KR980013233A - Method for recording and reproducing voice data in a facsimile system - Google Patents

Abstract

end. TECHNICAL FIELD [0001] The present invention relates to a technique for recording voice data on a recording sheet and reproducing the voice data in a facsimile having an automatic response function.

I. In order to prevent the loss of voice data due to the limitation of the capacity of the memory provided in the automatic answering device of the facsimile machine, the voice data is recorded on the recording paper and then reproduced.

All. SUMMARY OF THE INVENTION A key point of the solution of the invention is to convert a voice signal received through a telephone line into a digital signal and compress it, print the compressed voice data on a recording paper in a predetermined form, Data is scanned, the stunning image data is compensated for vertical ringing, and the compensated data of the vertical ringing is converted into voice data, and voice synthesis is performed to reproduce the voice signal.

la. An important application of the invention is applied to a facsimile machine having an automatic response function.

Description

Method for recording and reproducing voice data in a facsimile system

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of recording and reproducing audio data in a facsimile system, and more particularly, to a method and apparatus for recording audio data on a recording medium in a facsimile having an automatic response function and scanning and reproducing recorded audio data.

A facsimile apparatus is an office automation apparatus that can transmit or receive various documents and photographs to a remote party using a telephone line. The office automation equipment is required to have various functions as a complex office device as well as a simple function according to the development of industry and information diversification. The demand for such a complex function is a tendency that a facsimile system is being developed as a complex office equipment including an automatic response function as well as a printer function of a personal computer.

The automatic answering device can be roughly classified into an analog automatic answering device using a magnetic tape and a digital automatic answering device using a semiconductor memory such as a DRAM (Direct Random Access Memory) or an ARAM (Audio Random Access Memory) .

2. Description of the Related Art In recent years, an automatic response device has been developed which facilitates recording or reading data due to the rapid development of semiconductor memories. Conventionally, an automatic answering device using a semiconductor memory uses a volatile memory device in which recorded data is lost if power is not supplied due to a price of a semiconductor part, so that the time limitation for data storage and the storage of a large amount of data The memory capacity of the memory element is limited.

It is therefore an object of the present invention to provide an audio data recording and reproducing method for recording and reproducing audio data on recording paper in a facsimile apparatus in order to solve the above problems.

Another object of the present invention is to provide a method and apparatus for recording and reproducing audio data in which a facsimile apparatus can print audio data on recording paper and use the recording paper as auxiliary storage means.

1 is a block diagram of a facsimile apparatus having an automatic response function according to an embodiment of the present invention;

2 is an exemplary view showing an image reproduction area in a facsimile system;

3 is an exemplary view showing a recording area for an A4 size recording sheet in a facsimile system;

4 is an exemplary view of an information recording area according to an embodiment of the present invention;

FIG. 5 is a format of voice data recording according to an embodiment of the present invention; FIG.

6 is a specific recording format of voice data according to an embodiment of the present invention;

FIG. 7 is an exemplary view of voice data printed on a recording sheet of an embodiment of the present invention; FIG.

8 is a control flowchart for recording voice data for a recording medium according to an embodiment of the present invention.

9 is a control flowchart for reproducing voice data printed on recording paper according to an embodiment of the present invention.

10 is a flow diagram of a vertical screen echo compensation subroutine in accordance with an embodiment of the present invention.

11 is a flowchart of a speech data restoration subroutine according to an embodiment of the present invention.

12 is a flowchart of a horizontal error compensation subroutine in accordance with an embodiment of the present invention.

