JPS5939135A - Signal transmitting system - Google Patents

Abstract

PURPOSE:To reduce effectively a quantized noise and to attain the signal transmission with high fidelity of reproducing waveform and energy of an original signal, by giving a schedule data corresponding to a mean value to a lower bit without an effective bit. CONSTITUTION:Suppose that an analog signal is A/D-converted at the transmission side into a spare converting data in 15-bit. In transmitting a main data in 8-bit based on it, the upper 8-bit of the effective bit is extracted as a main data and the lower 4-bit is thrown away. The main data is shifted for the bit to produce the reproducing data in 15-it according to the scale information in response to the bit location taking the main data in 8-bit in the spare converting data at the receiving side. In this case, a data mean value ''0111'' is applied to the lower 4-bit. Thus, the error with the original data is less and the error in waveform transmission and energy transmission is less.

Description

[Detailed Description of the Invention] [Technical Field of the Invention] The present invention digitizes analog signals to create a transmission path,
σ) relates to a signal transmission system that is applied to a transmission system such as a demodulation system or a recording/reproduction system, receives a transmission signal from this transmission system, and returns it to an analog signal.

[Technical background of the invention]

Analog signal - digitized by iPcM (pulse code modulation) and transmitted by a transmission system (in this case, "transmission")
``Transmission'' in a broad sense means not only transmission through a transmission line, but also transmission through a variable/distorted light, recording/reproducing system, etc. ) When receiving this signal and converting it back to an analog signal, 1. Quantization noise occurs during PCM conversion of the analog signal. In order to reduce this quantization noise, it is generally believed that there is no suitable method other than increasing the number of bits of the digital signal during digitization in PCM conversion.

On the other hand, in the case of the transmitted analog signal, for example Oti No. 1g, the idea is that if the noise is no longer noticeable to the auditory sense, then it can be considered that the noise has been substantially reduced. There is a method to be applied before and after the noise reduction method 2PCM transmission system for analog 1g.

That is, as shown in FIG.
l) The (analog-digital) converter 2 performs digital processing, and the transmitting side signal processing circuit 3 performs code error detection and
After being subjected to the necessary processing for retention and data compression, it is sent to a transmission system such as a transmission path, a v/demodulation system, and a recording/playback system. The PCM signal received from the transmission system is subjected to processing such as code error = r, +E processing, and optical processing to decompress compressed data in the receiving side processing circuit 4, and is converted into D/A (digital-analog). After being converted into analog by the converter 5, the noise reduction encoder 3 corresponding to the above noise reduction encoder 3 is
It is decoded by the 1st reduction decoder 6 and output as audio output from 〇 to the output terminal.No 1. It is used for playback from speakers, etc.

Quantization noise 2 is audible because it is less than 1.
/ Analog No. 16 manually input to D converter 2 is A/
It is preferable that the tl be below the maximum value that can be graded by the D converter 2 and close to the nuclear maximum (if an analog signal exceeding the maximum value is input to the A/D converter 2, the signal will be clipped. Clipping distortion occurs.)
0, that is, it is desirable that the input Ig of the A/D converter 2 has a substantially constant value and a value sufficiently close to the maximum value that is equal to or less than the above-mentioned maximum value.

On the other hand, if a conventional noise reduction method for an analog system is applied, the same problem occurs.

(If the principle of the conventional general noise reduction method is that the change in signal level is reduced by at least 1, the analog output level on the encode side remains unchanged due to the decoding fql function. However, in conventional noise reduction methods, the signal is transmitted by increasing the noise generated in the transmission system such as the recording medium (8-bit recording), but in this case, the encoding In 1411, the large signal remains the same and the level is increased by one small multiple, and the small signal is returned to the original state on the decoding side.At this time, the transmitted tW@ is only a signal whose dynamic range has been compressed, so the decoding is performed by adjusting the level of the transmitted signal. For this reason, even when the entire level of the small signal increases during encoding, it must be lower than that of the large signal.Therefore, the encoded signal 1kA/
1. ) converter, the input power of the A/I) converter cannot be kept approximately constant. For these reasons, conventional noise reduction methods cannot sufficiently reduce the entire shadow I# of quantization noise.

(Example) The control time constant must be made large because the decoder side obtains control information based on the level of the encoded signal, that is, the signal input to the decoder.For example, the control time constant must be large. In this case, in order to obtain the correct signal level without causing ripples in the control signal even at the lower limit of the 5J hearing frequency, it is necessary to have a time constant of sufficient value even for such a low frequency. The time constant is also determined according to this.If the control time constant is thick like this, for example, if the attack is a signal that is sudden, such as the sound of a piano, the encoder side will not be able to control it in time, and the A/D converter will be overloaded. A level signal is applied to generate clipping distortion.

