JPH03214955A - Incoming call transfer device - Google Patents

Abstract

PURPOSE:To improve the talking quality by decreasing an estimate error of bypass erasure quantity. CONSTITUTION:Talking state discrimination sections 5, 5' are provided, the discrimination section 5 takes delay information of a sender side A from an echo canceller into account and compares 0 cross number (that is, frequency) of voice signals at a reception output point Pout and a transmission input point Sin. when the 0 cross number is coincident in this case, a bypass erasure quantity estimate means 4 is active. However, when the zero cross number cannot be discriminated to be coincident, the bypass erasure quantity estimate means 4 is inactive. Thus, the estimate error of the bypass erasure quantity is decreased and the talking quality is improved.

Description

[Detailed Description of the Invention] [Summary] The purpose of the present invention is to improve call quality by reducing the estimation error of the amount of loop cancellation in a call transfer device used when providing a call transfer service. In a call transfer device including a gain control means, audio signal levels at reception input points and transmission output points on the A side and the B side are detected. A loop consumption amount estimating means for estimating and transmitting the amount of loop cancellation using the audio signal level detected when the operable signal is input; and delay information from the reception output point to the transmission input point sent from the echo canceller; Using the audio signals appearing at the reception output points and transmission input points on the A side and B side. The A side and B
The audio signal at the transmission input point on the side is the audio signal of only the B-side speaker or the A-side speaker. Or determine either the voice signal of the A-side speaker or the B@speaker, or the synthesized voice signal of the A-side speaker and the B-side speaker, and determine the voice signal of the B-side speaker or the A-side speaker. The system is configured to include a call status determination section that sends out an operational signal when it is determined that this is the case, and sends out an inoperable signal when it is determined otherwise. [Industrial Application Field] The present invention relates to a call transfer device used when providing a call transfer service. A call forwarding service is a service that transfers a subscriber's telephone calls to another preset telephone. When using this device, it is necessary to improve the call quality by reducing the estimation error of the amount of interference cancellation. [Prior Art] Fig. 4 is a block diagram of a conventional example, and Fig. 5 is an explanatory diagram of the operation of the lump-sum erasure amount estimation part in Fig. 4. Generally, the lines connected to both sides of the call transfer device are local lines.
Regardless of long-distance line. Automatic gain-Il (hereinafter referred to as A) that compensates line loss for lines with arbitrary different transmission losses.
(abbreviated as GC) is required. In addition, a hybrid transformer is installed in the 2-wire to 4-wire converter, but due to mismatching of this hybrid transformer and wraparound at the telephone, echoes are generated and ringing sounds occur.
Therefore, as shown in Fig. 4, a call transfer device consisting of an echo canceller (hereinafter abbreviated as EC) 14, 14' and an AGC amplifier 32, 32', etc., always maintains loop loss while It is necessary to prevent the line loop from making noise. The operation of the call transfer device shown in FIG. 4 will be explained below assuming that a voice signal is transmitted from side B to side A. First of all. The analog voice signal sent from the B-side speaker is transmitted through the B-side line. It is converted into a digital signal by an analog/digital converter (hereinafter abbreviated as A/D converter) 12° via a hybrid transformer 11". This digital signal is sent to a voice detector 22° via an I!C 14". Enter it. The remainder is amplified to a predetermined level by the AGC amplifier 32°. It passes through the EC 14, is converted back to an analog audio signal by the digital/analog converter (hereinafter abbreviated as D/A converter) 13, and then is converted to an analog audio signal by the hybrid transformer 1.
1. It is added to the handset via the line on the A side. However, part of the audio signal is transferred to this hybrid transformer 1.
Due to the mismatch of .1 and wraparound at the A-side telephone, the signal is converted back into a digital signal by the A/D converter 12.
EC 14. The signal is sent out to the B side through the AGC amplifier 32, and a portion is applied to the voice detector 2l. Voice detectors 22".21" and 22. 21 detects the level of the input audio signal, and this detection result is used to estimate the amount of wraparound cancellation at a wraparound cancellation amount estimation part (hereinafter abbreviated as ERL estimation part) 23°, 23. Note that the audio detectors 221 and 21 determine the presence or absence of an audio signal from the level of the detected audio signal, and turn on/off the operation of the AGC amplifiers 32° and 32 in accordance with the determination result. Further, the loop gain controller 31 uses the estimated amount from the wraparound cancellation amount estimation part to determine the level of the audio signal. Connected to the higher transmitting output point. In other words, the AGC amplifier on the dominant side is given a gain that compensates for the line loss of the corresponding line, and the AGC amplifier on the inferior side is given a gain so that no noise occurs in the 4-wire loop, that is, the loop gain is set to 1 or less. to control the gain.