13 is a flow diagram of a dual horizontal error compensation subroutine in accordance with an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a facsimile apparatus having an automatic response function according to an embodiment of the present invention. The CPU 10 controls the overall operation of the facsimile according to the set program, controls to record the received voice signal on the recording paper, and controls to reproduce the voice signal scanned from the recording paper. The memory 12 stores a program for recording and reproducing voice data and voice data, and temporarily stores data generated during program execution. The OPE (OPERATING PANEL) 14 has a plurality of keys, and applies different key data to the pressing of this key, and displays the operating state under the control of the CPU 10. [ The scanner 16 scans the original, converts the image of the original into binarized information, and applies the binarized information to the CPU 10. Under the control of the CPU 10, the modem 18 analog-modulates the transmission data and demodulates and outputs the analog reception data. An NCU (Network Control Unit) 20 is operated under the control of the CPI 10 to form a call loop of the telephone line and to interface the signal of the modem 18 with the signal of the telephone line. The printer 22 prints image data on a recording sheet under the control of the CPU 10. [ The codec 24 converts the analog voice signal input through the NCU 20 into a digital signal, and outputs the digital voice signal to the voice processing unit 26, converts the digital voice signal into an analog voice signal, and outputs the analog voice signal to the NCU 20. The audio amplifier 28 amplifies the audio signal converted from the analog signal from the codec 24 to a predetermined level, and outputs the amplified audio signal to the speaker 30. The voice processing unit 26 is implemented by a DSP (Digital Signal Processor), and can be implemented using, for example, D6375A of DSP. The voice processing unit 26 synthesizes a voice message stored in the memory 12 under the control of the CPU 10, , And compresses the digitized voice signal from the codec 24 and stores the compressed voice signal in the memory 12.

2 is an exemplary view showing an image reproduction area in a facsimile system.

3 is an exemplary view showing a recording area for an A4 size recording paper in a facsimile system.

4 is an exemplary view of an information recording area according to an embodiment of the present invention.

5 is a view showing a format of voice data recording according to an embodiment of the present invention. a is a data line, b is a margin between lines, c is a left margin, d is a start bit, e is a data bit, f is a stop bit (parity bit), and g is a right margin.

6 is a specific recording format of voice data according to an embodiment of the present invention.

7 is an illustration of audio data printed on a recording sheet of an embodiment of the present invention.

8 is a control flowchart for recording voice data for a recording medium according to an embodiment of the present invention.

9 is a control flowchart for reproducing voice data printed on recording paper according to an embodiment of the present invention.

10 is a flow chart of a vertical image jitter compensation subroutine in accordance with an embodiment of the present invention.

11 is a flowchart of a voice data recovery subroutine according to an embodiment of the present invention.

12 is a flowchart of a horizontal error compensation subroutine according to an embodiment of the present invention.

13 is a flow diagram of a dual horizontal error compensation subroutine in accordance with an embodiment of the present invention.

The operation of the preferred embodiment of the present invention will be described in detail with reference to FIGS. 1 to 13. FIG.

ISO (International Standards Originization) In the Group3 facsimile system that can send and receive A4 size (21mm × 297mm) originals, the image playback area recommended by CCITT Recommendation T.4 can be displayed in print and scan The horizontal loss and the vertical loss are considered, and it is recommended that the range of 196.6 mm × 281.46 mm is guaranteed as shown in FIG. And Table 1 and Table 2 below show CCITT REC. The horizontal loss and the vertical loss that can be generated when reading and recording the A4 size manuscript of the GROUP3 facsimile system recommended in T.4 are shown. Table 3 below shows the horizontal and vertical recording and reading resolutions.

[Table 1]

[Table 2]

[Table 3]

The reading of the image reflects the light generated from the light source on the document, and the reflected light is detected by the reading sensor and read. Charge coupled device (CCD) chips or contact image sensors (CIS) are mainly used as the read sensor, and optical sensors are arranged at the same interval as the horizontal resolution shown in Table 3. Optical sensors that read ISO A4 size (210mm × 297mm) documents typically have a read width of 216mm to read documents up to letter size (216 × 279mm), while read sensors have eight lt; RTI ID = 0.0 > mm / mm. < / RTI >

Fluorescent lamps and LED arrays are used as light sources. The recording method is classified into the thermal recording method and the plain paper method according to the recording paper to be printed. The recorded horizontal resolution is 3.815 lines / mm as standard and 7.7 lines / mm and 15.4 lines / mm as an option. Of Group3 facsimile systems are capable of recording up to 7.7 lines / mm of vertical resolution. The facsimile system according to the present invention having such a resolution has an automatic response function, and the automatic response function is the same as that conventionally used in an answering machine. According to the present invention, a voice signal received from the CALLING PART in the voice memory is stored in the voice memory by the automatic response function. The data stored in the voice memory is converted into a code form as shown in FIG.

