JPH07183856A - Sound signal band compression transmission system - Google Patents

Abstract

PURPOSE:To enable the compression transmission with high compression ratio by extracting plural band components from a sound signal and generating a carrier wave signal composed of the synthesizing signal of two or more different frequency signals. CONSTITUTION:A sound signal is inputted in a sampling circuit 1 and is sampled by a sampling frequency fs. As for this sampled signal, a required band component is extracted in a filter circuit 2 and the signal is inputted in a modulation circuit 3. In the modulation circuit 3, a modulation is performed by the signal where the sine wave and amplitude of fs/6 to be outputted from a carrier wave signal generation circuit 5 is the half and a carrier wave signal to be the sum with the sine wave of fs/2 is extracted, a narrow band signal is extrated by a low-pass filter 4 where a passing band is 0 to fs/6 after a compression is performed for the band of 0 to fs/6. in this case, the carrier wave signal generation circuit 5 may be composed of a first transmitter generating a sine wave signal of a frequency fs/6, a second transmitter generating the sine wave signal of a frequency fs/2 and a synthesizer synthesizing each output signal, for instance.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio signal band compression transmission system for compressing a frequency band of an audio signal and narrowing the frequency band for transmission.

[0002]

2. Description of the Related Art Conventionally, as this type of technology, for example,
As shown in Japanese Patent Laid-Open No. 50-99005, a comb filter is used to extract required band components from an audio signal at equal intervals, and the signals extracted at equal intervals are approximately half the maximum frequency of the audio signal. There is a method of amplitude-modulating a carrier signal having a single frequency, and compressing the bandwidth of a voice signal to about 1/2 for transmission.

[0003]

However, in the above-mentioned conventional technique, in addition to the drawback that the compression rate cannot be increased to more than 1/2, since the comb filter is used, the band extracted from the audio signal is reduced. Since the components can be extracted only at equal intervals and in the same bandwidth, there is a problem that important frequency components such as formants that influence the quality of voice cannot be extracted. Further, the hardware configuration requires a large-scale circuit element such as a multiplier, which causes a problem that the circuit scale becomes very large.

As a result of research on a band compression transmission system of an audio signal, the present inventor extracts a required band component from the audio signal, and modulates a predetermined carrier signal containing two or more different frequencies with the extracted signal. By doing so, we have found that the spectrum of the original speech signal is placed within the required narrow band. In addition, the inventor of the present invention narrows the band by sampling the voice signal at a frequency cycle lower than twice the highest frequency of the voice signal on the transmission side, and the reception side samples the reception signal on the same side as the transmission side. It has been discovered that by sampling at a frequency and adding a zero-value sample to the sampled signal, it is possible to compress a voice signal, perform narrow band transmission, and reproduce with an extremely simple configuration.

The present invention has been made in order to solve the above-mentioned problems of the prior art. The first object of the present invention is to achieve a compression rate of 1/2 or more (1 / N (N is 2 or more). (Integer)) to provide a high compression rate voice signal band compression transmission system. A second object of the present invention is to provide a voice signal band compression transmission system that can extract not only equal intervals and equal bandwidths of band components extracted from a voice signal but also required band components. Further, a third object of the present invention is to provide a voice signal band compression transmission system which is much more economical in circuit scale than the prior art and is economical.

[0006]

In order to achieve the above first and second objects, the present invention requires, for example, modulation by using a plurality of band pass filters having different pass band characteristics. A configuration including means for extracting from the audio signal a plurality of band components that are transited (frequency-arranged) within a narrow band and different from each other and generate a carrier signal composed of a composite signal of two or more different frequency signals; It was done. As a result, by modulating using a carrier signal composed of a composite signal of two or more different frequency signals, the component due to modulation and the component due to intermodulation can be concentrated within a required narrow band, and the band compression rate can be reduced. It is possible to increase it to more than 1/2. Further, the band components extracted from the audio signal are not necessarily limited to the equal intervals and the equal bandwidths as in the conventional case, so that the important frequency components that influence the quality of the audio can be surely extracted.

