JPH0237852A - Voice secret call device - Google Patents

Abstract

PURPOSE:To improve the signal-to-noise ratio of a reproducing sound signal and to improve secrecy on a transmission path by inverting the spectrum of a transmitting sound signal on a transmitting side and restoring the spectrum of a signal extracted by an adaptive echo canceler on a receiving side. CONSTITUTION:The title device is equipped with a spectrum inverting circuit 2 to output an inverted sound signal into which the spectrum of the sound signal to be transmitted is inverted and a noise generating circuit 3 to output a noise signal in a frequency band occupied by the sound signal, and a secret call signal into which the inverted sound signal and the noise signal are added is transmitted to a transmission path 10. The secret call signal is guided from a receiving side input terminal 11 to the input terminal of an adaptive echo canceler 13, on the other hand, the output signal of a reference noise generating circuit 12 is supplied to the other input terminal of the adaptive echo canceler 13 as a reference noise signal. The spectrum of the signal extracted by the adaptive echo canceler 13 is restored by a spectrum restoring circuit 14 and outputted from a receiving side output terminal 15. Thus, the signal-to-noise ratio of the extracted sound signal can be improved and the secrecy on the transmission path 10 can be enhanced.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a voice polarization device, and in particular, on the transmission side, transmits a privacy signal in which a noise signal is superimposed on a transmitted voice signal,
The present invention relates to a voice confidential communication device that extracts an audio signal from a received confidential voice signal on a receiving side.

[Conventional technology]

A conventional voice privacy device includes a transmission noise generation circuit that periodically generates a noise signal with a substantially flat power spectrum and the same waveform within the entire frequency band of the transmission audio signal, and a transmission It has an addition circuit that adds the transmission noise signal of the noise generation circuit and the transmission audio signal, and sends out the added signal to the transmission path as a secret signal. The receiving side has a reference noise generating circuit that has the same function as the transmitting noise generating circuit on the transmitting side, and an echo canceller that extracts the transmitted voice signal from the confidential signal by referring to the reference noise signal. Ta.

[Problem to be solved by the invention]

In order to polarize the transmitted voice signal, the conventional voice encrypting device described above maintains the single-tone intelligibility of the transmitted voice signal included in the confidential signal at 0% when noise is added to the transmitted voice signal. The signal-to-noise ratio is set to -20 dB. At this time, if the reception side uses an echo canceller with a cancellation amount of 35 dB to cancel the noise, the signal-to-noise ratio of the extracted audio signal is as small as about +15 dB, resulting in a very audible signal. In order to improve the signal-to-noise ratio of this extracted audio signal, it is said that if the signal-to-noise ratio between the transmitted audio signal and noise on the transmitting side is changed to -10 dB, etc., the secrecy on the transmission path will be reduced. There are drawbacks.

[Means to solve the problem]

The voice secret communication device of the present invention includes a spectrum inversion circuit that outputs an inverted voice signal by inverting the spectrum of a voice signal to be transmitted, a noise generation circuit that outputs a noise signal within a frequency band occupied by the voice signal, and the inverted voice an addition circuit that transmits a secret signal obtained by adding the signal and the noise signal to a transmission line; and an adaptive echo canceller that refers to the noise signal and extracts the received inverted voice signal included in the secret signal from the transmission line. and a spectrum restoration circuit for restoring the spectrum of the received inverted audio signal.

〔Example〕 Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram of one embodiment of the present invention.

The present embodiment shown in FIG. 1 includes a transmission side input terminal 1 for inputting a transmission audio signal to a voice privacy device, a spectrum inversion circuit 2 for inverting the spectrum of the transmission audio signal, and a noise signal having the same waveform periodically. a transmission noise generation circuit 3; an addition circuit 4 that adds the spectrally inverted audio signal from the spectrum inversion circuit 2 and the noise signal from the transmission noise generation circuit 3 to output a confidential signal; and a transmission side output of the audio confidential communication device. Terminal 5 and f transmission path 10, the receiving side input terminal 11 of the voice privacy device, the reference noise generation circuit 12, the adaptive echo canceller 13, and the spectrum for restoring the spectrum of the signal extracted by the adaptive echo canceller 13. Restoration circuit 1
4, and a receiving side output terminal 15 of the voice confidential communication device.

FIG. 2 is a signal waveform diagram for explaining the operation of this embodiment. The operation will be described below with reference to FIGS. 1 and 2.