Next, detect presence/absence. URL estimation will be explained in detail. (1) Detection of voice/silence (see Figure 5) Voice detector 22. 21, the reception input point Ran and transmission output point S on the A side
0■ level (hereinafter referred to as moving average power) is detected, and the audio detectors 22° and 2l° detect the moving average power of the transmission output point and reception input point on the B side as shown below. Note that since the detection of voice/silence and the estimation of URLH are the same on side A and side B, side A will be taken as an example. That is, as shown in FIG. 5(b), for example, 16 bits of sampling data (amplitude value) is sent to the audio detector 22 every 125 μs. Enter in 21. Voice detector 22.21 is al+ at+ am+ 84
Find the sum of a certain interval (for example, 2ws) by shifting one sampling at a time. Pma"P- + P-
pn-11 and Psa=P. + pR- p,, calculate the moving average power P at the receiving input point and the transmitting output point. Then, from the formula shown in Figure 4(a) 10, E
Find RL (see Figure 5 (a) 10, ■). Note that +Ps is the moving average power up to now, IPR is the latest power, and IPII-11 is the oldest power. Also, the levels of the transmitting output point and the receiving input point are determined simultaneously. Additionally, a voice detector 22. In 21, the obtained moving average power is compared with the threshold value for identifying voice/silence. If it is greater than the threshold, there is a sound. If it is smaller than the threshold, it is determined that there is no sound, and the determination result is sent to the AGC amplifier. (2) Estimation of URL The URL estimation part 23 is processed by the voice detector 21. When the moving average power from 22 is input. As shown in FIG. 5, 6》10, the moving average power at the receiving input point R{m is calculated from the transmitting output point S. ,, Subtract the moving average power and get l! Find RL. Send it to the loop gain control section. ? , ERL is estimated at the transmission output point S. , ■ is less than Y dolls, and the moving average power at the receiving input point Ri is more than X dime, that is, A
The side speaker is not speaking. Once it is confirmed that the B-side speaker is speaking (this is called the update condition), the previously estimated ERL is updated. (3) Loop gain controller Now, when the wraparound cancellation amount ERL . The moving average powers of the B-side AGC amplifier 32", which has a large power, are compared with each other, and a control signal is applied to increase the gain by the AGC " to achieve a predetermined output level. AGC+A for the AGC amplifier 52 of
GC ”-ERL-ERL” < 1 (
The gain is controlled to satisfy 1). This prevents noise from occurring within the 4-wire loop. Note that if the left side of equation (1) is greater than 1, a noise will occur. ? [Problem to be Solved by the Invention]] Here, at the transmission output point S. The voice signal of the B-side speaker appearing in ① is sent to the EC 14, the D/A converter 13. Hybrid transformer 11, A side transmitting/receiving device. Hybrid transformer 1l^/D converter 12. EC 14 and DEC
14.0/A converter 13, hybrid transformer 11.
This is because it is an echo component that has passed through the A/D converter 12 and EC 14. Reception input point R! It is delayed compared to the R audio signal. However, the voice detector 22. 21, the moving average power of the receiving input point and the transmitting output point is detected simultaneously as described above. The latter moving average power includes not only the echo component of the B-side speaker. The audio signal of the A-side speaker. When the A-side speaker and the B-side speaker are talking together, the moving average power of their audio signals may also be sent to the ERL estimation section. On the other hand, when the ERL estimation part satisfies the above update conditions. ERL is estimated using these moving average powers. There are problems in that the error in the estimated URL increases, the loop gain increases, quasi-ringing or ringing occurs, and there is a high possibility that the call quality will deteriorate. The purpose of the present invention is to improve call quality by reducing the estimation error of the amount of wraparound cancellation. [Means for solving the problem] Figure 1 shows a block diagram of the principle of the present invention. In the figure, 14.
14" is inserted between the reception input point and reception output point and the transmission input point and transmission output point of the A side and B side in the 4-wire loop connecting the A side and the B side, and the transmission from the reception output point 3 is an echo canceller that cancels the echo passing through the transmission path leading to the input point using an estimated echo value generated internally. 3 uses the input A-side and B-side detour cancellation amount to cancel the transmission on the A-side and B-side. Among the output points, give the gain necessary to compensate for line loss to the amplifier connected to the transmission output point with a higher audio signal level, and give the amplifier connected to the transmission output point with a lower audio signal level. This is a loop gain control means that controls the gain so that no noise occurs in the 4-wire loop for the amplifier that is in the 4-wire loop. Also, 4,4" is the reception input point and transmission output of the A side and B side. A wraparound consumption amount estimating means that detects the audio signal level at a point, estimates and sends the wraparound cancellation amount using the audio signal level detected when the operation enable signal is input,
5.5° is calculated using the delay information sent from the echo canceller from the reception output point to the transmission input point and the audio signal appearing at the reception output point and transmission input point of the A side and B side.
The audio signals at the transmission input points of the A side and B side are the audio signals of only the B side speaker or the A side speaker, or the audio signals of the A side speaker or the B side speaker. Or, it determines which one of the synthesized speech signals of the A-side speaker and the B-side speaker, and when it is determined to be the speech signal of the B-side speaker or the A-side speaker, it sends an operation enable signal, and the other The time of judgment is the call status judgment part that sends out an inoperable signal. [Function] The present invention is provided with a call status determining portion 5.5°.