Further, in order to achieve the third object, a first sampling means for sampling a voice signal at a frequency lower than twice the highest frequency of the voice signal is provided on the transmitting side, and the sampling means. A first extracting means for extracting a plurality of band components from the extracted signal and a first converting means for converting the extracted signal into an analog signal, and the receiving side receives the received analog signal as described above. Second sampling means for sampling at the same frequency as the first sampling means, interpolation means for interpolating 0 sampled values in the sampled signal, and a plurality of bands from the output signal of the interpolation means The second extraction means for extracting the component and the second conversion means for converting the extracted signal into an analog signal to restore the audio signal are provided. As a result, as will be described later, it is possible to reduce the modulation circuits and the like of the transmission unit and the reception unit, which are required in the conventional voice band compression method, and it is possible to greatly reduce the amount of hardware.

[0008]

[Embodiment] According to Shannon's sampling theorem, the original signal can be reproduced from the sampled value if it is sampled at a frequency twice or more the highest frequency of the original signal. Therefore, unless otherwise specified, the sampling frequency is f.
When s, the maximum frequency of the input audio signal is fs / 2 or less. In addition, the band compression ratio is 1 / N (N is 2
The above integer), and taking N = 3 as an example,
An example of the present invention will be described.

First, a plurality of required band components are extracted from the input audio signal by, for example, band pass filters having different pass band characteristics. The carrier signal modulated by the extracted signal has, for example, a frequency of fs /
2N, that is, a composite signal of a sine wave of fs / 6 and a sine wave having an amplitude of 1/2 and a frequency of fs / 2. If the frequency of the audio signal is fv, the audio signal fv
The signal obtained when the carrier signal is modulated by (i.e., when the sound signal is multiplied by the carrier signal) contains the following four components, as is clear from the modulation theory. fs / 6-fv fs / 6 + fv fs / 2-fv fs / 2 + fv That is, in the carrier signal, a fs / 6 ± fv spectrum is obtained by modulating the fs / 6 frequency component and the audio signal fv. On the other hand, by modulating the frequency component of fs / 2 and the audio signal fv, a spectrum of fs / 2 ± fv can be obtained.

This is shown in FIG. 1A shows the spectrum of the audio signal fv, FIG. 1B shows the spectrum of the above fs / 6-fv, and FIG.
The spectrum of the component in which the 6-fv component is folded back within the band of 0 to fs / 2, (d) is the above fs / 6 + fv spectrum, and (e) is the cross modulation of the fs / 6 + fv component with the sampling frequency fs. , The spectrum of the component folded back within the band of 0 to fs / 2, (f) is the spectrum of fs / 2 ± fv, and (g) is the fs / 2.
The ± fv component represents the spectrum of the component folded back by the cross modulation with the sampling frequency fs.

As can be seen from this figure, all the components of the sub-bands (1), (2) and (3) of the original voice signal are 0 to f at the same level.
It is folded back into the narrow band of s / 6. However, in the case of this example, in order that all the sub-bands (1), (2) and (3) are folded back at the same level, the carrier signal is a sine wave of fs / 6 and an amplitude of fs of 1/2 of that. It is necessary to make a composite signal with a sine wave of / 2. In addition, the voice signal f
v (0 to fs / 2) is within 1/3 narrow band (0
~ Fs / 6), the plurality of band components extracted from the audio signal by the filter means are:
When they are arranged in the narrow band of 1/3 after modulation, it is necessary to have band components that do not overlap with each other (different from each other).

