JPS58225751A - Sound multiplex transmission system - Google Patents

Abstract

PURPOSE:To carry out multiplexing , by performing time compression regarding at least voiced-sound blocks, and utilizing voiceless sound blocks and a free time of said time compression. CONSTITUTION:When an input sound is made into blocks in an array shown in a figure, time periods t1, t2- are utilized for the transmission of other sounds. The input sound signal is divided into a voiced sound, a voiceless sound, and a nonsound part by a detection part 7 and a control part 13 controls a switch part 8. Namely, a contact (a) is selected for the voiced sound and the output of a compression processing part 5 is sent to a multiplexing part 9. A contact (b) is selected for the nonsound part and no signal is supplied to the channel multiplxing part 9. A contact (c) is selected for the voiceless sound and a unit 6 makes the input into blocks all the time, but a signal corresponding to the nonsound part is supplied to the multiplexing part 9 during this period. The multiplexing part 9 receives similar signals from other channels to multiplex them.

Description

[Detailed Description of the Invention] fAl Technical Field of the Invention The present invention uses an audio multiplex transmission system, which divides an I\jK audio signal into cases where it is voiced and voiced, cases where it is voiced and unvoiced, and cases where it is silent. 7. Transmit the southern voice block and the unvoiced sound block in a corresponding manner at predetermined time intervals, and time-compress at least the voiced sound block. This invention relates to an audio multiplex transmission system in which analog signals are multiplexed as they are by using at least the silent sections and the free time of the time-compressed voiced sound blocks.

Background and Problems of fBl Technology A method for multiplexing voice as an analog signal through a transmission path having a telephone voice band is a combination of a bandpass filter and frequency division multiplexing (FDM). However, with this method, the audio bandwidth per channel may become too narrow. On the other hand, TAS I is known as a method that utilizes the pause period (silent period) during a call, but in conventional TASH, the analog waveform is separated and transmitted when it reaches the koto or a value close to it. Because of this, the signal duration is not constant and the control system is complicated.

For this purpose, the input audio signal is divided into blocks at predetermined time intervals, and only the blocks containing audio levels higher than a predetermined value are transmitted. However, even with this method, the line The problem is that the number of audio channels that can be accommodated is only about twice as large.

(C1 Object and Structure of the Invention The present invention aims to take the above proposal one step further, and divides each block of the input audio signal into voiced sound blocks, unvoiced sound blocks, and silent blocks, and then divides each block of the input audio signal into voiced sound blocks, unvoiced sound blocks, and silent blocks. Perform time compression on the above, and use at least the above-mentioned silent blocks and the time-compressed free time.

The purpose is to perform multiplexing.

Therefore, the audio multiplex transmission system of the present invention is as follows.

Using a transmission line with a telephone voice band, the voice signal is divided into blocks at predetermined time intervals, and then the frequency band e of each block is
In the audio multiplex transmission method, which compresses the time of blocks and transmits them in blocks, and multiplexes a number of audio channels greater than the number of actual transmission paths, an audio signal blocking processing section is provided on the transmitting side. The blocking processing unit divides the input audio signal into blocks at predetermined time intervals after passing it through a low-pass filter when the input audio signal is both voiced and voiced. The time length of the block is compressed so as to expand the width to the entire telephone voice band and sent to the transmission path, and if the input voice signal is both voiced and unvoiced, the voice signal is transmitted at predetermined intervals. Separate and create blocks.

Either send it to the transmission path as it is, or delete a certain proportion of the time portion of the block and send it to the transmission path as a block with a short time length, and furthermore, if the input audio signal is not active, the corresponding input The system is configured to pause the transmission of the audio signal, restore the time length of the block corresponding to the voiced sound to the original time length on the receiving side, and restore the block corresponding to the unvoiced sound to the original time length according to the deletion ratio. The feature is that the data is restored to a length of time. This will be explained below with reference to the drawings.

fDl Embodiment of the Invention FIG. 1 (!I) to ffl are explanatory diagrams explaining time compression for a mausoleum sound block according to the present invention, FIG. 2 is an explanatory diagram explaining a transmission mode of an embodiment according to the present invention, FIG. 3 is an explanatory diagram illustrating the transmission mode of another embodiment of the present invention, FIG. 4 shows an embodiment of the transmitting side configuration of the present invention, and FIG. 5 shows an embodiment of the receiving side configuration.