Referring to FIG. 8, an operation of converting a voice signal stored in the voice memory into a code form as shown in FIG. 7 and printing on a recording paper will be described. First, in step 101, the CPU 10 checks whether a ring signal is detected from the NCU 20, The flow advances to step 102. In step 102, the CPU 10 checks whether the voice input mode is set by the key input unit of the OPE 14. If the voice input mode is not set, the process proceeds to step 103. [ In step 103, the CPU 10 performs a function corresponding to the set corresponding mode. That is, when the TEL mode is set, the handset is hooked off to make a telephone call. If the mode is set to the FAX mode, the FAX document is received. If the mode is set to the ANS / FAX mode, the automatic response and the FAX document are received. However, if the mode is set to the voice receiving mode in step 102, the CPU 10 controls the voice processing unit 26 to read the voice signal stored in the memory 12, and then transmits a voice synthesis voice guidance message (OUT GOING MESSAGE: OGM) . The transmitted voice guidance message is converted into an analog signal through the CODEC 24 and transmitted to the CALLING PART through the NCU 20. When the user of the calling party inputs a voice message after transmitting the voice guidance message, the CPU 10 receives the voice data in step 105. [ At this time, the received voice data is received through the NCU 20, converted into a digital signal from the codec 24, and applied to the voice processing unit 26. Then, in step 106, the CPU 10 controls the voice processing unit 26 to compress the digitally-converted voice data and store the compressed voice data in the memory 12. Then, in step 107, it is determined whether or not the reception of the voice data has been completed. If the voice data reception has not been completed, the procedure returns to step 105 to repeat the above-described operation to receive and store the voice data. However, if the voice data reception has been completed, the CPU 10 proceeds to operation 108 and reads the voice data stored in the memory 12, and controls the printer 22 to perform printing. In this case, the audio data is printed on A4-size recording paper as shown in FIG. 3, where a is an upper margin, b is a lower margin, c is a left margin, and d is a right margin. The recording unit, f, is designed to record the audio data recording unit. The printing section is composed of an information recording section in which a title and information for recorded audio data are recorded, and a data recording section in which audio data is recorded. The margin is determined by taking into consideration the effective printing area and the effective scanning area given by the structural characteristics of the facsimile system.

As shown in Fig. 4, the information recording section is constituted by an information recording reference line of a, a heading of print contents which is an information recording area of b, page information of recorded data, a data recording start line of c, . The actual audio data recording area of the recording sheet as shown in FIG. 3 is printed with the data of the form shown in FIG. 5, and the audio data is recorded in the space data of 0 for the binary data of 0 and 1, Data limits the method of recording with bar (BAR). A line consisting of a space and a bar is recorded as data combined with 0 and 1, and data is continuously recorded on a new line when the recording of consecutive data exceeds a line. In recording the data on a new line, a blank space b is inserted between the lines on which data is recorded for discrimination from the data on the previous line. One data line a has a left margin c and a right margin g as margins, and the data recording area consists of a start bit d, a data bit c, and an end bit f. The start bit and the end bit each have information of one bit, and the start bit is indicated by the information of the bar to indicate that the data of one line is started. The end bit indicates that the data of one line is terminated and the bit error that may occur when the image is read as a parity bit is detected and the bit error of the even parity (EVEN PARITY) for the bar data of the voice data area is corrected Information. As shown in FIG. 5, the voice data area is directly printed with spaces for 0 and 1 saturation data, respectively. The size of one bit recording unit for the data printed as shown in FIG. 5 was applied in consideration of the recording resolution of the facsimile system, and the left and right and up and down rings of the original in the image reading.

The size of one-bit recording unit with respect to the horizontal and vertical resolutions of the printed data is shown in FIG. In FIG. 6, the size of the recording unit for one bit is 4 pels as a unit of one bit for a resolution of 8 pels / mm horizontally, and 5 lines for a resolution of 7.7 Lines / Respectively. In order to distinguish between the data lines, 3 lines were applied vertically to the inserted margins. FIG. 7 shows a chart of actual data printed on the recording sheet by the printer 22 after reading the audio data recorded in the memory 12 in this way.

