CN101594123B - Method for establishing equivalent parallel filter, voice transmission method, device and system - Google Patents

Abstract

The invention relates to a method for establishing equivalent parallel filters of a cascade filter, a transmission method of voice data, the equivalent parallel filters and a transmission system of the voice data, wherein the method for establishing the equivalent parallel filters comprises the following steps: establishing a channel model of a cascade vocoder; performing time domain analysis and frequency domain analysis on the channel of the cascade vocoder to obtain accurate amplitude-frequency response and phase-frequency response; according to the amplitude-frequency response characteristic and the phase-frequency response characteristic, the channel of the cascade vocoder is equivalent to the parallel connection of a weak noise interference low-pass filter and a strong interference band-pass filter; and the voice coding and decoding process of the cascade vocoder channel is equivalent to a parallel model for performing low-pass filtering and band-pass filtering on the signal. The invention solves the problem of encrypted voice transmission in a mobile voice service channel, provides a basis for designing a quasi-stable modem of a voice band, and improves the reliability and effectiveness of the system for encrypted voice data transmission.

Description

Method for establishing equivalent parallel filter, voice transmission method, device and system

technical field

The invention relates to digital signal processing technology, in particular to a method for establishing an equivalent parallel filter, a voice transmission method, device and system.

Background technique

With the development of communication technology, mobile communication has become a part of people's daily life. Users pay more attention to communication security issues in special occasions while paying attention to communication quality. Due to the defects in the security mechanism of the existing mobile communication system in terms of encryption objects, encryption range, encryption strength, and encryption decision rights, etc., wireless interception or wiretapping incidents have occurred frequently, causing certain losses to users. In order to solve the problem of voice However, the encrypted data presents random characteristics and no longer has the characteristics of voice signals, so it cannot be directly transmitted on the original mobile voice service channel, thus proposing The transmission scheme of encrypted voice in the mobile network mainly includes two types at present: transmission based on data service channel and transmission based on voice service channel.

Among them, for the transmission scheme based on the data service channel, since the mobile data service channel adopts the feedback retransmission mechanism, the end-to-end delay is relatively large, and it is difficult to meet the real-time transmission requirements of the voice service.

For the transmission scheme based on the voice traffic channel, it is limited by the performance of the mobile vocoder and the cascaded working mode. In this process, the mobile vocoder adopts speech parameter coding, only extracts and transmits the parameters that characterize the key features of the speech signal, and the receiving end synthesizes the speech through these parameters. Although the synthesized speech at the receiving end sounds very similar to the input speech signal, it may be completely different in waveform samples from the original speech waveform. In addition, the mobile voice service adopts the working mode of the cascaded vocoder (Tandem), that is, the sender codes in the mobile terminal MS and decodes in the base station BTS (one coding and one decoding), and the receiver codes in the base station BTS and in the mobile terminal MS The decoding (one encoding, one decoding) cascaded vocoder working mode makes the voice signal go through four (GSM-GSM connection) or two (GSM-PSTN connection) code conversions in the network, which inevitably introduces distortion . Therefore, the cascaded vocoder in the mobile voice traffic channel becomes the bottleneck of this transmission scheme, and it is necessary to find a class of analog signals that can be efficiently and reliably transmitted in the mobile voice traffic channel and can be reliably converted to random data (called quasi-signals). smooth voiceband signal).

Currently, there are mainly voice-like synthesis methods and conventional modulation methods. The quasi-speech synthesis method regards the encrypted data as speech coding parameters, and realizes the conversion of data and analog signals through the existing speech synthesis analysis method. Although the analog signal synthesized by this method is similar to speech signals, the modulation rate is low, only It can reach about 1200bps, and the bit error rate is too high, so it is difficult to meet the zero bit error requirement of the deciphering module at the receiving end. Conventional modulation methods use FSK, QPSK and other modulation methods to realize the conversion of encrypted data and analog signals, but the analog signals synthesized by this method are difficult to meet the requirements of efficient and reliable transmission in mobile voice service channels. The distortion is large, and the demodulation error rate is usually greater than 10e-2, which cannot meet the performance requirements.

During the research and practice of the prior art, the inventors of the present invention found that the transmission of encrypted voice in the mobile voice service channel in the existing implementation methods cannot meet the requirements of efficient and reliable transmission in the mobile voice service channel.

Contents of the invention

Embodiments of the present invention provide a method for establishing an equivalent parallel filter, a voice data transmission method, an equivalent parallel filter, and a voice data transmission system to solve the problem of encrypted voice transmission in a mobile voice service channel and improve the system Transmission efficiency for encrypted voice.

In order to solve the above technical problems, the present invention provides a method for establishing an equivalent parallel filter in an embodiment, the method comprising:

Establishing a channel model of cascaded vocoders;

Perform time-domain analysis and frequency-domain analysis on the channel of the cascaded vocoder to obtain accurate amplitude-frequency response and phase-frequency response;

According to the amplitude-frequency response characteristic and the phase-frequency response characteristic, the channel of the cascaded vocoder is equivalent to a parallel connection of a weak noise interference low-pass filter and a strong interference band-pass filter;

The speech coding and decoding process of the cascaded vocoder channels is equivalent to a parallel model of performing low-pass filtering and band-pass filtering on signals.