FIG. 2 (a) is an example of the waveform of a transmitted audio signal. This signal is supplied to the spectrum inversion circuit 2 via the transmission side input terminal 1 in FIG. From Fig. 2(a), (
In e), the horizontal axis represents time and the vertical axis represents the amplitude of each signal, corresponding to the time in FIG. 2(f). Furthermore, the section from time tJ-1 to time 1 in FIG. 2 (f>) is expressed by the subscript "j-1°".

FIG. 2(b) shows the output signal of the transmission noise generation circuit 2. The frequency spectrum is approximately flat over the frequency band targeted for voice signals (for example, 300 Hz to 3400 Hz), and the noise signal is N, - This is a periodic signal that generates a waveform of 1 repeatedly in each section. The adder circuit 4 in FIG. 1 simply adds the transmitted audio signal shown in FIG. 2(a) from the transmitting side input terminal 1 and the transmitted noise signal shown in FIG. 2(b).The signal appears at the output end of the adder circuit 4. is called a secret signal and is shown in FIG. 2(C).

When adding the signals in this adder circuit 4, the level difference between the level S (dB) of the transmitted audio signal included in the secret signal and the level N (dB) of the noise signal is calculated by 20 dB (-Finally, the S/N
= -20 dB). This value masks the noise signal or the transmitted voice signal and maintains the phonetic intelligibility of the transmitted voice signal included in the confidential signal at 0%. In other words, a normal audio signal has an S/N of -20
At dB, short sound intelligibility deteriorates to 0%. Therefore, it is extremely difficult to recognize even the intonation of a low-level voice signal compared to the secret speech signal. Furthermore, if this secret speech signal is divided by a suitable bandpass filter, it is extremely difficult to recognize the presence of a voice signal in the same manner as described above.

In FIG. 1, this private communication device is led to a receiving side input terminal 11 via a transmission line 10. Here, for ease of understanding, the description will be made assuming that there is no delay time and linear distortion of the transmission path, and no frequency deviation between the transmitted signal and the received signal due to modulation/demodulation of the carrier system.

The confidential signal is guided from the receiving side input terminal 11 to the input terminal of the adaptive echo canceller 13. On the other hand, it is assumed that the output signal of the reference noise generation circuit 12 and the output signal of the transmission noise generation circuit 3 have the same repetition period, time relationship, and frequency spectrum. The output signal of this reference noise generation circuit 12 is supplied to the other input end of the adaptive echo canceller 13 as a reference noise signal. This reference noise signal is shown in FIG. 2(d).

Here, the functions of the adaptive echo canceller 13 will be briefly described (Reference: Echo Canceller Technology 61.12.2
0 Published by the Industrial Technology Center.

- The onboard adaptive echo canceller consists of a transversal filter with variable tap coefficients and a subtractor, and synthesizes pseudo echoes by passing a reference signal through the transversal filter. The residual error of this pseudo echo is determined from the actual echo using a subtractor, and the tap coefficients of the transversal filter are adaptively modified so that this residual error is minimized. The adaptive echo canceller follows changes in the characteristics of the echo path by adaptively modifying the tap coefficients of the filter to compensate for the linear distortion in the echo path and its fluctuations.

To explain the correspondence between a general adaptive echo canceller and the signals of each part in this embodiment, - The reception signal of the onboard adaptive echo canceller is the output signal of the reference pressure sound generation circuit 12 in FIG. The echo signal is an echo component included in the secret signal introduced into the receiving side input terminal 11 in FIG. Furthermore, the output signals of the reference noise generation circuit 12 in FIG. 1 and the transmission noise generation circuit 2 in FIG. This corresponds to the path from the terminal to the receiving side input terminal 11 in FIG. Such a general adaptive echo canceller is
The tap coefficients of the built-in transversal filter are adaptively adjusted so that the echo component present at its output end is minimized. The adaptive echo canceller 13 shown in FIG. 1 adaptively adjusts the tap coefficients of the built-in transversal filter so that the transmission noise signal component present at the receiving side output terminal 15 is minimized. In this way, the audio signal extracted by the adaptive echo canceller 13 in FIG.
The restored audio signal shown in FIG. 2(e) is supplied to the receiving side output terminal 15 of the audio secret communication device.

Several assumptions have been made above to facilitate understanding, and these points will be explained below.