Next, FIG. 2 shows an example of band components extracted from this audio signal. (A) shows an arrangement example of bands A, B, C, and D extracted from the audio signal. For example, the audio signal component is 0 to 3600 Hz, and the band A is 300 to 700 H.
z, band B is 1200 to 1500 Hz, band C is 210
It is set to 0 to 2400 Hz and the band D is set to 3100 to 3300 Hz, and is extracted from the audio signal by the four band pass filters each having the above pass band. Further, (b) is a component of the band A that has been changed by the modulation of the fs / 6-fv (corresponding to (c) of FIG. 1), and (c) is the fs / 6-fv.
Components of bands B and C that have changed due to the modulation of
1) and (d) represent the components of band D (corresponding to (f) and (g) in FIG. 1) that have been changed by the modulation of fs / 2 ± fv.

As shown in FIG. 2, the band components of A, B, C, and D extracted from the audio signal are 0 to fs / 6 (0 to 0) as a result of modulating the carrier signal which is the composite signal of the sine wave.
They are rearranged in a narrow band (1200 Hz) that does not overlap each other. Therefore, the original voice signal (0 to
This means that the band has been compressed to 1/3 compared to 3600 Hz). It should be noted that FIG. 2 shows an example of band components extracted from the audio signal.
C and D are not necessarily limited to equal intervals and equal bandwidths.
Of course, if frequency components that are important for the quality of speech are obtained at equal intervals and equal bandwidths, it goes without saying that they may be extracted at equal intervals and equal bandwidths. Is possible.

FIG. 3 shows the overall configuration of the present invention. The audio signal is input to the sampling circuit 1 and the sampling frequency f
is sampled at s. The sampled signal is filtered by the filter circuit 2 to extract a required band component and input to the modulation circuit 3. In the modulation circuit 3, the sine wave of fs / 6 output from the carrier signal generation circuit 5 and the amplitude of 1/2 are f
A carrier signal, which is the sum of s / 2 sine waves, is modulated by the extracted signal, compressed to a band of 0 to fs / 6, and then a low pass filter 4 having a pass band of 0 to fs / 6. A narrow band signal is extracted by. The carrier wave signal generation circuit 5 includes, for example, a first oscillator that generates a sine wave signal having a frequency fs / 6, a second oscillator that generates a sine wave signal having a frequency fs / 2, and the first oscillator. This can be easily realized by using a combiner that combines the output signals of the second oscillator.
Further, a value obtained by sampling the carrier wave signal at the sampling frequency fs in advance is stored in a storage means such as a ROM and the carrier data is generated by sequentially reading the stored data, which is extremely simple. A carrier signal generation circuit can be realized with the configuration.

Next, a specific example of the carrier signal in the present invention will be described with reference to FIGS. 4 and 5. Figure 4 (a)
, The signal represented by the black dots is a discrete sine wave signal with an amplitude of 1 and a frequency of fs / 6, and in (b), the signal represented by the black dots is a discrete sine wave signal with an amplitude of 1/2 and a frequency of fs / 2. The sine wave signal, (c) is the sum signal of (a) and (b). As shown in this figure, the sum of the sine wave of fs / 6 and the sine wave of fs / 2 whose amplitude is 1/2 is +1, +1/2, +1, -1, -1 / at each sampling point. 2, -1, +
It becomes a repetitive signal of 1 ...

Similarly, FIG. 5A shows fs of amplitude 2/3.
A discrete sine wave signal with a frequency of / 6, (b) is a discrete sine wave signal with a frequency of fs / 2 having an amplitude of 1/3, and (c) is a signal of the sum of (a) and (b). As shown in this figure, fs
The sum of the sine wave of / 6 and the sine wave of fs / 2 whose amplitude is 1/2 is 0, +1, 0, 0, -1, at each sampling point.
It becomes a repeating signal of 0, 0, +1, 0, .... The waveform of the carrier used in the above description is

[0017]

[Equation 1]

Although it has a form such that it is an odd number of N = 3 or more,

[0019]

[Equation 2]

In the general formula, N = 3 is set. Therefore, as a carrier wave,
Even if a generalized formula such as the formula (1) is used, the same effect as the above description can be obtained. In addition, even if the carrier represented by the equation (2), (3), or (4) below is used, a band with a high compression rate of 1 / N or more (N is an integer of 2 or more) is similarly obtained. It can be compressed.

