JPS60134658A - Voice and data multiplex transmission system - Google Patents

Abstract

PURPOSE:To prevent deterioration of voice and improve articulation (understandability) in a system that multiplex transmits voice and data frequency dividing in a voice zone by superposing data as high speed as possible. CONSTITUTION:An aural (speaking voice) signal of 0.3kHz-1.8kHz is taken out by passing information of a voice zone 23 through a band pass filter 7. The aural signal is inputted to an analog adder (ALADD) 10 and inputted also to a basic wave extracting circuit (BWS) 8. Basic wave that constitutes voice is extracted from component of the first formant area which is main component of the aural signal. The basic wave is multiplied to integer times of the first formant frequency, and the second formant component and third formant component are formed. Thus, control is made to generate voice having the first - third formant components and consequently, aural signals of high articulation small in deterioration can be obtained for received voice though multiplex transmission is made expanding the data transmission area of a voice zone to a limit.

Description

Detailed Description of the Invention (al) Technical Field of the Invention The present invention relates to a multiplexing system for frequency-dividing audio and data in one audio band and multiplex transmission, and a method for reproducing the multiplexed audio. Regarding.

佽）Technical Background One of the technologies related to the present invention is the technical standards regarding specific communication lines and leased lines established by Nippon Telegraph and Telephone Public Corporation.

To summarize the contents related to the present invention among the technical standards,
It will be as follows. In other words: Specified communication lines that can be used by companies and dedicated lines of D-1 standard, ■-1 standard, and J-1 standard can be used by dividing the frequency within the range specified by the above technical standards. .

The principle of this level regulation for divided usage is based on ``the total value of each divided level must be less than or equal to the same value as the level limit before division.''

However, when using a D-1 standard line that is divided into phone calls and other communication methods (e.g. data communication),
The frequency band for calls is at least 0.3KHz ~ 1,8
It stipulates that the KHz range must be secured.

This means that if the frequency band is cut down any further for other systems, the quality of the call will be reduced to 10d in terms of transmission loss.
It is known that the condition will be worse than B, and this is the limit.

Therefore, if other systems are to be simultaneously transmitted to an area that has already deteriorated to this extent, it will be necessary to further reduce the transmission level since the overall level is specified, but the human voice level is based on technical standards. Since it is difficult to regulate the transmission level, the transmission level of each call voice is specified so that at least the level state before division is maintained.

In other words, when the frequency of the output from the transmitter is analyzed, it is 1.8K.
It is known that the energy up to Hz is 93.5%, and the remaining 6.5% is contained in the frequency band above that.

Using this value, the level when superimposing other systems on the upper band of one voice band line is 6.5% of the specified level before division, and the telephone output is the same as normal. 9
The technical standard value is defined as being at a level of 3.5%.

Another matter related to the present invention is the human voice generation mechanism.

Looking at the mechanism by which human speech is generated, the speech waveform can be seen as the radiated output from the vocal tract, which is excited by a sound source such as vocal cord vibration.

In other words, human vocalization is a process in which a driving sound source is changed into an acoustic signal with a unique spectral distribution by the vocal tract system (a type of resonator made up of parts of the mouth and nose).

Therefore, when we take the spectral distribution of human speech per unit time, there are several peaks that indicate the state of the vocal tract system (resonator), and these peaks are caused by the spectral energy of the harmonics of the fundamental wave. The area where it is concentrated is called the formant region.

Then, from the lowest frequency, the first formant region, the second formant region,
This is called a formant region or third formant region, and its peak energy is attenuated at approximately 16 dB/octave.

The present invention focuses on the feature that the energy of the voice shows a formant distribution and is composed of multiples of the fundamental wave, and frequency-divides the voice and data in one voice band according to the above transmission standard. It is related to the audio reproduction method when multiplexed transmission is performed.

To be more specific, the 0°3KH of the call voice
This method frequency-multiplexes 93.5% of voice energy in the 2 to 1.8 KHz band and a data modem signal, transmits it over one voice band line, and reproduces it on the receiving side.

(C) Prior Art and Problems Conventionally, systems for transmitting voice calls and data include the Graphink audio conference system and electronic conference systems, which transmit voice calls and data using specific communication lines. , which outputs audio and still images using a specific terminal device.

In this case, in order to improve the quality of voice transmission (intelligibility), it is necessary to occupy one general voice band for transmission, and it is necessary to provide a side line for data.

