JPH0255868B2 - - Google Patents

Description

[Detailed description of the invention]

When encoding an audio signal and transmitting or recording/playing it as a digital signal, conventional methods to reduce the amount of data as much as possible include logarithmically compressing the signal amplitude, taking the difference, or performing delta modulation. It is well known that various methods such as the following have been adopted, but these conventional methods propose a reduction in data in the amplitude direction. By the way, while the above-mentioned digital encoding device can easily encode a signal, since the sampling interval is random, distortion components caused by errors from the original signal are generated when the signal is reproduced. A problem has arisen in that when it is included in the reproduced frequency band, it cannot be removed and is included in the reproduced signal. In other words, if a signal is encoded with a sampling interval such that the zero point interval is divided into approximately equal parts for each zero point interval of the signal, the sampling period will naturally be shorter in the signal part with the longer zero point interval. However, according to the sampling theorem, the frequency band in which the reproduced signal corresponding to the signal part with a long sampling period (low sampling frequency part) is distortion-free is a frequency band of 1/2 or less of the sampling frequency. However, the passband of the low-pass filter installed in the reproduction system is based on the sampling period determined corresponding to the part with the shortest zero point interval in the signal to be encoded. Because of this setting, distortion components in the reproduced signal corresponding to the signal portion with a long sampling period may be included in the passband of the low-pass filter installed in the reproduction system. The problem is that signal distortion inevitably increases. Therefore, in order to solve the above problem, the applicant company first converts the signal to be encoded into a signal of T for a certain time length (signal of one frame).
And the number of zero points for each signal of one frame is
By determining Zc and determining the average zero-point interval in the signal period of one frame, high-compression encoding can be performed while harmonic components in the reproduced signal for each frame are filtered through the low-pass filter during playback. proposed a recording/reproducing device that could remove the wafer by using a method (see specification of Japanese Patent Application No. 57-43649). The present invention provides a signal encoding, storage and reproducing device that can realize encoding with a higher degree of compression than the previously proposed devices. The specific contents of the signal encoding, storage and reproducing device of the invention will be explained in detail. FIG. 1 is a block diagram of the configuration of an example of a signal encoding storage/reproduction device according to the present invention. In FIG. 1, MIC is a microphone, LPF is a low-pass filter, ADC is an AD converter, and CG is a A clock pulse generator, CCT is a control circuit including a microcomputer, OP is an operation section, Br is a record button, Bp is a play button, and Bs is a stop button. In addition, in Figure 1, M ₁ is the first storage device (first memory), M ₂ is the second storage device (second memory), DAC ₁ and DAC ₂ are DA converters, and VLPF is a variable pass. Band-type low-pass filter, AMP is an amplifier,
SP is a speaker. FIG. 2 is a waveform example diagram of the signal Sa before encoding, and the 0-0 line in FIG. 2 is an AC axis line shown for reference. In the waveform diagram shown in FIG. 2, T is the time length T of a signal portion obtained by dividing the signal Sa into each fixed time length on the time axis, and each signal portion of each time length T described above is divided into two parts. This is called a frame signal. The digital data to be decoded in the signal encoding/storage device of the present invention has a sampling period for each one-frame signal depending on the state of change in the signal on the time axis in each one-frame signal. The encoding is performed as determined, but the encoding method is to have a specific relationship with the number of zero points in the signal of one frame, and to encode the samples for each frame. The period of conversion is determined. In the signal Sa shown in the second diagram, each one frame signal having a predetermined time length T is usually
A state where the AC axis crosses the 0-0 line multiple times within the time length T, that is, the time length T
However, the number of zero points in each frame of the signal differs depending on the frequency components of the signal in each frame. In addition, depending on the case, there may be a period in which there is no crossing over the AC axis 0-0. For example, to explain the waveform Sa shown in FIG. 2, there are eight zero points in one frame signal from time t ₁ to t ₂ , and one zero point from time t ₂ to t ₃ .