(c) Due to the cause of (b) above, there is a problem on the encoder side.
If clipping occurs in the transmission system, the level relationship cannot be transmitted correctly, and not only will the clipped waveform be played back on the decoder side, but in many cases, the level of the non-clipped portion will also be distorted. This is caused by the fact that the correct control voltage cannot be obtained on the decoder side due to the clipping phenomenon, which causes problems in sound quality such as piano attack.

Therefore, when transmitting a signal by converting the analog line analog multiplier into a digitized signal and transmitting it back to the original analog signal, we decided to
The following method was previously proposed as a signal transmission method capable of effectively reducing noise components such as 9'1# child noise associated with M transmission (% Application No. 57-130620).

In other words, the method involves converting an analog signal into a digital value and applying first level variable control so that the final digital value is a substantially constant value to obtain a first digital official name, that is, king data. At the same time, the control information in the first level variable control (hereinafter referred to as "scale information") is converted into a second digital signal consisting of a digital domain with a smaller number of bits than the first digital signal, that is, a scale/digital signal. L, ttv information data, synthesize these main data and sfur information data in a ratio of the number of data such that the scale information data is sufficiently smaller than the main data, give it to the transmission system, and transmit this synthesis from the transmission system. The signal reception is carried out by carefully separating and extracting the main data and the scale information data, converting the extracted main data into analog data, and based on the extracted scale information data, substantially converting the signal into the first level β. This signal transmission method is characterized in that the original analog signal is restored by applying a second level variable control having adverse characteristics.

When specifically implementing this signal transmission method, it is considered that a method in which the first and second stop/bell variable control is performed digitally is both practical and preferable.
In the above proposal, a configuration as shown in FIG. 2 was presented as an example of such a system.

In this case, on the transmitting side, the input analog 4N signal consisting of, for example, an audio signal is converted into digital preliminary conversion data of a sufficient number of bits, for example, 15 bits, at scheduled time intervals using the all-A/D converter 7, and then the digital level detecting means 8 Digitally detect the maximum level within the predetermined intercostal space or a level approximately equivalent thereto, obtain, for example, 4-bit scale information data, and then perform digital level control ten steps 9.
Now, based on the scale information data, digitally control the level of the output preliminary conversion data of the A/D converter 7 and compress the data to obtain, for example, 8-bit king data, and combine this main data with the scale information data. In the synthesizing means 10, one piece of scale information data corresponds to a large number of main data and is sent to the transmission system. on the other hand,
The receiver 1Ill separates and extracts the main data and scale information data from the transmission signal received from the transmission system in the separation means 11, and extracts the main data and scale information data from the transmission signal received from the transmission system in the digital level variable control means 12 based on the above-mentioned scale information data. Digital level variable control (data expansion) with control characteristics opposite to
1) /A converter 3 converts it into an analog signal to obtain an output analog signal.

The digital level detection means 8 detects the digital level by detecting the number of effective bits in the pre-modified/compressed data, that is, the most significant bit position IW of the effective bits excluding the code billatra. Te1 in digital level variable system 6'll1 means 9
The digital level variable control is as described above, and the most significant effective bit position is the corresponding bit position. It seems practical and effective to do this by extracting the converted data and creating a king data association. In this case, for example, if we assume that the diagonal lines in the 15-bit preliminary conversion data are valid bits as shown in Figures 3(a) to (C), then the valid bits are valid in the case of (=l) in the same figure. If 6 bits of the preliminary conversion data are occupied and 8 bits of main data are used, the lower 8 bits may be used as raw data.

At this time, the position where the main data is taken is the 8 bits below the 8th bit counting from the most significant bit, so this corresponds to shifting the preliminary conversion data 7 bits to the left (toward the upper part) and extracting only the upper 8 bits. The control level, that is, the scale information at this time is the above-mentioned shift) ILr7J. As you can see from this example, the maximum shift 1 is 7 bits, so when the effective number of bits is 8 or less, the scale information is uniformly "7".
Select. In addition, in the case of the laJ diagram (b), the number of effective bits is 9 bits, so as is clear from the figure, the scale information is 16''. lowest pitch)
(1,8B) is ignored and this part is considered an error. In the case of (C) in the same figure, the number of effective bits is 15 bits, the scale information is "0", and in this case, the lower 8 bits of the preliminary conversion data are ignored. In this way, when the number of effective bits is large, the effective bits ignored and discarded become an error, but the value is sufficiently small compared to the value of the main data. In this case, the maximum level of scale information is 8 (=2”)
Since it is a type, the scale information data only requires three pits.