Preferably, performing time-domain analysis and frequency-domain analysis on the channel of the cascaded vocoder to obtain accurate amplitude-frequency response and phase-frequency response specifically includes:

Perform time-domain analysis on the channels of the cascaded vocoders to obtain a description of waveform distortion;

According to the description of the waveform distortion, the processing error of the cascaded vocoder channel is equivalent to a parallel noise source of weak noise interference in the low frequency band and strong interference in the high frequency band;

Frequency domain analysis is performed on the channels of the cascaded vocoders to obtain amplitude-frequency response characteristics and phase-frequency response characteristics of the cascaded vocoder channels.

Preferably, the channel model of the cascaded vocoder includes: when the Global System for Mobile Communications GSM-GSM is connected, 4 cascaded full-rate vocoder GFR-4 channel models; and when the GSM-public switched telephone network PSTN is connected , the channel model of 2 cascaded full-rate vocoders GFR-2.

Preferably, the channel model of GFR-4 includes: the vocoder pair of the mobile terminal and the base station at the sending end, the vocoder pair of the base station and the mobile terminal at the receiving end, and all links between the sending end and the receiving end.

Preferably, the GFR-2 channel model includes: all links of the vocoder pair of the mobile terminal and the base station, wireless links, core network transmission links, and PSTN transmission links.

Preferably, the waveform distortion is described in the following ways: non-similarity measure description and similarity measure description.

Preferably, the method further includes: constructing an input signal that can be identified by the channel model of the cascaded vocoder, using the input signal to excite the channel model of the cascaded vocoder at the sending end, and recording the input signal at the receiving end output signal.

The embodiment of the present invention also provides a voice data transmission method, the method comprising:

The equivalent parallel filter receives the encrypted and modulated voice data sent by the sending end;

The equivalent parallel filter performs low-pass and band-pass filtering on the voice data respectively, and sends the filtered data to the receiving end, so that the receiving end can demodulate and decrypt the received data in order to obtain the original voice data.

Preferably, the method further includes: establishing an equivalent parallel filter, and the method for establishing an equivalent parallel filter is as described in any one of claims 1-7.

Correspondingly, an embodiment of the present invention also provides a device for establishing an equivalent parallel filter, including:

Establishing a unit for establishing a channel model of the cascaded vocoder;

An analysis unit, configured to perform time-domain and frequency-domain analysis on the channel of the cascaded vocoder established by the establishment unit, to obtain accurate amplitude-frequency response and phase-frequency response;

The first equivalent unit is used to equivalent the channel of the cascaded vocoder into a weak noise interference low-pass filter and a strong interference band-pass filter according to the amplitude-frequency response characteristics and phase-frequency response characteristics obtained by the analysis unit parallel connection;

The second equivalent unit is used for the speech encoding and decoding process of the cascaded vocoder channel established by the establishing unit to be equivalent to a parallel model of performing low-pass filtering and band-pass filtering on the signal.

Preferably, the analysis unit includes:

The first analysis unit is configured to perform time-domain analysis on the channel of the cascaded vocoder established by the establishment unit to obtain a description of waveform distortion;

The third equivalence unit is configured to equate the processing error of the cascaded vocoder channel to a parallel connection of weak noise interference in the low frequency band and strong interference in the high frequency band according to the description of the waveform distortion obtained by the first analysis unit noise source;

The second analyzing unit is configured to perform frequency domain analysis on the channel of the cascaded vocoder established by the establishing unit, and obtain the amplitude-frequency response characteristic and the phase-frequency response characteristic of the cascaded vocoder channel.

Preferably, the channel model of the cascaded vocoders established by the establishment unit specifically includes: channel models of 4 cascaded full-rate vocoders GFR-4 and 2 cascaded full-rate vocoders GFR-2.

Preferably, it also includes:

A construction unit, configured to construct an input signal that can be recognized by the channel model of the cascaded vocoder.

Correspondingly, an embodiment of the present invention also provides an equivalent parallel filter, including: a receiving unit, an equivalent low-pass filter, an equivalent band-pass filter, and a sending unit, wherein

The receiving unit is used to receive the encrypted and modulated voice data sent by the sending end;

an equivalent low-pass filter, configured to filter the received voice data, and send the filtered voice data to the sending unit;

an equivalent bandpass filter, used to filter the received voice data, and send the filtered voice data to the sending unit;

The sending unit is used to add the received voice data filtered by the equivalent low-pass filter and the equivalent band-pass filter, and then send it to the receiving end.

Correspondingly, an embodiment of the present invention also provides a speech data transmission system, including: a receiving end, an equivalent parallel filter, and a receiving end, wherein,

The sending end is used to sequentially compress, encrypt, encode and modulate the received voice data, and input the modulated data to the equivalent parallel filter;

The equivalent parallel filter is used to filter the received voice data, and send the filtered voice data to the receiving end;

The receiving end is used for sequentially modulating, decoding, deciphering and decompressing the received data to obtain the original voice data.

Preferably, the equivalent parallel filter is the equivalent parallel filter according to claim 14 .