Referring to FIGS. 9 to 13, the operation of scanning the printed document after printing the audio data in the space or bar form on the recording paper as described above and reproducing it again as a voice signal will now be described. In step 201, It is determined whether a voice reproduction key is input from the key input unit. At this time, if the voice reproduction key is inputted, the CPU 10 proceeds to step 202 and checks whether the document is inserted in the document table. If the original is not inserted into the original glass, the CPU 10 proceeds to step 203 and generates an alarm sound. In step 204, the CPU 10 displays a document loading request message to load the document on the display device of the OPE 14. However, if the original is inserted into the original glass, the CPU 10 controls the scanner 16 in step 205 to read the image data of the original and receive the binarized image data. At this time, the image reading of the document is generally carried out by a method of reading the image data by transferring the document to the reading device in which the document feeding roller driven by the stepping motor is fixed. In the mechanism in which the original is transferred to the reading device by the original conveying roller, the original sound of the original is generated due to the mechanical characteristics of the system, so that the audio data can not be restored using the read binary image data. Accordingly, in step 206, the CPU 10 compensates for vertical noise of the read-out binarized image data. In order to compensate for the readout of the binarized image data, there is a reference line for detecting the echo information of the image data. As shown in FIG. 4A, Compensate. A specific operation for compensating for the vertical echo is disclosed in Fig. In step 301, the CPU 10 stores the image data of the scanned original in the memory 12. Here, the image storage area of the memory 12 can store image data of 28 vertical lines. Then, in step 302, the CPU 10 detects a compensating reference line from the stored image data, and in step 303, detects vertical bell information for each horizontal pixel, that is, a compensation reference value for each horizontal pixel. The maximum compensatable reference value for the vertical ringing of the data stored in the memory 12 is the CCITT REC. It is possible to compensate for up to 3.6mm of vertical sounding as recommended in T.4. The detection of vertical sounding detects the compensation value by taking the vertical position value of the compensation reference line as the value of the echo. In step 304, the CPU 10 checks whether the detected vertical ringing noise compensation value exceeds a maximum compensatable reference value, and proceeds to step 305 if it exceeds the maximum compensatable reference value. In step 305, the CPU 10 displays an alarm sound for re-reading the image or a message on the display device of the OPE 14 so that the user can re-insert the original to re-read the image of the original. However, if the echo compensation value does not exceed the maximum compensatable reference value, the CPU 10 compensates the jitter for the 28 vertical line image data stored in the memory 12 in step 306. In order to compensate for such a bell, the jitter of the data appearing in the read image data by transferring the original to the original conveying roller continuously occurs during the entire process of passing the original through the reading unit. Therefore, The degree of echo is not constant even though the echo is compensated. As a result, it is possible to compensate for the data recovery by compensating the jitter using the compensation reference line located at the top of the document. Therefore, in step 307, the CPU 10 stores the compensated data of the vertical ringing into the buffer for the picture window to convert the data into voice data. Then, in step 308, the CPU 10 checks whether or not 15 lines of data are stored in the image window buffer, and proceeds to step 312 if 15 lines of data are not stored. In step 312, the image data stored in the buffer for the window is checked for the last line, and if it is the last line, the vertical buzzer compensation is terminated. If it is not the last line, the CPU 10 proceeds to step 313 and the CPU 10 reads the next line of image data and proceeds to step 314. In step 314, the CPU 10 compensates for the vertical ringing of the read image data, and stores it in the memory 12 in step 315. Then, in step 316, the CPU 10 stores the data compensated for the vertical ringing into the image window buffer for conversion into voice data, and returns to step 308 to repeat the above-mentioned operation to compensate for the vertical ringing of the read image data . The area for storing the 15 vertical lines of the window buffer is a value set to compensate vertical lines for each of the upper and lower five lines for one data line composed of five vertical lines. After compensating the vertical ringing of the read image data, the CPU 10 proceeds to step 207 in FIG. 9 to restore the voice data.