There are four zero points in the frame signal, and time t ₃
There is no zero point in the one-frame signal from time _t4 to time _t4 , there are three zero points in the one-frame signal from time t4 to time _t5 , and there are three zero points in the one-frame signal from time _t5 to time _t6 . It can be seen that the number of zero points is different for the signals of successive frames on the time axis, such that there is no zero point that occurs when the signal intersects the AC axis. Now, in the existing proposal, for a signal portion of a predetermined constant time length T in the signal Sa, that is, for each one frame signal, the time is expressed as a number related to the number of zero points existing in the one frame signal. In order to reduce the amount of data by performing encoding such that a sequence of sample values is obtained for the signal of one frame using a sampling period Tc whose length is divided into equal parts, it is necessary to reduce the amount of data one after another on the time axis. When the number of zero points of the signal of a frame is Zc, for each signal of one frame, the sample value sequence from the signal of each frame is obtained at a sampling period such that the time length T is equally divided by (Zc × K). Recording and playback is performed with the amount of data reduced so that the amount of data can be obtained.
For example, from time t ₃ in Fig. 2, the signal of one frame is
In the case of a DC signal, such as up to t ₄ , or in the case of a no-signal period, such as from time t ₅ to t ₆ in FIG. 2, unnecessary encoding will be performed. Therefore, in the present invention, when the signals of one frame all have the same sample value, data is composed of one sample value and a code representing the repetition of the sample value, and when playing back, the sample value For a single-frame signal in which there is a code representing the repetition of In the case of a signal with a large number of silent periods, encoding with a high degree of compression can be performed particularly effectively.In the present invention, when the number of zero points in one frame of the signal is Zc, the signal of each frame is The sampling period Tc such that the time length T is divided equally into {(Zc+1)K}
to obtain a sequence of sample values from each frame of the signal, and also determine whether the sample values for each sampling period Tc within the signal of one frame are all the same, and period Tc
If the sample values for each sample are all the same, data is obtained by combining one sample value and a code representing the repetition of that sample value, and the sampling period
Tc, the number of samples N in one frame of the signal, and data that is a combination of sampled values for each sampling period Tc in one frame of the signal, or one sampled value and a code representing the repetition of that sampled value. , and a frame number assigned to distinguish each frame, are created for each successive frame, and stored in a storage device sequentially, and when playing back, successive frames are read out from the storage device. data for each
The signal is reproduced through DA conversion. Next, the aforementioned encoding operation will be explained with reference to the block diagram of FIG. The microphone MIC shown in FIG. 1 converts sound waves into electrical signals (sound signals) and supplies them to a low-pass filter LPF. In the recording and reproducing apparatus shown in FIG. 1, a microphone MIC is used as a signal source, but the signal source may be a generator of other types of audio signals or a generator of other signals. In the following example description, the low-pass filter LPF is as follows:
Its cutoff frequency is said to be 3KHz.
The audio signal converted into a frequency band signal of 3KHz or less by the low-pass filter LPF is converted into a digital signal with the required number of bits (8 bits in the following explanation) by the AD converter ADC, and then sent to the microcomputer. The above-mentioned AD converter ADC performs AD conversion using pulses with a repetition frequency of 8 KHz from a clock pulse generator CG. The digital signal output from the AD converter ADC is an 8-bit digital signal in which the input audio signal is always sampled at a constant sampling period (1/8000 seconds in the example) and quantized. , which are sequentially stored in the first storage device M ₁ (first memory M ₁ or buffer memory M ₁ ) under the control of the control circuit CCT. The buffer memory _M1 described above is assumed to have a storage capacity of 512 bytes in the following explanation example, and it is divided into two parts each having half the storage capacity, and the two parts are used to alternately store data. used for writing and reading data. Now, the encoding/recording operation of the apparatus shown in FIG. 1 is performed according to the program shown in the flowchart shown in FIG. 3 by operating the recording button Br on the operation section OP. Operation section
When the record button Br in the OP is operated and the program starts ("Start" in Figure 3), in step (1) the 9-bit sample counter and the 8-bit zero point counter provided in the control circuit CCT are , the 16-bit frame counter, etc. are reset. Before the record button Br is operated, that is, the third
Even before the "beginning" in the illustrated flowchart, the control circuit of the recording and reproducing apparatus illustrated in the first figure
The CCT performs the interrupt operation in step (10) by receiving pulses from the clock pulse generator CG, and sequentially stores the digital signal output from the AD converter ADC in the buffer memory _M1 .