As can be seen from the above technical solutions, the embodiment of the present invention can not only construct the GFR-4 and GFR-2 channel models for the connection mode from the GSM end to the GSM end, but also the GFR-4 and GFR-2 channel models from the time domain and frequency domain. The channel model is identified and analyzed from two perspectives in the domain to obtain accurate amplitude-frequency response and phase-frequency response, and the low-frequency part of 0-1550Hz is equivalent to an easy-to-use low-pass filter affected by "weak noise interference", and the 1550Hz The -4000Hz high frequency part is equivalent to a difficult-to-use band-pass filter subject to "strong random interference". Therefore, the association between the cascaded vocoder and the low-pass and band-pass filters can be established, which can provide a basis for the design of the voice-band quasi-stationary modem, thereby solving the problem of encrypted voice transmission in the mobile voice service channel, thereby improving system transmission efficiency.

Description of drawings

Fig. 1 is a flow chart of a method for establishing an equivalent parallel filter provided in an embodiment of the present invention;

Fig. 2 is the structural representation of the GFR-4 black box channel model that provides in the embodiment of the present invention;

Fig. 3 is the structural representation of the GFR-2 black box channel model provided in the embodiment of the present invention;

Fig. 4 is the GFR-4 channel that provides in the embodiment of the present invention to the output time-domain wave form diagram of different frequency single-frequency sinusoidal signals;

Fig. 5 is the GFR-2 channel that provides in the embodiment of the present invention to the output time-domain wave form diagram of different frequency single-frequency sinusoidal signals;

Fig. 6 is the time-domain analysis result figure that adopts Euclidean distance measurement and cosine angle measurement in the embodiment of the present invention;

7 is a full-phase logarithmic magnitude spectrum and a phase spectrum diagram of the 600Hz single-frequency cosine input and output signals provided in the embodiment of the present invention;

Fig. 8 is the output phase shift and normalized amplitude of the GFR-4 and GFR-2 channels in the 0-4000Hz frequency range provided in the embodiment of the present invention;

Fig. 9 is the structural representation of the equivalent parallel filter model of the GFR-4 and GFR-2 black box system provided in the embodiment of the present invention;

FIG. 10 is a graph of the bit error rate of each data subcarrier provided in the embodiment of the present invention;

FIG. 11 is a flowchart of a voice data transmission method provided in an embodiment of the present invention;

FIG. 12 is a schematic structural diagram of a device for establishing an equivalent parallel filter provided in an embodiment of the present invention;

FIG. 13 is a schematic structural diagram of an equivalent parallel filter provided by an embodiment of the present invention;

FIG. 14 is a schematic structural diagram of a speech data transmission system provided by an embodiment of the present invention.

Detailed ways

Embodiments of the present invention provide a method for establishing an equivalent parallel filter based on cascaded vocoders in a Global System for Mobile communications (GSM, Global System for Mobile communications) voice traffic channel, a voice data transmission method, an equivalent parallel filter, and The transmission system of voice data, this scheme is aimed at the connection mode of GSM terminal to GSM terminal, for example, can construct the GFR-4 black box channel model (hereinafter referred to as GFR-4 channel) that is formed by cascading of 4 full-rate vocoders, Aiming at the connection mode from the GSM terminal to the Public Switched Telephone Network (PSTN, Public Switched Telephone Network) terminal, a GFR-2 black-box channel model (hereinafter referred to as the GFR-2 channel) formed by cascading two full-rate vocoders can be constructed , and then, identify and analyze the channel model from two perspectives of time domain and frequency domain to obtain accurate amplitude-frequency response and phase-frequency response, and the 0-1550Hz low-frequency part is equivalent to an easy-to-use channel affected by "weak noise interference". The low-pass filter of 1550-4000Hz is equivalent to a band-pass filter that is difficult to use due to "strong random interference". Therefore, the association between the cascaded vocoder and the low-pass and band-pass filters can provide a basis for the design of the quasi-stationary modem in the voice band, thereby solving the problem of encrypted voice transmission in the mobile voice traffic channel, and improving the system reliability and effectiveness.

In this embodiment, for ease of understanding and description, the GFR-4 black-box channel model or the GFR-2 black-box channel model is called an equivalent parallel filter.

In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, and to make the above-mentioned purposes, features and advantages of the embodiments of the present invention more obvious and understandable, the following describes the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings For further detailed explanation.

Referring to Fig. 1, it is a flow chart of a method for establishing an equivalent parallel filter provided in an embodiment of the present invention, which is applicable to the transmission of GSM voice services, and the method includes:

Step 101: Establishing a channel model of cascaded vocoders;

In this embodiment, for the connection mode from the GSM end to the GSM end, the equipment and the link between the voice signal input and the output are equivalent to a black box system with unknown transfer function, and a GFR-4 black box channel model is established; The black box system includes: the mobile terminal at the transmitting end and the base station's vocoder pair, the receiving end's base station and the mobile terminal's vocoder pair, and all links between the transmitting end and the receiving end. The structure diagram of its GFR-4 black box channel model is shown in Figure 2.

It is also possible to establish a GFR-2 black-box channel model for the connection mode from the GSM terminal to the PSTN terminal, and the device and link between the voice signal input and output are equivalent to another black-box system with unknown transfer function. The black box system includes: a mobile terminal and a vocoder pair of a base station, a wireless link, a core network transmission link, and all links of a PSTN network transmission link. The structure diagram of its GFR-2 black box channel model is shown in Figure 3.

