JPS60242744A - Voice-data multiplex transmission system - Google Patents

Abstract

PURPOSE:To ensure the switching from voice transmission to data transmission by inputting only a noise in a room to a transmitter, extracting intensity information so as to form a threshold value, comparing this value with the intensity information of an output of a speech analyzer so as to use the result as silence detection information. CONSTITUTION:A switch SW is closed before the start of voice communication so as to attain reflection connection, only the noise in a room is inputted from the transmitter MIC and the intensity information being an output of a synthesizer SYN at that time is inputted to a room noise setting board SET. The intensity information being an output of the device SYN is added by some value in this case and th result is stored in an internal storage circuit. The room noise level before the start of communication is set automatically to the set board SET as mentioned above. Then the SW is opened. A serial/parallel converter CONV selects only the intensity information of the output of an analyzer AN, converts it into a parallel signal and inputs the result to a comparator circuit COMP. The comparator circuit COMP compares this input with the said noise level of the set board SET, and when the former is smaller than the latter, it is decided as a silence part and the silence detection information is transmitted to a data terminal equipment DTE.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention is a communication system that multiplexes and transmits low-speed encoded audio signals and data using a communication line in the audio band (0.3 to 3.4 KHz). The present invention relates to an improvement in a threshold setting method when inserting data during silence in a low-speed encoded audio signal.

[Conventional technology]

When an audio signal is digitized into a low-speed encoded signal and transmitted using a voice band communication line, the audio signal naturally has silent parts, so it is necessary to insert other data into these silent parts and transmit it. is being carried out. Statistical multiplexing transmission like this is an effective method for making effective use of lines.

FIG. 2 is a schematic diagram of a conventional transmitting side in the case of multiplexing transmission of a low-speed encoded voice signal and data.

In the figure, DTE is a data terminal, C0DER is a voice low-speed encoder, and MPX is a voice/data multiplexer.

The analog voice signal output from the telephone device is converted into a low-speed code by the voice low-speed encoder C0I)ER and sent to the voice/data multiplexer MPX.

The audio/data multiplexing unit MPX has a function of detecting silent parts included in audio, and when detecting a silent part, sends silence detection information to the data terminal DTE.
Upon receiving this silence detection information, it immediately sends the data to the voice band line via the voice/data controller, electrification unit MPX, and modem (not shown).

This method of transmitting data using the silent portion of the audio signal is an economical communication method, especially when the line price is high.

[Problem that the invention seeks to solve]

In the conventional communication system of this type, a function is added to the audio/data multiplexing unit MPX to detect silent parts included in audio, and when a silent part is detected, data is transmitted. When there is a sound part, the audio signal is transmitted sideways.

However, since the level of audio fluctuates from moment to moment, conventional PCs
It detects that the output of the audio low-speed encoder C0DEH, which is composed of an A-D converter used in M communication equipment, is smaller than a set value manually set before the start of a call, and detects that the output is a silent part. In order to solve the problem of conventional methods in which silence cannot be detected no matter how long you wait if the noise level in the room on the transmitting side is high, the method is to transmit the voice as a low-speed encoded signal to the line in the voice band. However, in a communication method that inserts and transmits other data during silent periods of a call, a voice analysis device using a linear predictive coding method is used in the low-speed encoding section,
Before transmitting data, operate the changeover switch, input only room noise to the transmitter of your station, connect the transmission output of your station back to the reception input, and synthesize the linear predictive coding method of your station. The intensity information is extracted from the output of the device, a predetermined margin is added, and it is automatically stored as a threshold value in the noise setting panel.Then, the changeover switch is returned to the normal communication state, the call is started, and the call is automatically started. The intensity information of the output of the voice analyzer of the station is extracted and constantly compared with the threshold value of the noise setting board, and there is no sound when the intensity information of the output of the voice analyzer of the own station is smaller than the threshold value of the noise setting board. This is achieved by a voice/data multiplex transmission method that transmits detection information and sends data.

[Effect]

According to the present invention, before starting data transmission, with only room noise input to the transmitter, the changeover switch is operated to loop back and connect the transmission output of the own station to the reception input. The intensity information is extracted from the output of the synthesizer of the linear predictive coding method, a predetermined margin is added, it is automatically set as a threshold value on the noise setting board, and this value is combined with the output of the voice analysis device of the local station. Since the system transmits silence detection information and starts transmitting data by comparing the strength information of This effect occurs.

〔Example〕

In the present invention, the linear predictive coding method (LPC method), which is the most efficient in terms of ensuring sound quality, is used as the audio low-speed encoder.

This LPC method is a method recently developed in the field of speech recognition, which converts analog audio into A-4D, extracts spectral envelope information and sound source information, and then extracts voiced/unvoiced information and pitch from the sound source information. This is one of the voice analysis methods that extracts period information and intensity information.

As is well known, the frame format of the L:PC system is spectral envelope information (parameter); 34 to 35 bits.
Voiced/unvoiced switching, pitch period information; 6-bit intensity information; 6-bit synchronization signal Hlbit. The voiced/unvoiced switching information is composed of 1 bit, and the P/7 cycle information is composed of 5 bits.

Also, the frame period is 20 m (50 Hz), so the transmission speed of the low-speed code is 2400 bps (48 bits).
x50Hz).