The operation of restoring the data will be described in detail with reference to FIG. 11. In step 401, fifteen vertical line data are compared with respect to each pixel from the first pixel to the 1728th pixel on the horizontal side. In step 402, the CPU 10 determines that a bit is detected when black data is successively detected in five vertical lines with the comparison result of the 15 vertical line data. And determines that the start bit is detected when the detected vertical black data is horizontally positioned four or more times continuously. Then, in step 405, the CPU 10 detects the data bits using the sequential decimal numbers of black and white continuously located in the horizontal direction after the start bit is detected. At this time, if a data bit is detected, the CPU 10 analyzes the position of the vertical line in which data exists among the 15 vertical lines with respect to 1728 horizontal pixels from the window buffer in step 404. When the position of the vertical line for each horizontal pixel is detected by the above analysis, the values of the uppermost vertical line Ymin and the lowest vertical line Ymax at which the data are located are respectively analyzed. The center value Ycen of the position of the data is calculated by averaging the values of the highest vertical line Ymin and the lowest vertical line Ymax as shown in Equation 1 below. Then, in step 411, the CPU 10 sets the center value (Ycen) and the shift value (Yshift) of the buffer line as mathematical expressions 1 and 2.

[Equation 1]

&Quot; (2) "

Yshift = Ycen

Here, the shift of the buffer line shifts the buffer by a shift value calculated from top to bottom with respect to the vertical line of the image data by moving the window for reading the data. When the buffer is moved, new Mars data is read, stored in the buffer, and the voice data is read. If the data bit is detected in step 403, the process proceeds to step 404 to detect the vertical position of the data, and the process proceeds to step 411. If a data bit is detected in step 403, the process proceeds to step 405, where one bit of data information is composed of four horizontal pixels as shown in FIG. 6, so that the number of consecutive pixels is counted. Then, in step 406, the CPU 10 compensates for the horizontal reading error.

12, the CPU 10 divides the value obtained by counting the number of consecutive pixels by 4 in step 501, and in step 502, the CPU 10 checks the rest. At this time, if the remainder is 0 in step 503, the process proceeds to step 510, and the CPU 10 divides the value obtained by counting the number of consecutive pixels by 4 and converts it into 0 or 1 audio data information corresponding to the quotient. However, if the remainder is 1 in step 504, the CPU 10 proceeds to step 505, and the CPU 10 subtracts 1 from the count value of the consecutive pixels, and then proceeds to step 510 and divides the subtracted value by 4 to obtain 0 or 1 And converts it into voice data information. If the remainder is 3 in step 506, the process proceeds to step 509. In step 510, the CPU 10 adds 1 to the count value of the consecutive pixels, and then divides the added value by 4 to obtain 0 or 1 And converts it into voice data information. If the remainder is 2 in step 506, the flow proceeds to step 507 to perform the double compensation subroutine, and then proceeds to step 510. Referring to FIG. 13, the double compensation subroutine operation is referred to as an N-th successive pixel, and the compensation value is determined with reference to the number of consecutive N-1th and N + 1th pixels.

In step 601, the current continuous pixel value is determined as the Nth pixel value, and in step 602, the (N-1) th continuous pixel compensation value is checked. If the compensation value is 0 in step 603, the process proceeds to step 604 to check the (N + 1) th successive pixel compensation value. However, if it is determined in step 602 that the N-1th burning high calcination compensation value is less than 0 in step 605, the flow advances to step 606. In step 606, 2 is added to the count value of the number of consecutive pixels. Then, in step 510, the value obtained by adding 2 to the count value of the number of consecutive pixels is divided by 4 and converted into audio data information of 0 or 1 corresponding to the quotient. However, if the continuous pixel compensation value is greater than 0, the process proceeds to step 608 where the count value is subtracted from the count value of consecutive pixels. Then, in step 510, the value obtained by subtracting 2 from the count value of the number of consecutive pixels is divided by 4 and converted into 0 or 1 audio data information corresponding to the quotient. When the data bit is detected in step 511, the number of detected bits is counted and it is checked whether the number of bits recorded in one data line is completed. That is, it searches whether N is 0 or not. If N is not 0, the information of 1 bit of the voice data converted into 0 or 1 is stored in the memory 12, and the process proceeds to step 513. In step 513, the CPU 10 decrements N by 1, returns to step 511, and repeats the above-described operations until the number of bits detected in one data line is completed.