For example, an audio signal ti input by a microphone
, it has a waveform as shown in Figure 1 (fa1), and its spectrum can be considered to be as shown in Figure 1 (b1). When extracted by a low-pass filter, the signal shown in Fig. 1 (C1) is obtained, and the signal shown in Fig. 1 (d+
It has a spectrum as shown in the figure.

Telephone voice band (e.g. several hundred Hz to several hundred Hz)
), when transmitting a voiced sound part as shown in Fig. 1 f01, 2
For example, compressing the time axis of a block with time T to about 1%
Instead, it is possible to transmit the highest frequency component at a higher frequency of 1700×2 (Hz). By doing it this way.

As shown in FIG. 1 (el), about half of the time in the voiced sound block is empty, and this empty section can be used for transmitting other voices.

In one embodiment of the present invention, as shown in FIG. 2, since unvoiced sounds have components of, for example, 1700 Hz or higher, unvoiced sound blocks are transmitted without being subjected to time compression processing. Ru.

That is, when the input speech is divided into predetermined time intervals according to time t and becomes voiced, unvoiced, silent, unvoiced, voiced, etc. as shown in Figure 2 (at1), voiced As explained with reference to Fig. 1, the signal is compressed into a time of about 1 and transmitted, unvoiced sounds are transmitted as they are, and silences are not transmitted.This state is shown in Fig. 2 fblK. Cage 2 indicated time t□, t2
. . . becomes available for transmitting other audio.

FIG. 3 shows an explanatory diagram illustrating a transmission mode of another embodiment according to the present invention. In the case of the illustrated embodiment, the idea is that unvoiced sounds with high frequency components are, for example, parts where the characteristics of the speaker are not matched to that extent, and that it is sufficient to transmit what kind of unvoiced sounds they are.

FIG. 3 fal corresponds to the case shown in FIG. 2.

In the embodiment shown in FIG. 3, during transmission, the period of unvoiced sound is cut down to, for example, % as shown in FIG. The free time obtained by compressing the voice period, (along with the free time corresponding to the I+1 silent period, the above flli+
To use the free time obtained by compressing the period of unvoiced sound for other voice transmission.

In Fig. 2, it is not specified what to do on the receiving side based on the transmitted result as shown in Fig. 2 (b11). However, for the section of voiced sound shown in Fig. 2 (b11), In the case shown in Fig. 3 (as shown in Fig. 3 (c11), the previously compressed time axis is similarly restored! Also, for unvoiced sections.

In the case shown in Figure 2 (+), the received unvoiced sound is used to reproduce the unvoiced sound, and in the case shown in Figure 3 (b11), the section of the received unvoiced sound is repeated twice and output, and the unvoiced sound is reproduced. Needless to say, a silent period is taken as shown in FIG. 3 (C1 (also K in FIG. 2)) corresponding to the silent section.

FIG. 4 shows an embodiment of the transmitting side configuration of the present invention.

In the figure, 1 is an audio signal source, 2 is an analog-to-digital converter, 5 is a low-pass filter, 4 is a voiced sound block processing unit, 5 is a time compression processing unit, 6 is an unvoiced sound block processing unit, and 7 is an unvoiced sound block processing unit. 8 is a changeover switch section, 9 is a channel multiplexing section, 10 is a digital-to-analog conversion section, 11 is a transmission line group, and 12 is a voice characteristic detection section that discriminates whether input speech is voiced, unvoiced, or silent. A common control i-line 15 represents a control section.

The input audio signal is discriminated between voiced, unvoiced, and silent by the audio characteristic detection section 7, and is supplied to the unit 4 via the filter 3 and also to the unit 6. The output of the unit 4 is time-compressed by a time compression processing section 5 as shown in FIG.

When the voice characteristic detecting section 7 determines whether the sound is voiced, unvoiced, or silent, the control section 15 controls the changeover switch section 8 as follows. That is.

(1) Select contact point a corresponding to the voiced sound. The time compression processing unit 5 is supplied with components of the input audio at frequencies below, for example, 1700 Hz, which correspond to voiced sounds according to the present invention as shown in FIG. 1 (c11).

Therefore, the channel multiplexer 9 is supplied with a signal as shown in FIG. 1 (C1).

(11) Select a contact point in response to silence. Therefore, no signal is supplied to the channel multiplexer 9.

fill) Select contact C in response to unvoiced sound.

Although the unit 6 blocks the input audio signal over all periods, the period during which the contact point C is selected corresponds to the unvoiced period of the input audio.

Therefore, the channel multiplexer 9 is supplied with a signal corresponding to unvoiced sound via the switch 8.