For example, in this judgment part 5, after considering the delay information of the A-side transmission path from the echo canceller, the reception output point R
. ．． , and the number of zero crossings of the audio signal at the transmission input point Sin (
In other words, the frequencies) are compared. At this time, if the number of 0 crosses matches, it is determined that only speaker B is speaking and the audio signal at the reception output point has been routed to the transmission input point. Turn on the operation of the wraparound cancellation amount estimation means 4. However, if the number of zero crosses cannot be determined to be a match, it is determined that updating the wraparound cancellation amount is inappropriate because the A-side speaker or the A-side speaker and the B-side speaker are talking together. Turn off the 1,000-stage 4 operation for estimating the amount of wraparound cancellation. This reduces the error in estimating the amount of wraparound cancellation and improves call quality. (Embodiment) FIG. 2 is a block diagram of an embodiment of the present invention, and FIG. 3 is an explanatory diagram of the operation of the call state determination portion in FIG. 2. Here, the loop gain control section 31. eight GC amplifiers 32,
32° is a component of the loop gain control means 3, the audio detector 41.
42. The ERL estimation section 43 is a component of the wraparound cancellation amount estimation means 4. Also, the same symbols indicate the same objects throughout the figures. Below, assuming that the audio signal is transmitted from the B side to the A side, the third
The operation shown in FIG. 2 will be explained with reference to the drawings.

still. Since the operations on side A and B are the same, we will mainly explain the operation on side A. First, the audio signal sent from the B-side speaker is converted into a digital signal by a hybrid transformer 11''.
The signal is converted back to an analog signal by the D/^ converter 131 via the C14, and is applied to the receiver via the hybrid transformer 11. However, due to the mismatch of this hybrid transformer and the wraparound at the A-side phone, some of the audio signals are transferred to the A/D
Converter 12 converts it into a digital signal again? and EC
14, is sent to the B side through the AGC amplifier 32.

Here, after a certain amount of time has passed since the call forwarding service was started. Reception output point R. ut, D/A converter 1
3. Hybrid transformer 11. Handset, hybrid transformer 11. A/B converter 12. Transmission input point Si
n and the receiving output point R. ut, D/A converter 13
．． Hybrid transformer 11. The delay time between the A/D converter 12 and the transmission input point Sin is determined by the EC 14.