In the present invention, as described above, an LPC type voice analysis device is used as a voice low-speed encoder, and the above-mentioned LP in serial format is
A frame format signal of the C format is input to the audio/data multiplexing section.

FIG. 1 is a schematic diagram showing an embodiment of a data insertion threshold setting method in audio data multiplex transmission according to the present invention.

In the figure, MIC is the transmitter, SP is the receiver, and A/D is A-D.
Converter, D/A is D-A converter, AN is analyzer, SY
N is a synthesizer, 5END is a transmitter, REC is a receiver, M
PXs is a voice/data multiplex transmitter, and MPXr is a voice/data multiplex transmitter.
In the data multiplex receiving section, SW is a changeover switch.

In this explanation, the A-D converter A/D and the analyzer AN
Collectively referred to as CO'DER, D-4A converter D/A
and the synthesizer SYN are collectively called the decoder DEC.

The A-7-D converter A/D converts analog audio signals into ordinary 2
Using a converter for converting into a binary code, the analyzer AN converts this binary code into an LPC code.

Similarly, the synthesizer SYN converts the LPC code into a binary code, and the DA converter D/A converts the binary code into analog audio.

The transmitter 5END also has a function of arranging information such as spectrum envelope information, voiced/unvoiced switching, pitch period information, intensity information, synchronization, etc. in a predetermined frame format and transmitting it at a predetermined transmission rate. has a function of receiving the above frame format, decomposing it into each piece of information, and outputting it to the synthesizer SYN.

In the present invention, when starting voice communication, the data transmitting side manually switches the changeover switch S as shown in FIG.
Operate W to connect the output of the voice/data multiplex transmitter MPXs to the input of the local voice/data multiplex receiver MPXr via the contact of the changeover switch SW. That is, a loopback connection is performed.

In this state, no sound is input from the transmitter MIG. Therefore, the noise in the room where the transmitter MIC is placed enters from the transmitter MIC, is reflected back, and comes out at the output of the synthesizer SYN.

The intensity information (6 bits) output from the synthesizer SYN is input to the room noise setting board SET and stored in the internal memory circuit. In this case, is there a synthesis device on the room noise setting board SET? &S
The intensity information of the YN output is incremented by +α and stored in the internal storage circuit. This value α is set at the front.

By the above operation, the room noise level (to be precise, +α) before the start of the call is automatically set on the room noise setting panel SET.

Next, the changeover switch SW is returned to its original state, so that the output of the audio/data multiplex transmitter MPXs is connected to the transmission line, and the input of the audio/data multiplex receiver MPXr is connected to the receive line, respectively.

Also, a serial/parallel converter C0 connected to the analyzer AN
NV selects only the intensity information (6 bits) output from the analyzer AN, converts it into a parallel signal, and inputs it to one input terminal of the comparison circuit COMP. The other input terminal of the comparison circuit COMP is connected to the output of the internal storage circuit of the indoor noise setting panel SET, and the magnitude of the re-input is constantly compared.

If the intensity information of the analyzer AN output is smaller than the set value, it is determined that it is a silent section and silence detection information is transmitted.

Therefore, when voice communication is started and a silent portion occurs, the comparator circuit COMP transmits silence detection information.

This silence detection information is sent to the data terminal DTE, and the data terminal DTE starts transmitting data, which is sent out to the transmission line via the audio/data multiplex transmitter MPXs.

In this way, data can be transmitted using the silent portion of the audio signal.

〔Effect of the invention〕

As explained in detail above, according to the present invention, switching from voice transmission to data transmission is ensured, and the probability of switching back immediately after switching to data transmission as in the conventional method is greatly reduced. This has the effect of increasing stability.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram showing an embodiment of a data insertion threshold setting method in audio data multiplex transmission according to the present invention. FIG. 2 is a schematic diagram of a conventional transmitting side in the case of multiplexing transmission of a low-speed encoded voice signal and data. In the figure, DTE is a data terminal, C0DER is a voice low-speed encoder, MPX is a voice/data multiplexer, and M
IC is a transmitter, sp is a receiver, A/D is an A-4D converter, D/A is a D→A converter, AN is an analyzer, SYN is a synthesizer, 5END is a transmitter, REC is a receiver, MPXs
is an audio/data multiplex transmitter, MPXr is an audio/data multiplex receiver, and SW is a changeover switch. $1 country

Claims (1)

[Claims] Transmits voice as a low-speed encoded signal to a voice band line,
In a communication system that inserts and transmits other data during silent periods, a voice analysis device using a linear predictive coding method is used in the low-speed encoding section, and a changeover switch is operated before transmitting data to determine whether the local station Input only room noise into the transmitter, loop back and connect the transmission output of the own station to the reception input, extract intensity information from the output of the synthesizer of the linear predictive coding method of the own station, and add a predetermined margin. The threshold value is automatically stored in the noise setting board, and then the changeover switch is returned to the normal communication state to start a call, the intensity information of the voice analyzer output of the local station is extracted, and the noise is automatically stored as a threshold value. It is characterized in that it constantly compares with a threshold value of the setting board, and when the intensity information of the output of the voice analysis device of the own station is smaller than the threshold value of the noise setting board, it transmits silence detection information and sends data. Audio/data multiplex transmission method.