In reality, the scale information data is σ), which is one data for each of a large number of child 17fti converted data, and the maximum number of corresponding preliminary converted data is calculated in advance. The scale is detected and set to blue'42 by measuring or predicting, and common scale information (shift meter) is set for the corresponding large number of preliminary conversion data, and this scale information r is set for each of the above large number of preliminary conversion data. Detected and updated.

In addition, in the above, the main data consisted of only the data extracted from the pre-converted data by bit shifting, but this is because the analog signal handled is a unipolar signal with only one positive or negative signal, and there are no code pits in the pre-converted data. is not included, or even if code pits are included, there is no need to transmit them. On the other hand, if the input analog signal is a bidirectional signal such as an audio signal in which both positive and negative signals are mixed, the preliminary conversion data itself usually includes a sign bit or a bit equivalent to it as at least the most significant bit (M2R). Since this is actually an important effective bit, it goes without saying that this one sign bit and the data that is obtained by the bit shift described above are used as king data. In other words, if the main data is 8 bits, 1 of them
Since bits are used as code pits, the code pits and 7-bit data obtained by bit shifting constitute the king data.

By the way, in this case, the digital level variable control means 9 on the receiving side shifts bits in the opposite direction to that on the transmitting side by the shift amount indicated by the scale information data separated from the transmission notes. In other words, in the example shown in Figure 3(a), the 8-bit main data is shifted to the right (lower) by 7 bits to obtain the reproduced data with the same number of bits as the preliminary conversion data. 15 bits of reproduced data equal to the pre-converted data are obtained. I
In the example of NN diagram (b), the main data of vc8 bits? 6
Bit shifted and 1 bit of additional data in the lower part? addition 1
Create 5-bit playback data, and add 6-bit additional data to the lower part of the μ8-bit main data in the example in the same figure (C).
Create all 5-bit playback data. Here, the data added to the lower order l-[general K tj "0", "oooo
oo” etc.) 0 data will be used.

However, if 0 data is used as additional data in this way, if there are many valid bits, the F pit will always be discarded, and if the original analog signal has a high level (large amplitude), the Nisono 1. Since a lower-level 1st pitch sound will be played, the reproduced analog signal will always be smaller than the human-powered analog signal except when the signal is small, and the energy transmission loss over a given period of time may become a value that cannot be ignored. . Furthermore, for the same reason, the rise of a steep waveform on the reproduction side is always small and accompanied by a time delay, resulting in deterioration of the waveform due to transmission. In particular, when transmitting a signal such as an audio signal that requires faithful reproduction of the waveform and signal energy, the above-mentioned problem cannot be ignored. Furthermore, if the scale information changes, the number of lower-order additional bits (the number of truncated bits during transmission) changes accordingly, and the amount of error does not change significantly as the scale information changes. On the fidelity of large y>qq=
There is.

These problems specifically affect the transmission of low frequency components of audio signals, for example.

[Purpose of the invention]

The present invention effectively reduces the digitization noise that accompanies PGM transmission when transmitting analog signal sound digitally and transmitting it back to the original analog signal, and also reduces the waveform and energy of the original signal. A signal transmission method with high reproduction fidelity? is intended to provide.

[Summary of the invention]

A first aspect of the present invention is to obtain preliminary conversion data by converting an analog signal into digital data at scheduled time intervals;
Scale information t1 that determines the extraction bit position when extracting a part of this preliminary conversion data to create main data The settings are updated so that they almost correspond, and the data of the scheduled number of bits is sequentially extracted from the bit position corresponding to the scale information of the preliminary conversion data, and the upper effective bits of the preliminary conversion data are generally given priority 7'. In addition to obtaining main data with a smaller number of bits than the preliminary conversion data, the scale information is made into scale + * surplus φ data consisting of digital data with a smaller number of bits than the main data, and the scale information for each setting history is changed. The data is inserted between a large number of main data and given to the transmission system, and upon receiving this transmission signal from the transmission system, the raw data and the scale information data are separated. Shift, and apply the planned data corresponding to the average value of all the values of all the bits that can be expressed by the number of bits to the lower bit part where no effective bit exists as a result of this bit shift, and perform the preliminary transformation described above. This is a signal transmission system characterized by obtaining reproduced data' with the same number of bits as the data, and converting this reproduced data into an analog signal to obtain a reproduced analog signal.