[0021]

[Equation 3]

As a special case, the above (1),
In equations (2), (3), and (4), φ _k = 0 (k = 0, 1 ·
··)think of. If the value of the above equation (1) is S ₁ ,

[0023]

[Equation 4]

Therefore, S ₁ is: when t = 0, when S ₁ = N / 2 t = n / fs (n = 1, 2, ..., N-1),
When S ₁ = 0 t = N / fs (since N = odd), S ₁ =
-N / 2 t = n / fs (n = N + 1, N + 2, ..., 2N-
In the case of 1), S ₁ = 0 If the amplitude is normalized to 1, the carrier S ₁
Is a sampling time point t = 0, N / fs, 2N / fs, 3N / fs ..., and takes values of +1, -1, +1, ..., and is 0 at other sampling time points t = n / fs. Takes the value of. Similarly, when φ _k = 0, (2), (3),
If the values of equation (4) are S ₂ , S ₃ , and S ₄ ,

[0025]

[Equation 5]

Therefore, S ₂ is, when t = 0, S ₂ = N / 2 t = n / fs (n = 1, 2, ..., N-1),
When S ₂ = 0 t = N / fs (since N = even), S ₂ =
-N / 2 t = n / fs (n = N + 1, N + 2, ..., 2N-
When 1), S ₂ = 0

[0027]

[Equation 6]

Therefore, S ₃ is, when t = 0, S ₃ = N / 2 t = n / fs (n = 1, 2, ..., N-1),
When S ₃ = 0 t = N / fs (since N = odd number), S ₃ =
N / 2 t = n / fs (n = N + 1, N + 2, ..., 2N-
In the case of 1), S ₃ = 0

[0029]

[Equation 7]

Therefore, S ₄ is as follows: When t = 0, S ₄ = N / 2 When t = n / fs (n = 1, 2, ..., N-1),
When S ₄ = 0 t = N / fs (since N = even), S ₄ =
N / 2 t = n / fs (n = N + 1, N + 2, ..., 2N-
In the case of 1), S ₄ = 0

As described above, since the carrier signal in the present invention is composed of repetitive signals of very simple numerical values, the so-called multiplier or the like is used as the structure of the modulator for multiplying the audio signal extracted by the filter and the carrier signal. A large-scale element is not necessary, and it can be realized by using a microcomputer, for example, and the circuit scale can be reduced. In the above-described embodiment, the case where the voice band compression signal to be transmitted is a digital signal has been described. For example, in the embodiment of FIG. 3, by providing a D / A converter after the filter 4, the above transmission is performed. The signal can be analogized, and the present invention can be applied to both the digital transmission line and the analog transmission line.

Next, another embodiment of the present invention in which the circuit structure is further simplified and the circuit scale is remarkably reduced as compared with the above-mentioned embodiment will be described below. 6 and 7 show an embodiment of the present invention in which the circuit configuration is simplified as described above. FIG. 6 is a block diagram showing the configuration of the transmitting side, and FIG. 7 shows the configuration of the corresponding receiving side. It is a block diagram shown. In FIG. 6, 11 is a voice signal input terminal, 12 is an analog-digital converter, 13 is a digital filter, 14 is a digital-analog converter, 15
Is an output terminal, 16 is a sampling signal generating circuit, and 17 is a frequency divider. Further, in FIG. 7, 21 is a reception input terminal, 22 is an analog-digital converter, 23 is a 0 sample interpolation circuit (INS circuit), 24 is a digital filter, 2
5 is a digital-analog converter, 26 is a restored audio output terminal, 27 is a sampling signal generating circuit, and 28 is a frequency divider.