In the case of the embodiment shown in FIG. 5, a two-hour cutting section 14 is interposed as shown by the dotted line in FIG.

It may be considered that the period of unvoiced sound is cut into %, for example, and then supplied to the channel multiplexing section 9.

Channel multiplexing section 9 receives similar signals from other channels, multiplexes them, and outputs them onto transmission line group 11. On the other hand, the decoration unit 13 transmits information on how the changeover switch 8 is selected for each channel via the common control line 12.

FIG. 5 shows an embodiment of the receiving side configuration of the present invention.

In the figure, reference numeral 11 is a transmission line group, 12 is a common control line, 15 is an analog-to-digital converter, 16 is a channel separation unit, 1
7 is a time expansion section, 18 is a changeover switch section, and 19 is a digital/analog conversion section.

20 represents a receiver, and 21 represents a common control unit.

1! sent via the transmission line group 11! ! The signals are led to the channel separation section 16 and separated into each channel. The separated signals for each channel can be considered to be as shown in FIG.
The time axis is expanded by about twice as much, and the signal is guided to the contact point a, and is also guided to the one-plane contact point C. The common control unit 21 receives the signal from the common control line 12 and controls the changeover switch 18 so as to correspond to the operation of the changeover switch 8 on the transmission side corresponding to the channel. That is, the changeover switch 18 is controlled to select contact point a in response to voiced sound, select contact point A in response to silent sound, and select contact point C in response to unvoiced sound. Therefore, for voiced sounds, the time axis is extended and returned to the original time axis as clearly shown in Figure 3 (cl).For silent sounds, no signal is supplied to the digital to analog converter. The output from the channel separator 16 is directly led to the digital/analog converter 19.The digital/analog converter 19 converts the input signal into an analog signal and supplies it to the receiver 20V.

In the case of the embodiment shown in FIG. 3, an iterative processing section 22 is interposed as shown by the dotted line in FIG.

As a result, for unvoiced sounds, the received unvoiced sounds are repeated, for example, twice, and then the digital-to-analog converter 19
It will be supplied to

fil As described in detail, according to the present invention, at least the silent period and the free time obtained by compressing voiced sounds can be used for transmitting other audio signals. .

[Brief explanation of the drawing]

1 +8+ to ffl are explanatory diagrams for explaining time compression for voiced sound blocks according to the present invention, FIG. 42 is an explanatory diagram for explaining the transmission aspect of one embodiment according to the present invention, and FIG. An explanatory diagram illustrating the transmission aspect of the embodiment, FIG. 4 shows an embodiment of the configuration on the transmitting side of the present invention, and FIG. 5 shows an example of the configuration on the receiving side. In the figure, 1 is an audio signal source, and 5 is a low-pass filter. 4 is a voiced sound block processing unit, 5 is a time compression processing unit, 6 is an unvoiced sound block processing unit, 7 is a voice characteristic detection unit, 8 is a changeover switch, 16 is a channel separation unit, 17 is a time expansion unit, 18 is a changeover switch represents. Patent applicant Fujitsu Limited

Claims (1)

[Claims] (1) Using a transmission line with a telephone voice band, the voice signal is divided into blocks at predetermined time intervals, and the blocks are time-compressed according to the frequency band each block has, in units of blocks. In an audio multiplex transmission system that transmits and multiplexes a number of audio channels greater than the number of actual transmission paths, -C1 includes an audio signal blocking processing section on the transmitting side. If the input audio signal is a voiced and voiced sound, e.After passing through a low-pass filter, the input audio signal is divided into blocks at predetermined time intervals, and the bandwidth of the block is expanded to cover the entire telephone audio band. The time length of the block is compressed and sent to the transmission path, and if the input audio signal is both voiced and unvoiced, the audio signal is divided into blocks at predetermined intervals and sent to the transmission path. and further configured to suspend the transmission of the input audio signal when the input audio signal is not voiced, and restore the time length of the block corresponding to the voiced sound to the original time length at the receiving side pressure, An audio multiplex transmission system characterized in that a block of unvoiced sound is restored to its original time length according to the deletion ratio. +21 Claim No. 1 characterized in that when the Aty input audio signal is both voiced and unvoiced, the block of the audio signal is sent as it is to the transmission path.
Audio multiplex transmission method described in ). (3) When the input audio signal is both voiced and unvoiced, a certain percentage of time in a block of the audio signal is deleted and sent to the transmission path. Audio multiplex transmission method described in section 1).