Further, a second invention according to the present invention is to obtain preliminary conversion data by converting an analog signal into digital data at scheduled time intervals, and to extract bit positions when extracting a part of this preliminary conversion data to create main data. The scale information sound to be determined is updated at a time interval that is sufficiently longer than the scheduled time interval so as to approximately correspond to the effective number of bits of the preliminary conversion data near the core point, and the above of the preliminary conversion data is updated. The data of the scheduled number of bits is extracted one by one from the bit position corresponding to the scale information, and the scheduled data of the same number of bits is subtracted from the remaining lower residual data taken out, and the obtained data is sequentially added to the accumulated data. If the absolute value of the value exceeds the scheduled value, add or subtract a predetermined value to the extracted data of the scheduled number of bits, and set the accumulated value t-IJ on the above-mentioned scale information every time the scale information changes, and perform the preliminary conversion. Is the main data with a smaller number of bits than the preliminary f-converted data, with the upper effective pits of the data generally given priority? With the above scale tyta'e above king data less pi 15
scale information data consisting of digital data,
The scale information data for each setting update is inserted between a large number of main data and given to the transmission system, and upon receiving this transmission signal from the transmission system, the main data and scale information data are partially separated, The bits are shifted based on the information data, and the resulting lower bit part where no valid bits are present is applied to the above scheduled data 1 to obtain reproduced data with the same number of bits as the preliminary conversion data, and this reproduced data is converted into an all-analog signal. This is a signal transmission method that is characterized by converting and reproducing analog signals.

The third aspect of the present invention is to obtain preliminary conversion data by converting analog signals into digital data at scheduled time intervals, and extracting a part of this preliminary conversion data to create main data. The scale information that determines the bit position is added to the effective number of bits of the preliminary conversion data near the relevant point in time at a time interval that is sufficiently longer than the scheduled time interval.
He updated the settings to correspond, and in reality, the data of the planned number of bits was sequentially extracted from the bit position corresponding to the scale information of the preliminary conversion data, and the same number of bits was planned from the lower residual data. The data r is subtracted and the obtained data is added to the subsequent preliminary conversion data to obtain main data having a smaller number of bits than the preliminary conversion data in which the upper effective pits of the preliminary conversion data are generally placed first. At the same time, the scale consisting of digital data with a smaller number of bits than the upper self-scale information all generated data is used as the front axis data, and the scale information for each upper self-setting middle new data is set between the large number of main data. After receiving this transmission number r from the transmission system, the main data is separated from the Kamisagi 1 scale information data, and the main data is bit-shifted based on the scale information data. The above scheduled data t is placed in the lower bit part where no valid bit exists.
This signal transmission method is characterized in that the reproduced data is applied to obtain reproduced data with the same number of bits as the preliminary converted data, and the reproduced data is converted into an analog signal to obtain a reproduced analog signal.

[Embodiments of the invention]

First, the m1 embodiment of the present invention will be described with full reference to FIG. This embodiment is the first embodiment of the first invention.

In other words, the 15-bit preboiling data obtained by A/D converting the analog signal on the transmitting side is shown in Fig. 4 (a).
). When transmitting 8-bit king data based on this (we are considering the case where the sign bit is not considered), the upper 8 bits of the effective bits are extracted as main data, and the lower 4 bits are truncated, as shown in the figure. It will be done. This truncated portion is 4 bits, so the maximum f of data that can be truncated in this way is
1~ is II 11111+, and the minimum value is oooo"
It is. Therefore, the average value of the truncated data is II
0111 II, the receiving side receives the above main data, extracts the above 8-bit main data from the preliminary conversion data, and shifts the above king data by bits to 15 bits with a scale change according to the bit attachment. Create all playback data. At this time, if no special measures are taken, normally "oooo" is applied to the lower 4 bits as shown in FIG. 4(b), but in the present invention, as shown in FIG. 4(e), the above Data 01 i 1 '', which is the average value of the 4-bit data, is applied to the lower 4 bits.Therefore, errors with the original data are reduced, and errors in waveform transmission and energy transmission are also reduced.

This is particularly important when the scale (bit position) is switched on the receiving side according to scale information as in the present transmission system. In other words, the bit position (scale) of the 8 bits that make up the main data group in the preliminary data is switched on the communication side, but at this time, the average of the data that is truncated due to data reduction (the peak is always "0111"
...” (2 digits: 01”, 3 digits: “011”, 4 digits: 0111”, etc.). Therefore, no matter how the scale changes, the above information is used as additional data on the receiving side. By adding an average value of ``0111...1'' to the data, the error caused by the data compression using the scale vI can be minimized.