First, the audio signal input from the audio signal input terminal 11 is digitally sampled by the analog-digital converter 12. The sampling frequency fs is set to be twice or more the highest frequency f _m included in the audio signal. Next, the digital filter 13 extracts only some sub-band components, as shown in FIG. The output of the digital filter 13 is an example of a signal having frequency redundancy handled by the present invention.

The feature of this embodiment is that it is included in the following digital-to-analog converter 14, in which the sampling frequency of this digital-to-analog converter 14 is a value lower than twice the highest frequency of the audio signal. The main point of the present embodiment lies in the place set to. That is, in the conventional sampling, when the maximum frequency of the audio signal is f _m in order to reproduce the original signal from the sampled signal, the sampling frequency f
s is a frequency more than twice the maximum frequency f _m (that is,
Although fs ≧ 2f _m ), in the present embodiment, the sampling frequency fs is set to a frequency lower than twice the highest frequency f _m of the audio signal (fs <2f _m ), and the audio signal band is compressed and transmitted. At the same time, the receiving side is configured to reproduce the original audio signal by using the sample interpolating means, thereby greatly simplifying the circuit scale of the entire transmitting / receiving circuit.

In this way, the output of the digital filter 13
From the sample, the highest frequency f of the audio signal _mLess than twice the
A sample is selected at the sampling frequency, and this is digitally sampled.
The analog converter 14 converts the analog signal. now,
Samples are selected from the output of the digital filter 13 at Δt second intervals.
Let's start. f₀= 1 / Δt is the sampling frequency
Therefore, the maximum frequency of the audio signal is f_mThen, f₀<2
f_mIs. This sampling frequency f₀Timing signal
No. is, for example, as shown in FIG.
The sampling timing signal of the data 12 is divided by the frequency divider 17.
It can be easily obtained by dividing by a predetermined frequency.

[0036] The audio signal _{Acos (2πf υ t + φ υ} )
Let S _n be the sample value at time t = nΔt (n = 0, 1, 2, ...), S _n = Acos (2πf _υ nΔ
t + a _φ υ), m, when the n is an integer, cos (because it is _{2πmf o n △ t) = cos} (2πmn) = 1 sin (2πmf o n △ t) = sin (2πmn) = 0, S n _{= Acos (2πf υ n △ t} + φ υ) cos (2π
mf _o n Δt) + A sin (2πf _υ nΔt + φ _υ ) s
_{in (2πmf o n △ t)} = Acos {2π (f υ -m
f _o ) _nΔt + φ _υ }

When the sampled value S _n is converted into an analog signal by the digital-analog converter 14, its output cannot include frequency components _equal to or more than 1/2 of the sampling frequency fo, so its output S is 0 ≦. f _upsilon <when _{f o / 2, m = 0} S = Acos (2πf υ t + φ υ) f o / 2 ≦ f υ < when _{3f 0/2, m = 1} S = Acos {2π (f υ -f ₀ ) t + φ _υ } 3f _o / 2 ≦ f _υ <5f _o / 2, m = 2 S = Acos {2π (f _υ −2f ₀ ) t + φ _υ } (same below).

That is, the frequency spectrum 0 of the original signal
˜f _m is folded between 0 and f _o / 2 and converted into a narrow band signal. The original signal is the digital filter 13
Is divided into several sub-band components by
By choosing f ₀ , these sub-band components can range from ₀ to f ₀ /
It is possible to arrange them so that they do not overlap each other in the two zones.

FIG. 8 is a diagram showing an example of the frequency spectrum in this embodiment. In the figure, (a) shows the spectrum of the original voice signal, and (b) shows the spectrum of the voice signal sampled at the frequency fc. Further, (c) represents the output spectrum of the digital filter. In FIG. 8, it is divided into three pass bands, but it is sufficient that the signal spectrums do not overlap with each other in the subsequent sub-sampling, and the frequency characteristics as in this case are not necessarily required. Next, (d) shows the spectrum when the sub-sampling frequency f ₀ is now fc / 2. Further, the signal shown in (e) is a signal that includes all the components of the passing signal having the characteristics shown in (c), although it is folded back, so that if it is the transmitting side, the signal in this band is transmitted. Send it out as a signal. That is, in the case of this example, the band is compressed to 1/2.