When transmitting voice and data by frequency multiplexing in one voice band, the frequency range from 0°3KHz to 1KHz specified by the above-mentioned law is required.
．． Outside the 1KHz band, for example 3.0KHz to 3.4KHz
The KHz band is used to transmit voice energy in the third formant region and transmit data between 2.0 KHz and 2.6 Kt.
There is a method of transmitting using lz.

In this case, although the intelligibility of the voice is improved, the data transmission rate is limited to, for example, 1200 baud, and there is a problem that the transmission rate of the data terminal cannot be increased.

1dl Purpose of the Invention In view of the above-mentioned drawbacks of the conventional technology, the present invention provides a system in which voice and data are frequency-divided in one voice band and multiplexed for transmission. The purpose of this is to provide an audio/data multiplexing method that improves communication comprehension (intelligibility).

(e) Structure and object of the invention According to the present invention, in a system in which audio and data are frequency-divided and multiplexed in one audio band, the main components of audio are transmitted between 0.3 KHz and 1.8 KH2. A first method in which the signal is transmitted only in the band, and on the receiving side, only the main components of the voice are extracted from the received signal of one voice band using a bandpass filter, and the voice fundamental wave is extracted by the first formant of the voice signal. means, a second means for generating a second formant and a third formant from the fundamental wave, and a third means for adding an analog signal; The second formant and the third formant are generated by the second means, and the third formant is analog-added by the main component of the voice extracted by the bandpass filter and the second formant and the third formant. This is achieved by providing a method for generating speech having one formant, second formant, and third formant components, and allows high-speed data to be multiplexed by frequency division and multiplexed in one audio band while preventing audio deterioration. There are advantages.

(fl Embodiments of the Invention To summarize the gist of the present invention, the present invention focuses on the characteristic that the energy of voice shows a formant distribution and is composed of multiple waves of a fundamental wave, and furthermore,
In line with legal restrictions when frequency-dividing and multiplexing transmission in one voice band, 0.3KHz to 1.8KHz of call voice
When 93.5% of the voice energy in the KHz band and the data modem signal are frequency multiplexed, transmitted over one voice band line, and played back on the receiving side, the band is divided from the received one voice band signal. The filter extracts the main components of the voice, the fundamental wave of the voice is extracted from the first formant of the voice signal, the second formant and third formant components are generated from this fundamental wave, and the bandpass filter extracts the fundamental wave of the voice from the first formant of the voice signal. The extracted audio signal is subjected to analog addition to generate audio having the first to third formant components.

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a diagram explaining the principle of a multiplexing system when implementing the present invention, FIG. 2 is a diagram explaining the spectrum of human voice, and FIG. 3 is a block diagram of an embodiment of the present invention. This is a diagram shown in .

In FIG. 1, 1 indicates a fundamental wave area, 2 a voice transmission area, 3 a data transmission area, and 23 one voice band. Normally, the transmission band of a specific communication line or dedicated line is legally defined as 0.3 KHz to 3.4 KHz (this is the above-mentioned 1 audio band 23).

Therefore, the fundamental wave area 1 containing the fundamental wave of the audio frequency will not be transmitted.

When implementing the present invention, the above-mentioned 0
．． Voice transmission is performed only in a band of 3 KHz to 1.8 KHz, and the remaining band of 1.81 [Hz to 3.4 KHz is used as a data transmission band to improve the data transmission speed. The audio energy distribution at this time is 93.5% for the audio transmission area 2 and the remaining 6.5% for the data transmission area.

By adopting the above-mentioned voice and data multiplexing method, data terminals include electronic blackboard 2 handwriting terminals with a transmission rate of 1200 baud level, telex terminals with a transmission rate of 5° baud level, remote monitoring and control equipment (
It becomes possible to connect low-speed terminals such as SV terminals).

Next, the spectral distribution of human speech will be explained with reference to FIG.

In this figure, the horizontal axis is the audio frequency (f), and the vertical axis is the spectral density.

As mentioned above, human voice is converted into an acoustic signal with a unique spectral distribution by the vocal tract system, whose driving sound source is a type of resonator, and as shown in Figure 2, the voice A spectral distribution with several peaks corresponding to the road system is shown, and when viewed limited to the one audio band, the spectrum distribution has three peaks.

Winter mountains are places where the spectral energy of the multiples of the fundamental wave is concentrated, and the formant region [spectral envelope (
formant) and the first formant region 4. Second formant region5. The third formant region 6 is called the second formant region 6 for the first formant component. The peak energy of the third formant component is attenuated at approximately 16 dB/octave.