In addition, although the above description has been made for the case where the code bit is not set, if the code bit is to be taken into consideration, one bit of the main data should be used as the code bit, and the other data should be processed in the same manner as described above. Even if the number of bits of the preliminary conversion data (raw data) or the main data changes, it is sufficient to carry out one process with the same purpose as described above. In addition, the average value data according to the number of bits is "0111...+1" in a normal binary code, but it may be other values in a special digital code, so in any case, the number of bits is It is enough to use the average value of the data or Fit or tr average (, etc.).

By the way, when a method like the above-mentioned embodiment @1 is adopted, the average of all truncated parts during data compression during transmission is 16. As a result, the cumulative error of continuous data is reduced.Although there is a considerable improvement in terms of the average representation of force and energy, the reduction of the cumulative error described above is not sufficient for reproducing the waveform. In terms of loyalty, it's 11i'
, · can only be improved, and it cannot be said to be sufficient when high fidelity reproduction of waveforms such as audio signals is required (zero digit / ↑W.Especially continuous 14). A major problem remains regarding the fidelity of low-frequency signals (such as the low-frequency components of audio signals) that have a large effect.

A second embodiment of the present invention to be described next is an embodiment of the above-mentioned second invention that deals with such a problem.

Now, if the playback side adds all scheduled data as additional bits, for example "0111...n" as in the first example above (in this second invention and second embodiment, the playback side It goes without saying that average value data is preferable for the bit data, but even if it is not average value data, sufficient effects can be obtained if the additional bit data of Engyuku 1 is predetermined scheduled data. ), on the transmitting side 11, when the main data is extracted from the preliminary conversion data, the truncated data and the above scheduled data are subtracted by the percentage (if the truncated data is large, the difference is positive; if it is small, it is negative). It is possible to predict how much error will occur in the sound calculation), reception (Ill), and if the absolute value of this cumulative f night exceeds the planned value by accumulating this error, it is possible to predict the error. C5, for example 1 or 11 preliminary conversion data
Addition to the value extracted from - (For example, add or subtract 1,)
By using this as main data and transmitting it, it is possible to reduce the cumulative error on the receiving and reproducing side. Here, when the `$ product (ti!J) of the difference between the above truncated portion and the scheduled data becomes 1/2 or more of the minimum unit of evangelism information (1, sn of the main data), the data is extracted from the preliminary conversion data. It is desirable to use the data obtained by adding or subtracting 1 to the data as the main data, but if the accumulated E is more than 1, it can be set as appropriate how many additions or subtractions are made to the raw data.

Here, such a second embodiment will be described in more detail with reference to FIG. 5. Here, it is assumed that when the cumulative value (absolute f value of itA exceeds raw data J, 8 B)/2 is reflected in the main data.

In FIG. 5, reference numeral 14 is an offset f-mount register in which offset data corresponding to additional data to be added is stored on the reproduction side 1. In this case, for example, the maximum digit of the additional data, that is, the average value of 7 digits, is " 0111111”
' is holding. 15 is data in which the lower effective bits of the preliminary conversion data are truncated when the king data is created on the transmitting side, and the truncated data is bit-shifted so that the MSB matches the offset data (in this case, the bit (O data is added to the lower bits generated by the shift.) This is a 7-bit attenuator that subtracts the offset data from this data to obtain differential data. 16 is a 7-bit adder that cumulatively adds the offset data given from the offset value register J4 to the output difference data 1 of the energy reducer 15 as initial data. 7. The device J6 is reset to the offset data each time the scale information is updated. Carry and borrow output of this adder 16? Upper nC: L8B of 8 hit data lost from preliminary conversion data [Add and subtract, respectively, and use as A main data.

In such a configuration, a reset pulse is first given to the scooper 16, and the cumulative km of the scooper 16 is set to 7 digits (1): lr7*data ``011111111KIJ'' given from the offset register J4. Then, in the subtracter 15, the truncated data truncated from the preliminary conversion data for each sample is stored in the offset value register 1.
It is compared with the offset data of 117 of 4, a difference value is calculated, and this difference data is cumulatively added by the score adder 16. In this case, the truncated data is bit shifted to match the MSB position or the MSB of the offset data "0111111" (zero data is added to the lower part).
Then, a subtracter 15 performs a difference calculation. Here, the period during which the adder 16 performs the accumulation is the period until the scale value changes (4). When the scale value changes and the (iI!I) of the offset scale information data (3 bits for example 11) changes, a reset pulse is sent to the adder 16, and the adder 16 again outputs the offset data "0111111".
.

is reset, and the series of operations that correctly send the state frame value to the playback side within the same scale 1b is completed (-2, next scale f
Move on to the operation regarding the target. 14. When a carry output is generated in the converter 16, lk is added to the data extracted from the preliminary conversion data to create king data, and when the hollow output is off, the same data is subtracted by 1 and sent as main data. If the preliminary conversion data, which is the original data, is 15 bits and the main data is 8 bits, the truncated data is 7 bits or less, and in the above calculation, these 7 bits are used.