Next, description will be made on the receiving side. Receive input terminal 2
The received narrow band signal input from 1 is digitally sampled by the analog-digital converter 22. The sampling frequency is set equal to the sampling frequency f ₀ on the transmission side. The timing signal that determines the sampling frequency can be easily obtained by dividing the sampling signal of the digital-analog converter 25 by a frequency divider 28 by a predetermined frequency as shown in FIG. The received narrow band signal is Acos {2π (f _υ −m
f ₀₎ and _t + φ υ}, the sample value at time t = n △ t When _{S n0, Sno = Acos {2π} (f υ -mf o) n △ t +
φ _υ }

Next, in the 0-sample interpolation circuit 23, K zero-value samples S _n1 , S _n2 , ..., S _nk are inserted between each sample. That is, S _nk = 0 (k = 1, 2, ... K) Therefore, S ₀₀ , S ₀₁ ,
S ₀₂ , ..., S _ok , S ₁₀ , S ₁₁ , S ₁₂ , ...,
Sampling sequences such as S _1k , S ₂₀ , S ₂₁ , ... _Are added. S _nk is time t = _nΔt + {k / (k + 1)} Δ
It is a sample value at t. Also, if the sampling frequency is F ₀ , then F ₀ = (K + 1) f ₀ . By the way, if K is an even number, put L = K / 2, ω ₀ = 2πf _0, and

[0042]

[Equation 8]

Calculate Sin (ω ₀ t / 2) on both sides
If you multiply

[0044]

[Equation 9]

Here, k = 0, 1, 2, ... K, t
= NΔt + {k / (k + 1)} Δt, ω ₀ Δt =
Since it is 2π, according to the equation (11) of the above equation 9, when k = 0, R _E = K + 1 when k ≠ 0, R _E = 0 Similarly, when K is an odd number, L = (K−1) / 2 , Ω ₀ = 2
Put it as πf ₀ ,

[0046]

[Equation 10]

Calculate Sin (ω ₀ t / 2) on both sides
If you multiply by,

[0048]

[Equation 11]

Here, k = 0, 1, 2, ... K, t
= NΔt + {k / (k + 1)} Δt, then ωoΔt =
Since it is 2π, according to the equation (14) of Equation 11, when k = 0, when Ro = K + 1 k ≠ 0, and when Ro = 0 or more is used, the sampled value S _nk is K = even _{If, S nk = Acos {2π (} f υ -mf 0) n △ t + φ υ} × R E / (K + 1) = Acos {2π (f υ -mf 0) t + φ υ} × R E / (K + 1) ·· (15) When K = odd S _nk = A cos {2π (f _υ −mf ₀ ) nΔt + φ _υ } × Ro / (K + 1) = A cos {2π (f _υ −mf ₀ ). t + φ _υ } × Ro / (K + 1) ..... (16) can be written.

When K = even, S _nk has a form in which the received narrow band signal is multiplied by R _E / (K + 1). Since 1 / (K + 1) is just an amplitude coefficient, considering R _E ,

[0051]

[Equation 12]

It is equivalent to modulating with a carrier wave.
Therefore, S _nk is cos (lω ₀ t) [l =
1, 2, ..., L] is a signal in which upper and lower sidebands are superimposed, that is, a signal folded between ₀ and f 0/2 on the transmitting side is expanded again between ₀ and f _m. . Therefore, if this signal is passed through the filter 24 to extract only the desired sub-band component and then restored to an analog signal by the digital-analog converter 25, the original signal is reproduced. Also, when K = odd number is almost the same, R ₀ is (1
2) Take the form of formula.