Step 102: Perform time-domain analysis and frequency-domain analysis on the channel of the cascaded vocoder to obtain accurate amplitude-frequency response and phase-frequency response; in this step, it can be for the GFR-4 black box channel model, or for GFR-2 black box channel model, including:

1) Time-domain analysis is performed on the channel of the cascaded vocoder to obtain a description of waveform distortion;

Wherein, the description method of the waveform distortion is: description of non-similar measure and description of similar measure, specifically including: Euclidean distance, inner product, cosine angle, Tanimoto measure, and other available non-similar and similar measure descriptions The method is not limited in this embodiment.

2) According to the description of the waveform distortion, the processing error of the cascaded vocoder channel is equivalent to a parallel noise source of weak noise interference in the low frequency band and strong interference in the high frequency band; A correlation is established between the introduced error and the strength of the noise disturbance;

3) Perform frequency-domain analysis on the channels of the cascaded vocoders to obtain amplitude-frequency response characteristics and phase-frequency response characteristics of the cascaded vocoder channels.

Step 103: According to the amplitude-frequency response characteristics and phase-frequency response characteristics, the cascaded vocoder channel is equivalent to a parallel connection of a weak noise interference low-pass filter and a strong interference band-pass filter;

In this step, for the GFR-4 black-box channel model, or for the GFR-2 black-box channel model,

Step 104: The speech encoding and decoding process of the cascaded vocoder channel is equivalent to a parallel model of performing low-pass filtering and band-pass filtering on the signal.

In this step, either the GFR-4 black-box channel model or the GFR-2 black-box channel model can be used.

Preferably, the method may further include: constructing an input signal that can be recognized by the channel model of the cascaded vocoder, using the input signal to excite the channel model of the cascaded vocoder at the sending end, and recording the received output signal at the terminal. For example, the input signal can be used to stimulate the GFR-4 black-box channel model or the GFR-2 black-box channel model at the sending end, and record the output signals at the receiving end respectively.

Through the above analysis results of time domain and frequency domain, it can be verified that the voice channel has a consistent description. The correlation in the high frequency band is relatively large, and the correlation in the high frequency band is small. In other words, the GFR-4 black box channel or the GFR-2 black box channel has little influence on the sinusoidal signal in the low frequency band, and the signal waveform distortion is small, but has a great influence on the high frequency band signal, and the waveform distortion is large.

The implementation process of the present invention will be illustrated below through a specific identification example of the GFR-4 channel model of the GSM full-rate vocoder.

1. Construct the input signal for model identification;

Generally speaking, the voiceband frequency range is 300-3400Hz. Without loss of generality, a cosine function x(t)=A _x cos(2πft+θ ₀ ) is constructed within the frequency range of 0-4000Hz, where the signal frequency f is stepped by 50Hz and traverses the frequency range of 0-4000Hz; the initial The phase θ ₀ =300, half of the maximum value of the effective data of the vocoder is taken as the amplitude of the signal, that is, A _x =16376. Sampling and quantization are performed at a sampling frequency f _s =8000 Hz to obtain a digital signal x(n)=A _x cos(2πnf/f _s +θ ₀ ).

This step is optional.

2. Time domain transmission characteristic analysis;

Use the input signal x(n)=A _x cos(2πnf/f _s +θ ₀ ) to excite the GFR-4 channel, and record the output signal y(n) at the receiver, where the output of the GFR-4 channel to different frequency single-frequency sinusoidal signals The time-domain waveform diagram is shown in FIG. 4 . In this embodiment, frequencies of 50 Hz, 300 Hz, 600 Hz, 1000 Hz, 1200 Hz and 1600 Hz are taken as examples. Among them, the abscissa represents the sample point, and the ordinate represents the normalized amplitude. It can be seen from this figure that the response of GFR-4 channel to input signals of different frequencies is different, and the transition time of the system from transient state to steady state is also different. For low-frequency signals, there is a large distortion in the 1-2 frame signal, and the system is in a transient state, and the signal distortion is small after the 3rd frame, and the system tends to a steady state. For higher frequency signals, the system transient time is extended to 3-4 frames, and then tends to a steady state.

Similarly, FIG. 5 is an output waveform diagram of the GFR-2 channel to single-frequency sinusoidal signals of different frequencies. It can be seen from the figure that it is similar to the process of the GFR-4 channel, that is, the response of the GFR-2 channel to different frequency input signals is different, and the transition time of the system from the transient state to the steady state is also different. For low-frequency signals, there is a large distortion in the 1-2 frame signal, and the system is in a transient state, and the signal distortion is small after the 3rd frame, and the system tends to a steady state. For higher frequency signals, the system transient time is extended to 3-4 frames, and then tends to a steady state.

In order to further describe the time-domain transmission characteristics of the GFR-4 channel for cosine signals of different frequencies, this embodiment uses a similarity measure as an example to evaluate the correlation between the output signal and the input signal. This embodiment uses Euclidean distance and cosine angle as examples to illustrate:

First, let vector x=(x ₁ , x ₂ , . . . , x _l ), and vector y=(y ₁ , y ₂ , . . . , y _l ).