In the above description, a hardware configuration has been shown for the convenience of accumulating hits, but it goes without saying that similar processing may be performed by software processing based on routkoii1kiku.

In this way, signal transmission more faithful to ¥ than in the first embodiment can be realized.

In addition, after obtaining the truncated data, fW, '? ? The timing difference until the carry/borrow output due to Keihei is reduced by 11 to the main data is 0 in J111 theory, that is, the carry/borrow due to the truncated data is extracted when the truncated data is obtained. Is it desirable to reflect this in the data?Jll] It is considered that a delay of one sample (one data) is unavoidable, and a significant effect can be obtained even if the delay is delayed by several samples. And a certain willow 1 like this
Allowing a delay of 1 (-) has the advantage that the device can be constructed at low cost.

Furthermore, the third ¥MfII of the present invention can further improve the second embodiment, and a large effect can be obtained with relatively fFrl vehicle processing. This third embodiment is an embodiment of the third invention. In the second embodiment, when the value of scale information, that is, the scale value changes, the warmer 16 is reset each time, and the cumulative total offset is set. However, there is no need to reset iL according to the third embodiment described here.
With pH, 15-bit preliminary conversion data can be transmitted to the receiving side with high fidelity and reproducibility while transmitting, for example, 8-bit main data. In this case, of course, there is no way that 15-bit precision information can be realized by transmitting the 8-bit king data once, and the data that is sent and received by the transmission of the 8-bit main data of the digit (b) is Cumulative f of the original data (preliminary conversion data) on the sending side
It sends messages that match the target. As a result, high-fidelity transmission can be realized, such as being able to transmit low frequency components such as the bass region of an audio signal with extremely high accuracy. This is because the cumulative nI value is the same between sending and receiving data, so the average value of the received raw data is the same as the data on the sending side.
This is because when it comes to low-frequency signals that slowly change over a long period, the accuracy is almost the same as if the original data were sent as is.

Regarding this more detailed l@3 embodiment, Fig. 6(a) to (p)
). In this case, the data is 2'S:1
7 two's complements
~2's complement), and Preliminary 11i
The data length of the conversion data (original data) is 15 bits, the data length of the main data is 8 bits, and the MSB of the main data is the sign bit (therefore, the main data is the pre-11 internal conversion data excluding the sign bit. 7 bits are taken out and a sign bit is added to it.)

In advance, as shown in FIG. 6(a), a 15-bit register corresponding to the data length of the preliminary conversion data is prepared (assuming that the register is reset in advance and the register is set to 0). The first original data (pre-excitation conversion data) as shown in FIG. 6(b) obtained by the first □ sampling (νD transformation) is input (added) to this register. As shown in FIG. 4(C), an 8-bit king data r is created and transmitted from the original data according to the scale information and stored in the register. That is, based on the number of effective bits excluding the sign bit of the original data in FIG.

According to the scale information obtained, 7-bit data is taken out and a sign bit is added to the most significant part of this data to form 8-bit main data. Koiyo 1 voice, 2/ as binary code
Since we are using S combination t Itfi, we consider that the sign bits are continuous until the first valid bit is left (i.e., if the sign bit is positive, the sign bit is al+, the most significant]'', and the lower significant bits become the valid bits. If it is negative, the sign bit is 1111, and the lower bits from the most significant II□1 are valid bits), and the bit 1 bit above the most significant bit determined by the scale information
From this, 8 bits of data are extracted from the lower order and used as king data. When the data is retrieved from this register and transmitted, only the truncated portion, that is, the truncated data, remains inside the register as shown in FIG. 2(d). From this truncated data, the receiving side adds the scheduled data "01111", which corresponds to the additional data shown in FIG.
- to find the error with the reproduced data on the receiving side, and put the result in the register as shown in the same figure.
Find the error between. Next, the second original data from the second sampling as shown in (g) of the same figure is added to the register, and the result is shown as (h) of the same figure. put it in the register. In this case, as a result of the above-mentioned addition, a carry occurs from the error part of the lower five digits corresponding to the truncated part, and the value obtained by adding 1 to the lowest part of the part corresponding to the main data of the second original data becomes the main data. The greatest feature of this embodiment is that the errors are sequentially accumulated and the cumulative result is reflected in the main data and transmitted.
As a result? a. Transmission with low distortion (that is, high waveform reproducibility) can be performed.