After compensating for the horizontal error, the CPU 10 counts the number of bits of the audio data detected in the audio data line in step 407 of FIG. Then, in step 408, it is determined whether the end bit is detected from a value obtained by counting the number of bits of the voice data. If the end bit is not detected at this time, the process returns to step 403, and the above-described operation is repeated until the end bit is detected until step 407 is performed. The end bit is a bit in which the number of bits recorded in one data line is completed. If the end bit is detected, the CPU 10 proceeds to step 409 where the CPU 10 compares the number of 1's detected in the current data line to find whether a parity error occurs. When the parity error occurs, the CPU 10 displays an alarm sound for re-reading the image or a message on the display device of the OPE 14 so that the user can re-insert the original to re-read the image of the original. However, if the parity error does not occur, the center value Ycen of the data position is averaged by averaging the most significant vertical line Ymin and the lowest vertical line Ymax, and then, in step 411, Set the shift value of the line (Yshift).

After the audio data signal conversion for the read image signal for each data line is completed as described above, the audio data converted in step 208 of FIG. 9 is stored in the memory 12, and the process proceeds to step 209. FIG. In step 209, the CPU 10 reads the converted voice data from the memory 12, controls the voice processor 26 to voice-synthesize the voice data, and then converts the voice data into an analog signal through the codec 24. The analog signal-converted audio signal is amplified by the audio amplifier 28 and sent to the speaker 30 for reproduction.

Also, a personal computer and a facsimile system are connected and simulated in order to test whether a desired effect can be obtained in implementing the present invention. Experimental recording data were recorded using random function. The evaluation of the experiment was evaluated on the data recording capacity for one ISO A4 recording paper and the reproduction rate for the reproduction algorithm on the data recording algorithm. The readback rate was evaluated by comparing the recovered data with the initial recorded data and the bit error. The recording capacity of data for one recording sheet of ISO A4 size was determined by the recording area where data is recorded and the size of the recording unit. The recording area for the recording sheet of ISO A4 size is divided as shown in FIG. 3, and the size of each area applied in the present invention is shown in Table 4 below.

[Table 4]

[Table 5]

Table 5 shows the horizontal and vertical sizes of the unit bits of the data record and the size of the lag space.

The data recording capacity for the recording sheet of the ISO A4 size is shown in Table 6 below.

[Table 6]

As shown in Table 6, it is possible to record 392 bits of data per one data line with respect to the A4 recording sheet, and 269 data lines can be recorded with respect to one recording sheet, so that a single A4 size recording sheet has about 13.1 Kbytes Quot; is recorded. A voice processing unit 26 for performing a digital automatic response is configured to store voice signals in the memory 12 and store the voice data sampled at 8 KHz and modulated by the PCM method in an 8-bit digital signal into a 4-Mbit memory for about 15 to 17 minutes And a compression ratio that can store data of the data. Therefore, the voice data stored in the memory 12 using the DSP chip of the D6375A can record voice signal data of about 25 seconds for one recording sheet of the ISO A4 size. Fig. 7 is an exemplary view of print data that is recorded on an actual recording paper and enlarged to 208%. In addition, experiments of the regeneration algorithm recorded 100 kinds of random data on the recording paper and read the recorded data 10 times each time to reproduce the data. Also, the recording characteristics of the recording paper were compared using plain paper and thermal recording paper, respectively. Experimental results were evaluated by comparing the number of occurrences of parity errors generated during data reproduction with the bit error between the regenerated data and the initial recorded data. As shown in Table 7 below, the experiment was carried out to read out 1000 readings of each 100 recording data, 10 readings per recording paper.

[Table 7]

Experimental Results As shown in Table 7, the reading error rate was 0.6% due to 6 reading errors on the plain paper, and the reading error rate was 1.4% due to 14 reading errors on the thermal recording paper. However, when the reading error occurs, the paper is correctly read, and no reading error occurs when the paper is tested again. As for the difference in reading ratio among the recording papers, in the case of plain papers, the image data was less likely to be read due to the characteristics of the recording papers. In the thermal papers, Because. However, the algorithm of the present invention for data recording and reproduction has different reproduction results depending on the mechanical characteristics of the recording paper and the system, but is not a problem in practical use.

As described above, according to the present invention, voice signals received in a facsimile having an automatic response function are stored in a memory, and then voice data stored in a memory is printed and stored on a recording paper, and the stored recording paper is scanned and stored in a memory It is possible to store and reproduce all the received voice data regardless of the capacity of the memory for storing the voice and to prevent the loss of the voice data received due to the limitation of the memory capacity There is an advantage.

