Telephonic communications – Substation or terminal circuitry – For loudspeaking terminal
Reexamination Certificate
1997-12-18
2002-04-23
Woo, Stella (Department: 2643)
Telephonic communications
Substation or terminal circuitry
For loudspeaking terminal
C379S390010, C379S388010
Reexamination Certificate
active
06377679
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a speakerphone that estimates an echo canceling performance of an echo canceler through an easy processing and improves the transmit and receive switching performance of the voice switch, in a full-duplex communication system using a speaker and microphone.
2. Description of the Related Art
The speakerphone used for a telephone conversation using a speaker and microphone without using a handset has been applied widely to a teleconference system that connects plural locations, and to the automobile telephone system wherein the driver cannot free his hands from the steering for obvious safety reasons.
However, this speakerphone involves troublesome phenomena, such as an acoustic echo generated by sounds emitted from the speaker returning to the microphone while reflecting, and a line echo generated by a talker's uttered voice being reflected at the connections on the communication line due to the impedance mismatching thereat.
FIG. 6
is a chart for simply explaining the acoustic echo and the line echo.
What makes the problem acute is that the acoustic echo path and the line echo path coincide so as to make up a closed loop (formed of a microphone
61
, communication line SP, and speaker
62
), as shown in FIG.
6
. If the gain of the foregoing dosed loop exceeds 1, it will generate an oscillation (howling) inside the closed loop, which will in the worst case disable the conversation. Even if the howling does not occur, if there is a line echo, the talker's uttered voice will be emitted from the speaker
62
with a delay, and hence, the talker will be in a trouble of speaking.
Devices have been provided in order to avoid the influence of these echoes, which can be classified roughly into two. One of them is a half duplex voice switching system, wherein, when a near-end talker is speaking, an electric loss is inserted on the receive path of the talker (transmit state), when the talker is listening to, an electric loss is inserted on the transmit path of the talker (receive state). In this system, the switching of the transmit and the receive state is carried out on the basis of the voices uttered by the near-end talker or the far-end talker.
The other one is an echo canceling system, wherein an adaptive filter to estimate the characteristic of the foregoing echo is employed to produce a signal similar to the echo, and the signal is subtracted from the transmit and the receive paths to thereby remove the echo signal from the dosed loop. In the echo canceling system, echoes are removed in real time, both the transmit and receive paths are not closed, and hence, the full duplex communication is possible.
The technique relating to the speakerphone using the foregoing voice switch is disclosed, for example, in the Japanese Patent Application Laid-open No. 5-44221.
FIG. 7
is a functional block representation of the speakerphone disclosed in the foregoing document.
As shown in
FIG. 7
, a speakerphone
100
using the voice switch comprises a transmit section
200
, a receive section
300
, and a computer
110
. The transmit section
200
includes a multiplexer
210
for temporarily storing a plurality of input signals such as speech signals inputted from a microphone
111
, a mute control
211
to dose the transmit path in accordance with a control signal from the computer
110
described later, a high pass filter
212
for removing background noises contained in the foregoing speech signals, a programmable attenuator
213
(equivalent to receive state setting means) for giving attenuation to the foregoing speech signals passed through the high pass filter
212
in accordance with a control signal from the computer
110
, an envelope detector
214
for detecting an envelope of a speech signal outputted from the high pass filter
212
, a low pass filter
215
for reducing switching noises generated by the programmable attenuator
213
and shaping output waveforms to a communication line
101
, and a logarithmic amplifier
216
for logarithmically amplifying an output from the envelope detector
214
.
The receive section
300
contains functionally the same circuits as the transmit section
200
: a multiplexer
310
for temporarily storing a plurality of input signals such as speech signals received through a communication line
102
, a mute control
311
to dose the receive path in accordance with a control signal from the computer
110
, a high pass filter
312
for removing background noises contained in the foregoing speech signals, a programmable attenuator
313
(equivalent to transmit state setting means) for giving attenuation to the foregoing speech signals passed through the high pass filter
312
in accordance with a control signal from the computer
110
, an envelope detector
314
for detecting an envelope of a speech signal outputted from the high pass filter
312
, a low pass filter
315
for reducing switching noises generated by the programmable attenuator
313
and shaping output waveforms to a speaker
112
, and a logarithmic amplifier
216
for logarithmically amplifying an output from the envelope detector
314
.
And, the foregoing computer
110
(equivalent to state switching means) receives signals from the logarithmic amplifiers
216
,
316
through a multiplexer
117
and an A/D converter
115
, and controls the mute controls
211
,
311
and the programmable attenuators
213
,
313
. Further, the computer
110
is connected to a calibration circuit
113
as well. The calibration circuit
113
feeds a specific calibration tone to the multiplexers
210
and
310
to assist the estimation of system characteristics.
The operation of the foregoing speakerphone, specially a transmit break-in operation switching from the receive state to the transmit state will hereunder be described.
FIG. 8
is a flow chart for explaining the transmit break-in operation.
As shown in
FIG. 8
, when the process comes into step
1001
, the speakerphone enters the receive state. Then, the process advances to step
1002
where a determination is made as to whether a transmit signal TX-S inputted from the microphone
111
exceeds an expected transmit signal IX-E by a specific threshold Th. Here, the expected transmit signal TX-E is a transmit signal expected to be generated by the coupling of the receive signal RX-S from the speaker
112
to the microphone
111
. The reason to provide this step
1002
is to prevent a phenomenon that the device generates the self-switching by the receive signal RX-S emitted from the speaker
112
and the influence of an acoustic echo, while the near-end talker does not speak.
At step
1002
, if the transmit signal TX-S exceeds the expected transmit signal TX-E, the process advances to step
1003
where a determination is made as to whether the transmit signal TX-S exceeds a transmit noise TX-N by a specific threshold Th. The decision at this step is provided to determine whether the transmit signal TX-S is a voice signal or a noise signal.
At step
1003
, after the transmit signal TX-S is confirmed as a voice signal, the process advances to step
1004
where a comparison is made whether the transmit signal TX-S exceeds the receive signal RX-S by a specific threshold Th. And, if the transmit signal TX-S is greater than the receive signal RX-S at step
1004
, the process moves to step
1005
where the holdover timer is initialized, and then the process moves to step
1006
where it brings the device into the transmit state.
Thus, the foregoing speakerphone prevents an error switching due to the acoustic echo by comparing the transmit signal TX-S with the expected transmit signal TX-E. To prevent the error switching due to the line echo is performed substantially in the same manner as in the acoustic echo, and the description will be omitted
Incidentally, the threshold used in the foregoing expected transmit signal TX-E and the decision at step
1004
is determined by using a calibration tone actually outputted from the calibration circuit
113
. More concretely,
Hashimoto Hiroshi
Nakazumi Akira
Nishimoto Yoshiro
Nishisaka Masahiro
Takahashi Tetsuya
Kabushiki Kaisha Kobe Seiko Sho
Reed Smith LLP
Woo Stella
LandOfFree
Speakerphone does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Speakerphone, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Speakerphone will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2872115