In other words, the average value for 2 samples (2 data) is within the same error range as 9-bit transmission with 9 bits of king data. Well, 16 samples is equivalent to 12 bits...is that a large number of samples? If you think about it, it will be extremely accurate.

Then, as shown in the same figure, for the second sampling [t
, the same figure (
The planned data "01111"' shown in j) is subtracted, and the resultant data in the same figure (h) (which is negative data) is subtracted from the 30th Zangling's 3rd data in the same figure (R).
The original data are added and the result is stored in the register as shown in (m) of the same figure. Here, the scale information changes and the transmitter Chitake defeat 9 output bit position changes to the upper position (fl+
The bit is shifted by 1 bit and the truncated data becomes 6 bits.

The 6-bit scheduled data "01llll" k as shown in figure (0) is subtracted from the remaining data as shown in figure (n) as a result of all the main data being lost, and the result is shown in figure (n).
p) σ), and the same processing is performed one after another. As shown in the figure ((b)), if the scale information changes, the position from which the main data is extracted is changed accordingly.

In this way, in the third embodiment, not even a single bit of the original data is discarded. The data once discarded at the time of 5 signals is also stored in the register and sequentially added to the next sample σ) original data (tll), thereby reducing the cumulative error between transmission and reception.

Therefore, the average error for a large number of data becomes infinitely smaller as the number of data increases.

In addition, in the above, the additional data added on the receiving side 1, υ
Scheduled data [Figure 6 (e), (j), (0)
), we use the average I + i data of the corresponding number of bits, ``0111...'', and it is of course preferable to use the average I data or something similar to it, but As mentioned above, for a meeting in which errors are accumulated and reflected in the main data, the value of the scheduled data itself may be any preset value, rather than the value itself. For example, the above-mentioned scheduled data 'kO data may be changed to "ooo...". In this case, processing calculations such as temperature reduction on the transmitting side can be simplified, and processing of receiving and biasing and 11 can also be done easily. This applies not only to the third embodiment but also to the second embodiment.

Also, the third fruit 7ill! The order of the data addition process, the data transmission process, and the calculation process in i-1 can be changed as appropriate, and the point is that the lower bits (truncated data) of the preliminary variation data are not completely discarded. It is only necessary to compare it with the scheduled data to be added in the received bone 1, accumulate the error amount, and when the cumulative value reaches 1 scheduled data, add or subtract it to the main data and send it.

In addition, when the average value is sent with a small error as in the second and third embodiments above, it is better to insert a de-emphasis circuit on the 9 side to average the transmitted data. can be recycled. In this way, when inserting a tie-emphasis circuit on the receiving side, good results can be obtained by pre-emphasizing syllables on the transmitting side.

Furthermore, the digital code used for the data may be any code such as natural binary code, 2/S gongliment, offset binary code, etc., and is extracted from the preliminary conversion data when creating main data according to the characteristics of the code. Just change all bit positions.

In addition, the configurations of the data storage register, retouching device, etc. in the second and third examples are merely examples.
A large number of IIFIm processes may be performed in parallel, processing systems may be divided according to bit positions, processing may be performed in a time-division manner, or software processing may be used for some or all of the processing.

Furthermore, instead of performing all digital processing until the reproduction data is obtained, the processing of the received fll+ is carried out in an analog manner according to the gain scale information when converting the main data (i+)/A, -D/A
It is also possible to perform analog level variable control according to the converted t-scale information.

In addition, the present invention does not change the gist of the invention.1゜〔Effects of the invention To provide a signal transmission method that effectively reduces all the experimental noise associated with PCM transmission and reproduces the wave and energy of the original signal with high fidelity when transmitting the signal back to the original signal. be able to.

[Brief explanation of drawings]

Figure 1 is a block diagram showing an example of the conventional method, and Figure 2
The figure is a block diagram showing the configuration of an example of the lJ arrangement method of the present invention method, 31g+(+su~(C) is a diagram for explaining signal processing in N example, and FIG. 4(-)~( e) is a diagram for explaining the first embodiment of the present invention; FIG. 5 is a diagram for explaining the second embodiment of the present invention; FIG. 6 (-)
-(p) are diagrams for fully explaining the third embodiment of the present invention. Applicant's agent Patent attorney Takehiko Suzue 5th garden 6th ward (e) Vo〒〒〒-1-``IL-1--J (j) +01111+L
-J (p) Ward α-sense Ward I [Door ■ Door procedure amendment, 5110 ``'# 11M'' 8 Commissioner of the Patent Office Kazuo Wakasugi 1, Indication of case Patent application No. 150208/1982 2, Name of invention signal transmission Method 3, Relationship with the case of the person making the amendment Patent applicant (240) Sansui Denki Co., Ltd. 4, Agent address: No. 17 Mori Pill, No. 5, Toranomon 1-chome, Minato-ku, Tokyo 105 Phone: 03 (502) 3181 ( Oshiro Iri (i, Supplementary [13 Nori・] Illustrated Specification 1 Drawing 7, Contents of amendments IIB Book (no changes in 8) Drawings 1 (no changes in contents)