(1) Euclidean distance:

${\mathrm{<span\; class="notranslate">\; d</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\sqrt{{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; T</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\sqrt{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; l</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

d ₂ (x, y)≥0, and the distance between the vector and itself is equal to 0, and the correlation is the largest. As the value increases, the correlation between vectors decreases.

(2) Cosine similarity measure:

${\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; ine</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; x</span>}}^{\mathrm{<span\; class="notranslate">\; T</span>}}\mathrm{<span\; class="notranslate">\; the\; y</span>}}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; \&Center\; Dot;</span><span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>\mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

in $\left|\right|x\left|\right|=\sqrt{\underset{i=1}{\overset{\mathrm{<figure-callout\; id="2"\; label="l\; x\; i"\; filenames="H200910139601XE00031.png,H200910139601XE00032.png"\; state="\{\{state\}\}">l</figure-callout>}}{\mathrm{\&Sigma;}}}{x}_i^{}{}_2^{}},$ $\left|\right|\mathrm{the\; y}\left|\right|=\sqrt{\underset{i=1}{\overset{l}{\mathrm{\&Sigma;}}}{\mathrm{the\; <figure-callout\; id="2"\; label="y\; i"\; filenames="H200910139601XE00031.png,H200910139601XE00032.png"\; state="\{\{state\}\}">y</figure-callout>}}_i^{}}.$

S _cosine (x, y) reflects the angle between the vectors. The larger the value, the smaller the angle between the vectors. When the angle is 0, the maximum value is 1.

Second, calculate the Euclidean distance and cosine angle between the normalized steady-state output y(n) and the corresponding input sequence x(n), and the results are shown in black in Figure 6(1) and 6(2) respectively Shown is a time-domain analysis result diagram using the Euclidean distance measure and the cosine included angle measure in the embodiment of the present invention. From Figures 6(1) and 6(2), it can be seen that the correlation between the output signal in the low frequency band and the input signal is relatively large, and the correlation in the high frequency band is relatively small. In other words, the GFR-4 channel has little influence on the sinusoidal signal in the low frequency band, and the signal waveform distortion is small, but has a great influence on the high frequency band signal, and the waveform distortion is large.

Similarly, for the normalized steady-state output y(n) of the GFR-2 channel and the Euclidean distance and cosine angle of the corresponding input sequence x(n), the results are also shown in Figures 6(1) and 6( 2) is shown in light black.

Finally, take the frequency range 0-1550Hz with less waveform distortion. 0-1550Hz is defined as the low frequency band, and 1550-4000Hz is the high frequency band. Since the distortion of the waveform is caused by the GFR-4 channel transmission, and the waveform distortion in the low frequency band is small, and the waveform distortion in the high frequency band is large, therefore, the embodiment of the present invention converts the distortion introduced by the GFR-4 channel transmission into a weak noise interference in the low frequency band , Parallel noise sources with strong interference in the high frequency band. The noise source is actually the encoding and decoding error and encoding conversion error introduced by the encoding and decoding of the GSM FR cascaded vocoder, thus converting the encoding and decoding error of the cascaded vocoder into an interference noise source, and obtaining Qualitative disturbance strength description.

3. Analysis of frequency domain response characteristics;

First, analyze the frequency response of a certain frequency sine wave passing through the system, without loss of generality, select the frequency f = 600Hz. Perform spectrum analysis on the input and output signals, as shown in Figure 7, which is the full-phase logarithmic magnitude spectrum and phase spectrum of the 600Hz single-frequency cosine input and output signals. In FIG. 7 , the abscissa is the number of samples, and the ordinate is dB or degrees.

It can be seen from the amplitude spectrum diagram that there are many harmonic components with an amplitude of -50db in the amplitude spectrum of the input signal x(n), which is the error introduced by the cosine signal x(t) after sampling and quantization. The peak spectral lines of the input signal x(n) and the output signal y(n) are both at k=19, and no new frequency components are generated, indicating that the GFR-4 channel is a linear system. Among them, the signal-to-noise ratio of x(n) is -100db, and the signal-to-noise ratio of y(n) is about -40db. The system is disturbed by noise, and all harmonics of x(n) are submerged by noise.

It can be seen from the phase spectrum diagram that the phase spectrum of the signals x(t) and x(n) has a horizontal straight line characteristic, and the phase values at the position of the peak spectral line (k=19) and its surrounding spectral lines are equal to the initial phase. The phase spectrum of y(n) is disturbed by noise, and only the phase spectrum near the peak spectrum line (k=19) can be approximately considered as the initial phase of the signal.

It can be seen that the GFR-4 channel has frequency invariance to the 600Hz cosine signal, but the amplitude and phase of the signal are affected by the transmission noise. The transmission noise is actually the encoding and decoding errors and encoding conversion errors introduced by the encoding and decoding of the GSM FR vocoder.

The phase shift and normalized amplitude due to transmission noise are calculated below. Select N=512, and obtain the initial phase θ _z =30.00060, the phase offset Δθ=0.00060, and the normalized amplitude through the all-phase APFFT spectrum analysis method ${\hat{A}}_z=0.99935.$

For other frequencies f, calculate the initial phase, phase offset and normalized amplitude respectively, the results are shown in Figure 8(1) and 8(2) respectively, within the range of 0-4000Hz frequency band provided in the embodiment of the present invention Output phase shift and normalized amplitude for GFR-4 and GFR-2 channels. It can be seen from Figures 8(1) and 8(2) that the normalized amplitude change of 50-1550Hz is less than 0.1, the phase deviation of the 500-1550Hz frequency band is less than 0.5 degrees, and the phase deviation of the 50-450 frequency band is less than 2 degrees.