The center frequency and spectral distribution characteristics of each formant region indicate the characteristics of each voice.

Focusing on the characteristics of the audio spectrum described above, a method for generating a highly intelligible audio signal from the audio signal that has been multiplexed and transmitted by frequency division will be explained using the embodiment shown in FIG.

In FIG. 3, 7 is a bandpass filter, 8 is a fundamental wave extraction circuit (BWS), and 9 is a multiplication wave generation circuit (HWG).

10 is an analog adder (ALADD), and 11 is a specific communication line.

As mentioned above, the information within one voice band 23 transmitted through the specific communication line 11 includes call voice having 93.5% voice energy in band 2 of 0.3 KHz to 1.8 KHz, and
Consists of data with 6.5% audio energy in z-band 3 from 1.8 KHz to 3.4 Kl.

By passing this 1 audio band information through a bandpass filter 7, an audio (call audio) signal in the 0.3 KHz to 1.8 kHz band 2 is extracted, and the audio signal is input to an analog adder (ALADD) 10. At the same time, it is also input to a fundamental wave extraction circuit (BWS) 8, which extracts the fundamental wave constituting the voice from the components in the first formant region, which are the main components of the voice signal, and converts it into a multiplied wave. A generation circuit (HWG) 9 multiplies the fundamental wave by an integral multiple of the first formant frequency, and generates second and third formant components whose audio energy density is attenuated by the above 16 dB/octave. .

An analog adder (ALADD') 10 performs analog addition of the generated audio signal of the 2.3rd formant component with the audio signal of the 0.3 KHz to 1.8 KH2 band 2 that has been input in advance. According to the first
From the formant components, the second. It is possible to generate an audio signal including up to the third formant component, and the second formant component, which was lost during multiplexing, can be generated. Since the third formant component is generated by multiplying the fundamental wave of the voice, it is possible to obtain a voice signal 2 that is extremely close to the voice of the call before multiplexing.

(Effects of the Invention As explained in detail above, the voice/data multiplex transmission system of the present invention focuses on the feature that voice energy exhibits a formant distribution and is composed of multiples of the fundamental wave. , and in accordance with legal restrictions when transmitting voice and data by frequency division and multiplexing in the voice band, 93 in the band of 0°3KHz to 1.8KHz of call voice.
．． When 5% voice energy and a data modem signal are frequency multiplexed, transmitted over one voice band line, and played back on the receiving side, the received one voice band signal is filtered through a bandpass filter to extract the above voice. Extract the main component, extract the fundamental wave of the voice from the first formant of the voice signal, generate the second formant and third formant components from this fundamental wave, and add analog to the voice signal extracted by the bandpass filter. As a result, the received audio is controlled to generate audio having the first to third formant components, so even though the data transmission area in the audio band is expanded to the limit and multiplex transmission is performed, the received audio is This has the effect of making it possible to obtain an audio signal with high clarity.

[Brief explanation of drawings]

Fig. 1 is a diagram explaining the principle of the multiplexing method when the present invention is implemented, Fig. 2 is a diagram explaining the spectrum of human voice, and Fig. 3 is a block diagram showing an embodiment of the present invention. This is a diagram. In the drawing, 1 is the fundamental wave area, 2 is the audio transmission area, 3 is the data transmission area, 23 is the 1 audio band, and 4 is the first
formant region, 5 is a second formant region, 6 is a third formant region, 7 is a bandpass filter, 8 is a fundamental wave extraction circuit (
BWS), 9 is a multiplication wave generation circuit (H),
10 is an analog adder (ALADD). 11 indicates a specific communication line. 4 3rd] 2 -351-

Claims (1)

[Claims] In a method that frequency-divides and multiplexes voice and data in one voice band, the main components of voice are transmitted at 0.3KHz.
~1, transmits only in the 8KHz band, and the receiving side -
a first means for extracting only the main components of the voice from the received signal of one voice band using a bandpass filter, and extracting a voice fundamental wave from the first formant of the voice signal; a second means for generating two formants and a third formant; and a third means for adding the analog signals;
A second means generates a second formant and a third formant from the fundamental wave of the voice extracted by the first means, and a third formant generates the main components of the voice extracted by the bandpass filter and the above. A voice/data multiplex transmission system characterized in that a second formant and a third formant are added in analog form to generate voice having first formant, second formant, and third formant components.