Claims (3)

[Claims] (1) Analog signals are converted to digital data at scheduled time intervals to obtain preliminary conversion data, and scale information for determining extraction bit positions when extracting a part of this preliminary conversion data to create main data is used as described above. At a time interval that is sufficiently longer than the scheduled time interval, the settings are updated to approximately correspond to the effective number of bits of the preliminary conversion data near the relevant time, and the planned number of bits is calculated from the bit position corresponding to the scale information of the preliminary conversion data. The data are sequentially extracted to obtain main data with a smaller number of bits than the conversion data of the nucleon 41tt+, in which the upper effective bits of the preliminary conversion data are generally prioritized, and the scale information with the number of bits smaller than the above data. scale information data consisting of digital data, and the scale TFT information data for each setting update is inserted between a large number of king data and given to the transmission system,
Upon receiving this transmission key signal from the transmission system, the king data and the scale information data are separated, the main data is bit-shifted based on the scale information data, and the existence of valid bits is generated as a result of this bit shift. In the lower bit part that does not correspond to the number of bits in Table Q, the average of those values among all the immortals that are n A signal transmission method characterized by obtaining reproduced data and converting the reproduced data into an atelolog signal to obtain a reproduced analog signal. (2) Analog signal? Preliminary conversion data is obtained by converting it to digital data at scheduled time intervals, and extraction bit position t is used when extracting a part of this preliminary conversion data and creating main data.
T? 1'' scale information sound to be determined, at a time interval sufficiently longer than the scheduled time interval, tl is set and updated to match the effective number of bits of the preliminary conversion data near the relevant time, and the above of the preliminary conversion data is updated. The data of the planned number of bits is sequentially extracted from the bit position corresponding to the scale information, and the planned data of the same number of bits is subtracted from the remaining lower residual data taken out, and the obtained data is sequentially cumulatively added and added to the cumulative fji. If the absolute value exceeds the planned number of 41, a predetermined value is added or subtracted to the data of the scheduled number of bits taken out, and the cumulative t4 is added or subtracted each time the scale information changes.
K is reset to obtain main data with a smaller number of bits than the bulk preliminary converted data, in which the upper effective bits of the preliminary converted data are given priority, and the scale information is obtained from the digital data with a smaller number of bits than the main data. The scale information data for each setting update is inserted between a large number of main data and given to the transmission system, and the transmission system transmits the raw data and the scale information data through all transmission signals. The main data is bit-shifted based on the upper scale information data, and the planned data is applied to the resulting lower bit part where no valid bits exist, thereby converting all the reproduced data with the same number of bits as the preliminary conversion data. TI-
Characteristic signal transmission method. (3) Convert analog signals to digital data at scheduled time intervals to obtain preliminary conversion data, and use the scale information described above to determine the extraction bit position when extracting a part of this preliminary conversion data to create king data. At a time interval that is sufficiently longer than the scheduled time interval, the setting is updated so that the number of effective bits of the preliminary conversion data near the relevant time is approximately equal to 1;
Substantially, the data of the scheduled number of bits is sequentially extracted from the bit position corresponding to the scale information of the preliminary conversion data, and the scheduled data of the same number of bits is subtracted from the lower residual data, and the obtained data is used as a subsequent backup. In addition to the conversion data, the upper effective bits of the preliminary conversion data are generally given priority, and the king data having a smaller number of bits than the preliminary conversion data is obtained, and the scaled H information is also The scale information data consists of digital data with a small number of bits, and the scale information data is updated every time the above settings are updated? Is this transmission defect signal from the transmission system by pushing it between a large number of data and giving it to the transmission system? Then, the main data and the scale information data are separated, the main data is bit-shifted based on the scale information data, the resulting effective bit σ) is applied to the non-existent lower bit part, and the scheduled data is applied to the lower bit part that does not exist. A signal transmission method characterized in that reproduced data having the same number of bits as preliminary conversion data is obtained, and the reproduced data is converted into an analog signal to obtain a reproduced analog signal.