Therefore, the call status determination section 5 stores the information on the transmission path delay time from the EC in an internal memory (not shown), and also stores the information on the transmission line delay time from the EC in an internal memory (not shown), and also stores the information on the transmission path delay time from the EC within the time t at the reception output point R0■ as shown in FIG. Count the number of O-crosses with a counter (not shown). Moreover, the reception output point R is at the transmission input point Sin. The counter counts the number of O-crosses within time t, which is delayed by the above delay time from the time when the number of 0-crosses was counted in uL. Then, the two count values are compared by a comparator (not shown) to determine whether the two signals are the same or not, and the determination criteria are as follows. In other words, if the following conditions are satisfied: (number of 0 crosses at reception output point Rout - number of margins) <number of 0 crosses at transmission input point Si (number of O crosses at reception output point + number of margins), the two signals are simply It is determined that the signal at the reception output point is the same as that of the looped signal, and the operation enable signal is set as ER.
It is sent to the L estimation part 43. However, if the conditions are not satisfied, it is determined that they are not the same, and an inoperable signal is sent to the ERL estimation section. As shown in FIG. 2, the analog signal is sampled by an A/D converter and digitized, but a digital signal processor DSP (not shown) changes the sign of the sampled value from positive to negative. Alternatively, the time of change from negative to positive is monitored, and this is treated as a 0 cross and counted up by a counter, so that the number of O crosses indicates twice the frequency.

On the other hand, the voice detector 41. 42 is the receiving input point R! as in the conventional example. The moving average power at the Rl transmission output point SouL is sent to the ERL estimation section 43, and this estimation section estimates the ERL while the operation signal from the call status determination section is input, and loops the updated ERL. The signal is sent to the gain control section 31.

Here, ERL estimation requires the use of moving average powers that differ by the delay time as shown in FIG. 3, but this can be done by delaying the output of the voice detector 42.

Now, since the updated HRL from the ERL estimation part 4°, whose operation is controlled by the output of the call state determination part 5' as well as the A side, is also added to the loop gain control part, these can be used to The gains of the GC amplifiers 32 and 32' are controlled to stabilize the loop gain. That is. The error in estimating the amount of loop cancellation is reduced and the call quality is improved. [Effects of the Invention] As explained in detail above, the present invention has the effect of reducing the error in estimating the amount of interference cancellation and improving the call quality.

[Brief Description of the Drawings] Fig. 1 is a block diagram of the principle of the present invention, Fig. 2 is a block diagram of an embodiment of the present invention, Fig. 3 is an explanatory diagram of the operation of the call state determination portion in Fig. 2, Figure 4 is a block diagram of the conventional example, and Figure 5 is an explanatory diagram of the operation of the wraparound consumption estimation part in Figure 4. In the figure, 3 is the loop gain control means, 4.4 degrees is the loop cancellation amount estimation means, 5 and 5 degrees are the call status determination part, and 14.14'' is the echo canceller. Fig. 3

Claims (1)

[Claims] The reception input point (R_i_n) and reception output point (R_o_u_t) of the A side and B side in the 4-wire loop connecting the A side and the B side.
), the transmission input point (S_i_n) and the transmission output point (S_
echo canceller (14, 14
') and using the input A-side and B-side wraparound cancellation amounts, the amplifier connected to the transmission output point with the higher audio signal level among the transmission output points on the A side and B side is A loop that provides the gain necessary to compensate for line loss, and controls the gain of the amplifier connected to the transmission output point with the lower audio signal level so that no noise occurs within the 4-wire loop. In the call transfer device including the gain control means (3), the audio signal level is detected at the reception input point and the transmission output point of the A side and the B side, and the audio signal level detected when the operation enable signal is input is used. Delay information from the reception output point to the transmission input point sent from the echo canceller (14, 14') and the A-side and the audio signal appearing at the reception output point and the transmission input point of the B side, so that the audio signal at the transmission input point of the A side and the B side is an audio signal of only the B side speaker or the A side speaker, or Determine either the voice signal of the A-side speaker or the B-side speaker, or the synthesized speech signal of the A-side speaker and the B-side speaker, and combine it with the audio signal of the B-side speaker or the A-side speaker. An incoming call transfer device characterized in that it is provided with a call state determining section (5, 5') that sends out an operable signal when a determination is made, and sends out an inoperable signal when other determinations are made.