It can be seen that the GFR-4 channel has certain "transparent" transmission characteristics for 0-1550Hz single-frequency sinusoidal signals, which can be approximately equivalent to an easy-to-use low-pass filter interfered by weak noise; for 1550-4000Hz single-frequency A sinusoidal signal creates a large "block" that can be approximated as an unusable bandpass filter interfered with by strong noise. Thus, the process of extracting and synthesizing the speech parameters of the input signal by the vocoder is equivalent to a parallel model of performing low-pass filtering and band-pass filtering on the signal, as shown in FIG. 9 , which is the GFR- Schematic diagram of the structure of the equivalent parallel filter model of the 4 and GFR-2 black-box systems.

For the transmission of actual analog signals, the spectral characteristics are more complex than single-frequency sinusoidal signals. After encoding and decoding by the vocoder, the actual low-pass frequency range is reduced, and the transparent transmission characteristics will be reduced. It can be approximated that the noise interference Increase, as long as the demodulator at the receiving end can effectively resist the influence of these noises and reliably demodulate the received analog signal with certain distortion, then the low frequency band can still be equivalent to a low-pass filter interfered by weak noise . The "blocking" of the high-frequency part will be greater, or the noise interference will be greater, and it will be more difficult to use.

4. Simulation verification

Since OFDM sub-carrier frequencies increase at equal intervals, it is easy to evaluate the frequency response characteristics in a small frequency band in the continuous frequency band, so the present invention uses OFDM to realize the conversion of random data and analog signals and to verify the conclusion.

The bit error rate curve diagram of each data subcarrier received by the receiving end is shown in Figure 10 . From the verification results, it can be seen that the constellation diagram of data subcarriers 1-20 (frequency range 0-1050Hz) converges well, and the demodulation bit error rate is less than 0.1%, indicating that the noise interference in this frequency band is very small; as the frequency increases, the constellation The diagram gradually diverges, and the bit error rate increases. When the subcarrier is 25 (frequency 1250Hz), the constellation diagram is completely blurred and cannot be demodulated. It shows that for the actual analog signal, the range of the low-pass frequency band is narrowed, and the permeability is reduced. That is to say, the simulation results confirm the correctness of the cascaded vocoder equivalent filter model.

It can be seen that the embodiment of the present invention establishes a relationship between the cascaded vocoder and the equivalent low-pass and band-pass filters, which can provide a basis for the design of the voice-band quasi-stationary modem, thereby solving the problem of encrypted voice in mobile voice. The problem of transmission in the traffic channel improves the reliability and effectiveness of the system.

Correspondingly, the embodiment of the present invention also provides a voice data transmission method, the flow chart of which is shown in Figure 11, the method includes:

Step 201: The equivalent parallel filter receives the encrypted and modulated voice data sent by the sending end;

Step 203: After the equivalent parallel filter filters the voice data, it sends the filtered data to the receiving end, so that the receiving end can demodulate and decrypt the received data in order to obtain the original voice data.

Preferably, the method may further include: establishing an equivalent parallel filter. The implementation process of the method for establishing an equivalent parallel filter is shown in FIG. 1 and its preferred embodiment, and will not be repeated here.

Please also refer to FIG. 12 , which is a schematic structural diagram of a device for establishing an equivalent parallel filter provided in an embodiment of the present invention. An effect unit 124, wherein the establishment unit 121 is used to establish a channel model of the cascaded vocoder; the analysis unit 122 is used to perform time-domain analysis on the channel of the cascaded vocoder established by the establishment unit 121. and frequency-domain analysis to obtain accurate amplitude-frequency response and phase-frequency response; the first equivalent unit 123 is used to combine the cascaded The vocoder channel is equivalent to the parallel connection of the weak noise interference low-pass filter and the strong interference band-pass filter; the second equivalent unit 124 is used for voice encoding of the cascaded vocoder channel established by the establishment unit 121 And the decoding process is equivalent to a parallel model of low-pass filtering and band-pass filtering on the signal.

Preferably, the analysis unit may include: a first analysis unit, a third equivalent unit and a second analysis unit, wherein the first analysis unit is used for the cascade vocoder established by the establishment unit The channel is analyzed in the time domain to obtain a description of waveform distortion; the third equivalent unit is configured to equivalently convert the processing error of the cascaded vocoder channel according to the description of waveform distortion obtained by the first analysis unit to A parallel noise source of weak noise interference in the low frequency band and strong interference in the high frequency band; the second analysis unit is used to perform frequency domain analysis on the channel of the cascaded vocoder established by the establishment unit to obtain the cascaded vocoder The amplitude-frequency response characteristics and phase-frequency response characteristics of the instrument channel.

Preferably, the channel model of the cascaded vocoder established by the establishment unit may specifically include: channel models of 4 cascaded full-rate vocoders GFR-4 and 2 cascaded full-rate vocoders GFR-2, But it is not limited to this, and other models can also be adaptively established.

