JPH042014B2 - - Google Patents

Description

Detailed Description of the Invention [Summary] The present invention provides, for example, when a voice signal and a facsimile signal are respectively predictively encoded and transmitted, the transfer function of the predictor is selected to be suitable for each signal. In the case of facsimile signals, it conforms to the CCITT recommendations, and in the case of facsimile signals, communication conforms to public sector networks.

[Industrial application field]

The present invention allows analog signals such as voice signals and facsimile signals that pass through a modem to be
This paper relates to improvements in the encoding method used in the encoding section of the 32Kbps ADPCM system.

A system that encodes analog signals that pass through a modem, such as voice signals and facsimile signals.
An example is shown in Figure 2, where telephones 1 and G
A standard facsimile machine 2 (with a built-in modem and the output is an analog signal) is housed in the exchange 3, and these are connected to the output carrier line according to the request, and the analog signal to be sent is is encoded by the 32 Kbps ADPCM method in the encoding unit 4 and output.

The encoding unit 4 of this 32KbPs ADPCM system is based on the CCITT recommendations, and when considering connection with the Telegraph and Telephone Public Corporation network or other companies' networks, it is natural that it is better to use one that complies with the CCITT recommendations.

However, the predictor of the encoder of the 32Kbps ADPCM system according to the CCITT recommendation has a transfer function of voice formant characteristics from the perspective of voice quality.
Regarding modem signal communication, only 4800bps is guaranteed.

On the other hand, most recent facsimiles are of the G standard, and this G standard facsimile has a speed of 9600bps.
Therefore, the coding part of the CCITT recommendation cannot be used as is for facsimile.

Therefore, in the case of an audio signal, the encoder
What is desired is something that complies with the CCITT recommendations and can communicate at 9600 bps for signals from a modem, such as facsimile signals.

[Conventional technology]

FIG. 3 is a block diagram showing the configuration of a conventional 32 Kbps ADPCM encoding section.

In the figure, 5 is an analog-to-digital converter, 6 is a subtracter, 7 is a quantizer, 8 is an inverse quantizer, and 9 is a predictor.

In FIG. 3, the input audio signal or facsimile signal is converted into a digital signal at a rate of 64 Kbps by an analog-to-digital converter 5, and the difference between it and the output of the predictor 9 is taken by a subtracter 6. , is quantized by a quantizer 7 and output at a rate of 32 Kbps.

On the other hand, the output of the quantizer 7 is also input to the inverse quantizer 8, where it is inversely quantized, multiplied by a prediction coefficient at the predictor 9, and input to the subtracter 6.

In this way, the 64 Kbps code sampled by the analog-to-digital converter 5 is encoded using the ADPCM method to become 32 Kbps on the output side.

In this case, the predictor 9 uses the audio signal as well as the audio signal.
In order to be able to communicate 9600bps signals from facsimile, we use one that has an uncharacterized transfer function over the entire frequency band, rather than the one recommended by CCITT.

[Problem that the invention seeks to solve]

However, this degrades the audio quality and
Also, since it is not based on the CCITT recommendations, there is a problem that it cannot be connected to the Telegraph and Telephone Public Corporation network or other companies' networks.

[Means for solving problems]

The problem mentioned above is that in the ADPCM encoding section that encodes the audio signal and the analog signal sent via the modem at 9600 bps, the input section is equipped with a detector for training pulses sent from the modem and a modem. During use, a carrier monitoring unit is provided to monitor the transmitted carrier, and as a predictor, a first prediction unit (compliant with CCITT, G721 recommendation) of a transfer function of audio formant characteristics is provided.
A second prediction part of the transfer function that is not characterized over the entire frequency band (taking into account the variance of the amplitude distribution of the prediction residual signal, the prediction order is increased by applying CCITT and G721). , the first prediction unit that encodes the audio signal is always used, and when the training pulse is detected by the detector, the second prediction unit is used and the signal is sent. This problem is solved by the encoding method of the present invention in which the incoming carrier is monitored by the carrier monitoring section, and when the carrier monitoring section no longer receives the carrier, the first prediction section is used.