First, we will discuss absolute delay time as a problem with transmission paths. If there is a delay time in the transmission path 10, the time relationship between the transmission noise signal included in the secret signal in FIG. 2(C) and the reference noise signal in FIG. 2(d) changes. However, regardless of the delay time, this relative time difference is at most the repetition period of the reference noise signal. Therefore, if the time length of the filter provided in the adaptive echo canceller 13 is longer than the repetition period of this reference noise signal, there is no problem at all.

Variation in delay time in the transmission line 10, frequency shift between transmission and reception in the carrier system, frequency response, phase distortion, etc. are all equivalent to linear distortion in the transmission line 10. Therefore, these problems can be solved by adaptively modifying the tap coefficients of the l-lanceversal filter of the adaptive echo canceller 13 so as to minimize the noise signal included in the reproduced audio signal shown in FIG. 2(e). Ru.

Next, although the repetition period, time relationship, and frequency spectrum of the output signals of the transmission noise generation circuit 3 and the reference noise generation circuit 12 in FIG. It has the same properties as the linear distortion in , and should be solved.

There are two differences in the frequency spectrum.

One is the difference in bandwidth. If the frequency included in the transmission noise signal is not included in the frequency band of the reference noise signal, the frequency component not included will appear as a residual signal at the output of the adaptive echo canceller 13, and as shown in FIG. ) will be mixed into the reproduced audio signal. However, it is easy to set the frequency band of the reference noise signal to be higher than the frequency band of the transmission noise signal, and there is no particular problem. Second, the difference in frequency spectra caused by the frequency responses of the reference noise signal and the transmission noise signal is similar to the linear distortion in the transmission line 10 and does not pose a problem.

However, it goes without saying that this difference in frequency spectrum and the influence of the frequency response on the transmission path 10 causes a difference in frequency spectrum between the transmitted audio signal and the reproduced audio signal.

Next, the parameter for secret deciphering called -B key will be explained. The key to this audio secret communication device is shown in Figure 2 (b).
This can be obtained by simultaneously changing the repetition period of the transmission noise signal in FIG. 2 ((1)) and by changing the waveforms of the transmission noise signal and the reference noise signal. It is possible to arbitrarily change the time length within the time length, and it can be arbitrarily changed within a range that makes the frequency spectra of the transmission noise signal and the reference noise signal approximately flat.

Note that in this embodiment, the spectrum inversion circuit is
Although it is set at position 2 in the figure, it is also conceivable to install it at the subsequent stage of the adder circuit 4.

〔Effect of the invention〕

As explained above, the present invention includes a spectrum inversion circuit that inverts the spectrum of a transmitted audio signal on the transmitting side, and a spectrum restoration circuit that restores the spectrum of the signal extracted by the adaptive echo canceller on the receiving side. By adding , the signal-to-noise ratio of the reproduced audio signal extracted on the receiving side is improved while maintaining the single-tone intelligibility of the transmitted audio signal included in the confidential signal on the transmission path to 0%, Moreover, there is an effect of increasing confidentiality on the transmission path.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing one embodiment of the present invention, and FIG. 2 is a signal waveform diagram for explaining the operation of this embodiment. DESCRIPTION OF SYMBOLS 1... Transmission side input terminal, 2... Spectrum inversion circuit, 3... Transmission noise generation circuit, 4... Addition circuit, 5...
... Transmission side output terminal, 10... Transmission line, 11... Receiving side input terminal, 12... Reference noise generation circuit, 13...
- Adaptive echo canceller, 14... Spectrum restoration circuit, 15... Receiving side output terminal, a... Transmission audio signal, b... Transmission noise signal, C... Confidential signal, d...
-Reference noise signal, e...playback audio signal, f...section and time.

Claims (1)

[Claims] a spectrum inversion circuit that outputs an inverted audio signal obtained by inverting the spectrum of an audio signal to be transmitted; a noise generation circuit that outputs a noise signal within a frequency band occupied by the audio signal;
an addition circuit that transmits a secret signal obtained by adding the inverted voice signal and the noise signal to a transmission line; and an adaptation circuit that refers to the noise signal to extract the received inverted voice signal included in the secret signal from the transmission line. What is claimed is: 1. A voice secret communication device comprising: a type echo canceller; and a spectrum restoration circuit for restoring the spectrum of the received inverted audio signal.