Preferably, the device may further include: a construction unit, configured to construct an input signal that can be recognized by the channel model of the cascaded vocoder.

For the functions and functions of each unit in the device, refer to the corresponding implementation process of the above-mentioned method for details, and details will not be repeated here.

Correspondingly, an embodiment of the present invention also provides an equivalent parallel filter, whose structural diagram is shown in Figure 13 for details, and the equivalent parallel filter may include: a receiving unit 131, an equivalent low-pass filter 132, an equivalent band Pass filter 133 and sending unit 134, where

The receiving unit 131 is used to receive the encrypted and modulated voice data sent by the sending end; the equivalent low-pass filter 132 is used to filter the received voice data and filter the filtered voice data Send to sending unit 134; Described equivalent band-pass filter 133 is used for filtering the voice data received, and sends the voice data after filtering to sending unit 134; Described sending unit 134 is used for receiving Voice data filtered by the equivalent low-pass filter 132 and the equivalent band-pass filter 133 are summed and sent to the receiving end.

For details on the functions and functions of each unit in the equivalent parallel filter, refer to the corresponding implementation process of the above method, which will not be repeated here.

Correspondingly, the embodiment of the present invention also provides a voice data transmission system, its structural schematic diagram is shown in Figure 14, the system may include: a sending end 141, an equivalent parallel filter 142 and a receiving end 143, wherein,

The sending end 141 is used to sequentially compress, encrypt, code and modulate the received voice data, and input the modulated data to the equivalent parallel filter; the equivalent parallel filter 142 is used to The received data is filtered in the low frequency band and the high frequency band respectively, and the filtered data is sent to the receiving end; the receiving end 143 is used to sequentially modulate, decode, decrypt and decompress the received data , to get the original voice data.

Wherein, the equivalent parallel filter is the equivalent parallel filter as shown in FIG. 14 .

For the functions and roles of each unit in the system, refer to the corresponding implementation process of the above-mentioned method for details, and details will not be repeated here.

As can be seen from the foregoing embodiments, the embodiment of the present invention provides a method for establishing an equivalent parallel filter, a voice data transmission method, an equivalent parallel filter, and a voice data transmission system, regardless of the connection mode from the GSM end to the GSM end , or for the connection mode from the GSM end to the PSTN end, the GFR-4 and GFR-2 channel models can be constructed, and the channel model can be identified and analyzed from the perspectives of time domain and frequency domain to obtain accurate amplitude-frequency response and phase Frequency response, the low-frequency part of 0-1550Hz is equivalent to an easy-to-use low-pass filter affected by "weak noise interference", and the high-frequency part of 1550-4000Hz is equivalent to a difficult-to-use low-pass filter affected by "strong random interference". bandpass filter. Therefore, the association between the cascaded vocoder and the low-pass and band-pass filters can be established, which can provide a basis for the design of the voice-band quasi-stationary modem, thereby solving the problem of encrypted voice transmission in the mobile voice service channel, thereby improving system transmission efficiency.

Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware, but in many cases the former is a better implementation Way. Based on this understanding, the essence of the technical solution of the present invention or the part that contributes to the prior art can be embodied in the form of software products, and the computer software products can be stored in storage media, such as ROM/RAM, disk , CD, etc., including several instructions to make a computer device (which may be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments of the present invention.

The above is only a preferred embodiment of the present invention, it should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications should also be It is regarded as the protection scope of the present invention.

Claims (13)