The first and second terms on the right-hand side of the equation (12) are the same as in the case of K = even number, but for the third term, the modulation frequency (L + 1) ω ₀ / 2π = K + 1 /
2 is f ₀ = F _0/2, side bands of the modulated product overlaps the lower sideband folded by sampling frequency F _0,
The amplitude doubles. This corresponds to the fact that the second term on the right side of the equation (12) has the coefficient 2 but the third term lacks it, and as a result, the modulation by the second term is caused. The modulation according to the third term has the same level.

The operation on the receiving side in this embodiment described above will be supplementarily described with reference to the spectrum chart of FIG. On the receiving side, the received signal (e) is (f ₀ )
That is, when AD conversion is performed at the sampling frequency of fc / 2, the signal (d) is reproduced. The same spectrum is obtained even if a 0 value is inserted in the 0 sample interpolation circuit 23 for this signal, and when this signal is passed through the digital filter 24, the signal shown in (c) is extracted. A voice signal can be restored by DA converting this signal into an analog signal. The restored audio signal is not exactly the same as the original signal shown in FIG.
Even if limited to a specific frequency component as shown in (c), the clarity does not deteriorate so much. In the above example, the compression rate is 1
However, if f _{0 is set} to a much lower frequency, even greater compression is possible. However, in that case,
As described above, it is necessary to prevent the spectra from overlapping when the characteristics of the filter are subsampled.

As described above, in this embodiment, the voice band compression with a high compression rate is possible with the extremely simple structure as shown in FIGS. 6 and 7. Therefore, it is possible to reduce all of the modulation circuit and low-pass filter on the transmission side and the modulation circuit on the reception side, which were conventionally indispensable in this kind of voice band compression method, and the circuit scale is greatly reduced and the economic efficiency is significantly improved. Can be realized.

[0056]

As described above, according to the present invention, the signal processing process for rearranging a voice signal can be simplified and an economical device configuration can be realized. In other words, when performing the product of the extracted audio signal and carrier in the modulator, it is only necessary to multiply by a very simple numerical value as described above, so a large-scale element called a so-called multiplier is not required. Instead, simple hardware, such as a microcomputer and a ROM, can be easily implemented. In order to further reduce the amount of hardware, the audio signal is sampled at the transmitting side at a frequency cycle lower than twice the highest frequency of the audio signal, and the receiving signal is sampled at the receiving side as the same as the transmitting side. By sampling at a frequency and adding a zero-value sample to the sampled signal, it is possible to realize a voice signal band compression transmission system with a significantly reduced circuit scale and excellent economy.

Since the band can be compressed at a high compression rate,
It is possible to transmit N times as many channels (N is an integer of 2 or more) as compared with the case without compression. Furthermore, in the case of digital transmission, the transmission rate can be reduced to 1 / N, and in the case of a fixed transmission rate, N times as many channels of voice can be transmitted, and the transmission delay time can be reduced to 1 / N. Is also possible, and the effect is immeasurable. Further, the band components extracted from the voice signal are not necessarily limited to the equal intervals and the same bandwidth as in the conventional case, so that the important frequency component that influences the quality of the voice can be surely extracted. In addition, the present invention can be applied to both digital transmission lines and analog transmission lines, and its application range is extremely wide.

[Brief description of drawings]

FIG. 1 is a diagram showing an example of a spectrum when an audio signal is modulated according to the present invention.

FIG. 2 is a diagram showing an example of band division of an audio signal according to the present invention.

FIG. 3 is a block diagram showing an embodiment of a transmission circuit of the present invention.

FIG. 4 is a waveform diagram showing an example of a carrier signal used in the present invention.

FIG. 5 is a waveform diagram showing another example of a carrier signal used in the present invention.

FIG. 6 is a block diagram showing another embodiment of the transmission circuit according to the present invention.

7 is a block diagram showing an embodiment of a receiving circuit corresponding to the transmitting circuit of FIG.

