patent: 6389440 (2002-05-01), Lewis et al.
Omologo et al, “Acoustic Event Localization Using a Crosspower-Spectrum Phase Based Technique”, IEEE, 0-7803-1775-0/94, pp. II-273 through II-276, 1994.
Reed et al, “Time Delay Estimation Using the LMS Adaptive Filter-Static Behavior”, IEEE, 0096-3518/81/0600-0561, pp. 561-576, 1981.
Intel Corporation
Kacvinsky John F.
Lao Lun-See
Nguyen Duc
LandOfFree
Method of reference to echo time alignment for facilitation... 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 of reference to echo time alignment for facilitation..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method of reference to echo time alignment for facilitation... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3317933