Claims (22)

A facsimile system having an automatic response function, characterized by converting a voice signal received through a telephone line into a digital signal, compressing the compressed voice data, and then printing the compressed voice data on a recording paper in a predetermined form. Data recording method. The method according to claim 1, wherein the voice data is printed with data having a space and a bar section to be recorded. 3. The method according to claim 2, wherein the print form of the recording paper is printed so as to have an upper margin, a lower margin, a left margin, and a right margin in the recording paper. 4. The facsimile machine according to claim 3, wherein the print form of the recording paper comprises an information recording area for displaying information of the data to be printed on a top of the recording paper and a data recording part for printing audio data. Data recording method. 5. The method according to claim 4, wherein the information recording area comprises an information recording reference line, an information recording area, and a data recording start line. 6. The method according to claim 5, wherein the information recording area comprises a title for audio data, a date and time of receiving audio data, and page information of data to be recorded. 5. The method of claim 4, wherein the data recording unit has a margin between a data line and a line. 8. The method of claim 7, wherein the data line has a start bit, a data bit, and an end bit. 9. The method of claim 8, wherein the start bit is a bit of data represented by information of a bar. 9. The method according to claim 8, wherein 0 of the data bits is printed in a space and 1 is printed immediately. 9. The method of claim 8, wherein the recording unit size of the start bit, data bit, and end bit is 4 pels for a recording resolution of 8 pels / mm horizontally, and 5 lines for a recording resolution of 7.7 lines / And the information is printed. 8. The method of claim 7, wherein the laminar space is represented by a space line of 3 lines for a recording resolution of 7.7 lines / mm vertically. In a facsimile system having an automatic response function, audio data recorded on a recording paper in a preset recording format is scanned and stored in a memory, and the stored audio data is read, Characterized in that the method comprises the steps of: 14. The method of claim 13, further comprising compensating for vertical firing of the scanned image data when scanning the recorded voice data. 1. A facsimile system having an automatic response function, comprising: scanning data printed on a recording paper in a preset recording format; compensating vertical scanning of the scanned image data; Converting the converted voice data into voice data, and reproducing voice data by synthesizing the converted voice data. The method as claimed in claim 15, wherein the vertical echo compensation step comprises the steps of: detecting the scanned information recording reference line; detecting a vertical echo compensation value based on the detected information recording reference value; And compensating for the vertical ringing of the scanned image data when the value of the scanned image data is smaller than the set reference value. 17. The method of claim 16, further comprising the step of re-scanning the image data when the detected vertical ringing compensation value is greater than a preset reference value. Further comprising the step of compensating for the vertical buzzing of the upper and lower five lines for one data line composed of five vertical lines by covering the window with the image data compensated for the vertical buzzing, In the audio data. The method as claimed in claim 15, wherein the step of converting the voice data comprises the steps of: detecting a start bit and a data bit from the image data compensated by the secondary vertical sync ring; detecting a vertical position of the data line after detecting the data bit Converting the voice data into voice data; and converting the voice data into voice data. The method of claim 19, further comprising the steps of: counting the number of consecutive horizontal pixels after detecting the data bits and compensating for the horizontal error; and counting the number of audio data bits after the horizontal error compensation to detect the end bit Wherein the voice data is reproduced by the facsimile machine. 21. The method as claimed in claim 20, further comprising the step of detecting the presence or absence of a parity error after the detection of the end bit and rescanning the image data when a parity error is detected. The method as claimed in claim 19, wherein the step of detecting a horizontal error comprises the steps of: detecting a remainder obtained by dividing a value obtained by counting the number of consecutive horizontal pixel bits by 4; Subtracting 4 from the counted value, detecting a quotient by adding 1 to the counted value and dividing the counted value by 4 when the detected remainder is 3; Subtracts 2 from the counted value and divides by 4 if the pixel compensation value of the bit is smaller than 0, and subtracts 2 from the counted value and divides by 4 if the (N-1) And converting the detected quotient to binarized voice data. The method of claim 1, ※ Note: It is disclosed by the contents of the first application.