[Effect]

According to the present invention, the predictor includes a first predictor (CCITT,
G721 recommendation), and a second prediction part of the transfer function that is not characterized over the entire frequency range (taking into account the variance of the amplitude distribution of the prediction residual signal, CCITT,
G721 with an increased prediction order), and is always used to encode the audio signal.
If the training pulse is detected by the detector and the signal is from a modem such as a 9600 bps signal such as a facsimile, the second predictor is used. It does not degrade the signal, meets CCITT recommendations for voice signals, and allows the use of signals from modems of 9600 bps.

〔Example〕

FIG. 1 is a block diagram showing the configuration of a 32 Kbps ADPCM encoding unit according to an embodiment of the present invention.

In the figure, 10 is a training pulse detector, and 11 is a training pulse detector.
12 is a predictor with a transfer function that takes into account voice formant characteristics that meet the CCITT recommendations, 13 is a carrier monitoring unit, and SW1 and SW2 are switches. .

The difference between Fig. 1 and Fig. 3 is that when the facsimile sends a transmission request to the modem, the input section detects the training pulse that the modem on the other end sends to synchronize. A carrier monitoring section 13 is provided to monitor the carrier whose frequency is 1700 Hz while the modem is in use.
A predictor 11 that conforms to the CCITT recommendations and a predictor 12 that has a transfer function that is not characterized over the entire frequency band are provided, and in a normal case, a switch SW2,
Set SW1 to the solid line side to use the predictor 11, and if the training pulse is detected by the training pulse detector 10, the switch will be activated.
The predictor 12 is used with SW2 and SW1 on the dotted line side, and at this time, the carrier output from the modem is monitored by the carrier monitoring unit 13. When the modem is stopped and the carrier runs out, the carrier monitoring unit 13 is the point where the switches SW1 and SW2 are placed on the solid line side.

In this way, the encoder complies with the CCITT recommendations in the case of audio signals, can be connected to the Telephone and Telecommunications Public Corporation network or other companies' networks, and can also be used for G standard facsimile.

Note that there is only one predictor, and in the case of audio signals, the parameters of the transfer function are taken into consideration with the audio formant characteristics, and in the case of facsimile, the parameters of the transfer function are not characterized over the entire frequency band. Of course, it may also be possible to switch between them.

[Effects of the Invention] As explained in detail above, according to the present invention,
In the case of an audio signal, the encoding part of the ADPCM method is
Conforms to CCITT recommendations, can be connected to the Telegraph and Telephone Public Corporation network or other companies' networks, and in the case of a signal from a modem
It has the effect of being able to communicate signals of 9600 bps.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing the configuration of a 32 Kbps ADPCM encoding unit according to an embodiment of the present invention, and Fig. 2 shows an example of a system for encoding analog signals that pass through a modem for audio signals and facsimile signals. The block diagram, Figure 3, shows the conventional example.
FIG. 2 is a block diagram showing the configuration of a 32 Kbps ADPCM encoding unit. In the figure, 1 is a telephone, 2 is a facsimile,
3 is a switch, 4 is an encoder, 5 is an analog-to-digital converter, 6 is a subtracter, 7 is a quantizer, 8 is an inverse quantizer, 9, 11, 12 is a predictor, 10 is a training pulse detector 13 is a carrier monitoring unit, and SW1 and SW2 are switches.

Claims (1)

[Claims] 1 ADPCM that encodes the audio signal and the analog signal sent through the modem with a 9600 bps signal
In the encoding section of the system, a detector for training pulses sent from the modem and a carrier monitoring section for monitoring the carrier sent out while the modem is in use are provided at the input section, and the ADPCM predictor includes:
A first prediction unit having a transfer function of voice formant characteristics, a second prediction unit having a transfer function that does not characterize the entire frequency band, and means for switching between the first and second prediction units are provided. , when the training pulse is detected by the detector, controls the switching means to use the second prediction section, inputs the sent carrier to the carrier monitoring section, and inputs the carrier to the carrier monitoring section; An encoding system characterized in that the switching means is controlled to use the first prediction unit when a carrier is no longer received.