1. A method for setting up an equivalent parallel filter, characterized in that, comprising: Establishing a channel model of cascaded vocoders; Performing time-domain analysis and frequency-domain analysis on the channel of the cascaded vocoder to obtain an accurate amplitude-frequency response and phase-frequency response, specifically including: performing time-domain analysis on the channel of the cascaded vocoder to obtain a waveform Distortion description; according to the description of the waveform distortion, the processing error of the cascaded vocoder channel is equivalent to a parallel noise source of weak noise interference in the low frequency band and strong interference in the high frequency band; for the cascaded vocoder Perform frequency domain analysis on the channel of the cascaded vocoder to obtain the amplitude-frequency response characteristics and phase-frequency response characteristics of the cascaded vocoder channel; According to the amplitude-frequency response characteristic and the phase-frequency response characteristic, the channel of the cascaded vocoder is equivalent to a parallel connection of a weak noise interference low-pass filter and a strong interference band-pass filter; The speech coding and decoding process of the cascaded vocoder channels is equivalent to a parallel model of performing low-pass filtering and band-pass filtering on signals. 2. The method according to claim 1, wherein the channel model of the cascaded vocoder comprises: when Global System for Mobile Communications GSM-GSM is connected, 4 cascaded full-rate vocoders GFR-4 channels model; and the channel model of two cascaded full-rate vocoders GFR-2 when GSM-PSTN is connected. 3. The method according to claim 2, wherein the channel model of the GFR-4 comprises: a mobile terminal at the sending end and a vocoder pair at the base station, a vocoder pair at the receiving end base station and the mobile terminal, and All links between sender and receiver. 4. method according to claim 2, is characterized in that, the channel model of described GFR-2 comprises: the vocoder pair of mobile terminal and base station, wireless link, core network transmission link, PSTN transmission link All links. 5. The method according to any one of claims 1 to 4, characterized in that the waveform distortion is described in the following ways: non-similarity measure description and similarity measure description. 6. The method according to any one of claims 1 to 4, further comprising: constructing an input signal that can be recognized by the channel model of the cascaded vocoder, and using the input signal to transmit The channel model of the cascaded vocoder is excited at the end, and the output signal at the receiving end is recorded. 7. A transmission method of voice data, characterized in that, comprising: Establishing an equivalent parallel filter; the process of establishing an equivalent parallel filter includes: establishing a channel model of the cascaded vocoder; performing time-domain analysis and frequency-domain analysis on the channel of the cascaded vocoder to obtain an accurate The amplitude-frequency response and phase-frequency response specifically include: analyzing the channel of the cascade vocoder in time domain to obtain a description of waveform distortion; according to the description of the waveform distortion, the channel of the cascade vocoder is The processing error is equivalent to the parallel noise source of the weak noise interference of the low frequency band and the strong interference of the high frequency band; the channel of the cascaded vocoder is analyzed in the frequency domain to obtain the amplitude-frequency response characteristics and Phase-frequency response characteristics; according to the amplitude-frequency response characteristics and phase-frequency response characteristics, the cascaded vocoder channel is equivalent to a parallel connection of a weak noise interference low-pass filter and a strong interference band-pass filter; the described The voice coding and decoding process of the cascaded vocoder channel is equivalent to a parallel model of low-pass filtering and band-pass filtering on the signal; The equivalent parallel filter receives the encrypted and modulated voice data sent by the sending end; The equivalent parallel filter performs low-pass and band-pass filtering on the voice data respectively, and sends the filtered data to the receiving end, so that the receiving end can demodulate and decrypt the received data in order to obtain the original voice data. 8. The method according to claim 7, further comprising: An equivalent parallel filter is established, and the method for establishing an equivalent parallel filter is as described in any one of claims 2 to 7. 9. A device for establishing an equivalent parallel filter, comprising: Establishing a unit for establishing a channel model of the cascaded vocoder; The analysis unit is used to analyze the channel of the cascaded vocoder established by the establishment unit in time domain and frequency domain to obtain an accurate amplitude-frequency response and phase-frequency response, specifically comprising: a first analysis unit configured to analyze the The channel of the cascaded vocoder established by the establishment unit is analyzed in time domain to obtain a description of waveform distortion; the third equivalent unit is used to convert the cascade vocoder according to the description of waveform distortion obtained by the first analysis unit. The processing error of the coder channel is equivalent to the parallel noise source of the weak noise interference of the low frequency band and the strong interference of the high frequency band; the second analysis unit is used to perform frequency domain analysis on the channel of the cascaded vocoder established by the establishment unit Analyze and obtain the amplitude-frequency response characteristics and phase-frequency response characteristics of the cascaded vocoder channel; The first equivalent unit is used to equivalent the channel of the cascaded vocoder to a weak noise interference low-pass filter and a strong interference band-pass filter according to the amplitude-frequency response characteristics and phase-frequency response characteristics obtained by the analysis unit parallel connection; The second equivalent unit is used for the voice encoding and decoding process of the cascaded vocoder channel established by the establishment unit to be equivalent to a parallel model of performing low-pass filtering and band-pass filtering on the signal. 10. The device according to claim 9, wherein the channel model of the cascaded vocoder established by the establishment unit specifically includes: 4 cascaded full-rate vocoders GFR-4 and 2 cascaded full-rate vocoders Channel model for rate vocoder GFR-2. 11. The device of claim 9, further comprising: A construction unit, configured to construct an input signal that can be recognized by the channel model of the cascaded vocoder. 12. An equivalent parallel filter is characterized in that it comprises: a receiving unit, an equivalent low-pass filter, an equivalent band-pass filter and a sending unit, wherein The receiving unit is used to receive the encrypted and modulated voice data sent by the sending end; an equivalent low-pass filter, configured to filter the received voice data, and send the filtered voice data to the sending unit; an equivalent bandpass filter, used to filter the received voice data, and send the filtered voice data to the sending unit; The sending unit is used to add the received voice data filtered by the equivalent low-pass filter and the equivalent band-pass filter, and then send it to the receiving end. 13. A transmission system of voice data, characterized in that, comprising: a sending end, an equivalent parallel filter and a receiving end, wherein, The sending end is used to sequentially compress, encrypt, encode and modulate the received voice data, and input the modulated data to the equivalent parallel filter; The equivalent parallel filter is used to filter the received voice data, and send the filtered voice data to the receiving end; specifically includes: a receiving unit, an equivalent low-pass filter, an equivalent band-pass filter and The sending unit, wherein the receiving unit is used to receive the encrypted and modulated voice data sent by the sending end; the equivalent low-pass filter is used to filter the received voice data and filter the filtered voice data Send to the sending unit; the equivalent bandpass filter is used to filter the received voice data, and send the filtered voice data to the sending unit; the sending unit is used to receive the equivalent low-pass filter It is added to the voice data filtered by the equivalent bandpass filter, and then sent to the receiving end; The receiving end is used for sequentially modulating, decoding, deciphering and decompressing the received data to obtain the original voice data.

Images