FIG. 8 is a diagram showing an example of a spectrum in the embodiment of FIGS. 6 and 7.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Sampling circuit 2 Filter circuit 3 Modulation circuit 4 Low pass filter 5 Carrier wave signal generation circuit 6,16,27 Sampling signal generation circuit 11 Audio signal input terminal 12,22 Analog digital converter 13,24 Digital filter 14,25 Digital analog converter 15 Output terminal 17, 28 Frequency divider 21 Receive input terminal 23 0 Sample interpolation circuit 26 Restored audio output terminal

Claims (8)

[Claims] 1. Extracting a required band component from an audio signal,
In a voice signal band compression transmission system in which a carrier signal is modulated with the extracted signal and the modulated signal is transmitted, band components extracted from the voice signal are different from each other within a narrow band required by the modulation. A voice signal band compression transmission system characterized in that a plurality of frequency-arranged band components are selected and the carrier signal is a composite signal of two or more different frequency signals. 2. The voice signal band compression transmission system according to claim 1, wherein the voice signal is sampled at a sampling frequency which is at least twice the highest frequency of the voice signal, and then the plurality of band components are sampled from the sampled signal. And the carrier signal is (N-2) / 2N of the sampling frequency.
A first sine wave signal having a frequency of (N is an odd number of 3 or more) and a second sine wave having an amplitude of ½ of the first sine wave and ½ of the sampling frequency. A voice signal band compression transmission method characterized by comprising a synthesized signal with a signal. 3. The voice signal band compression transmission system according to claim 1, wherein the voice signal is sampled at a sampling frequency which is at least twice the highest frequency of the voice signal, and then the plurality of band components are sampled from the sampled signal. And the carrier signal is (N-1) / 2N of the sampling frequency.
A voice signal band compression transmission system comprising a composite signal of two sinusoidal signals having frequencies (N is an even number). 4. The voice signal band compression transmission system according to claim 1, wherein the voice signal is sampled at a sampling frequency which is at least twice the highest frequency of the voice signal, and then the plurality of band components are sampled from the sampled signal. And the carrier signal is (N-1) / 2N of the sampling frequency.
A sine wave signal having a frequency (N is an odd number of 3 or more),
A voice signal band compression transmission system characterized by comprising a synthesized signal with a DC signal having an amplitude of 1/2 of the sine wave signal. 5. The voice signal band compression transmission system according to claim 1, wherein the voice signal is sampled at a sampling frequency that is at least twice the highest frequency of the voice signal, and then the plurality of band components are sampled from the sampled signal. And the carrier signal is (N-2) / 2N of the sampling frequency.
A sine wave signal having a frequency (N is an even number of 4 or more),
A voice signal band compression transmission system characterized by comprising a synthesized signal with a DC signal having an amplitude of 1/2 of the sine wave signal. 6. The voice signal band compression transmission system according to claim 4 or 5, wherein the relative amplitude value at each sampling point of the carrier signal is +1, 0 ... 0, +1, 0 ...
A voice signal band compression transmission system characterized by being repeated 0. 7. The voice signal band compression transmission system according to claim 2 or 3, wherein the relative amplitude value at each sampling point of the carrier signal is +1, 0 ... 0, -1, 0 ...・
A voice signal band compression transmission system characterized by being repeated 0. 8. A first side for sampling a voice signal on the transmitting side.
Of sampling means, a first extracting means for extracting a plurality of band components from the sampled signal, and a first converting means for converting the extracted signal into an analog signal. Is a second sampling means for sampling the received analog signal, an interpolating means for interpolating 0 sampled values in the sampled signal, and a plurality of band components from the output signal of the interpolating means. A voice signal band compression transmission system having a second extracting means for extracting and a second converting means for converting the extracted signal into an analog signal to restore a voice signal, the first and second samples The audio signal band compression transmission system, wherein the sampling frequency of the conversion means is a frequency lower than twice the highest frequency of the audio signal.