It also counts up a 9-bit sample counter. Also, before the beginning, the 9-bit sample counter is repeatedly reset each time it reaches a full count. In step (2), the stored sample value is read from the buffer memory _M1 , and the 9-bit sample counter is counted up by 1. In step (3), it is checked whether the sign of the sample value read in step (2) is the same as the sign of the sample value immediately before it, and if there is no change in sign, it is assumed that the point is not zero and the process is continued. Return to step 2), and when there is a change in sign, proceed to step (4) assuming that the sample value read in step (2) is the zero point.
In (4), the 8-bit zero point counter is counted up by 1. In step (5), check whether the number of sample values sequentially read from buffer memory _M1 has reached 256 (or 512) using the counted value of the 9-bit sample counter, and check whether the number of sample values read from buffer memory _M1 has reached 256 (or 512). When the number of items reaches 256 (that is, step (2) ~
If (4) is repeated 256 times), proceed to step (6), and if the number of sample values read from buffer memory _M1 has not reached 256, return to step (2). Here, the number 256 of sample values read out from the buffer memory _M1 as described above is, for one frame signal of time length T of the signal Sa shown in FIG.
AD converter ADC has a constant sampling period (1/8000 seconds)
The number of sample values obtained by sampling at 256 is the time length T of one frame signal.
This corresponds to In step (6), using the count value Zc of the 8-bit zero point counter, the predetermined number K, and the number 256 representing the time length T of the signal of one frame, the sample in the signal of one frame is calculated. Calculate the sampling period Tc required to obtain the value sequence, and
Calculate the number of samples N. Sampling period = Tc = 256/(Zc+1)K Number of samples = 256/Tc Next, in step (7), the buffer memory M ₁
, a 9-bit sample counter (address N sample values are read out sequentially from the buffer memory _M1 using the count value of the counter) at intervals of Tc as an address signal. The sample value is appended with a code indicating that it is repeated, the frame number indicated by the count value Fc of the 16-bit frame counter, the number of samples N, and the sampling period.
Create a pair of data with Tc, store it in the second storage device M ₂ (second memory M ₂ ), and proceed to step (8). If not, create data that combines the N sample values, the frame number indicated by the count value Fc of the 16-bit frame counter, the number of samples N, and the sampling period Tc, and store it in the second storage device. It is stored in M ₂ (second memory M ₂ ) and proceeds to step (8). In step (8), the 16-bit frame counter is incremented by one. In step (9), check whether the 16-bit frame counter has reached full count or the stop button has been pressed.
Check whether Bs is being operated and if the frame counter has reached a full count or if the stop button Bs is being operated, the process is over; if not, return to step (2) and repeat each of the steps above. Repeat. As is clear from the above explanation, the second memory _M2 stores data with a sampling period Tc, data with a sample number N, and a sample value string or one sample value for each frame of the signal. A digital signal consisting of a code representing the repetition, a frame number Fc (frame counter count value Fc), etc. is stored, but this is stored in the first memory M ₁ (buffer memory M ₁ ). The amount of data has been greatly reduced compared to the original digital signal, and due to the above-mentioned encoding performed during recording, the amount of data has been reduced and can be stored in a small memory capacity. This makes it possible to record and play back audio signals for a long time. Next, the case where the signal stored in the second memory _M2 as described above is read out and the audio signal is reproduced will be described with reference to the flowchart shown in FIG. Playback button Bp of the operating section OP in the device shown in the first diagram
is operated, the program shown in Figure 4 starts ("Start" in Figure 4), and the first step (1P) is started.
At step (2P), the frame counter and sample counter are reset, and at step (2P), the number of samples N in the signal of one frame, the data of the sampling period Tc, the data representing the repetition of the sample value, etc. are read out from the second memory _M2 . Next, in step (3P), the second memory M ₂
The sample values are read out one by one in the order in which they were stored, or one sample value is read out repeatedly the required number of times and given to the DA converter DAC ₁ , and
Data with a sampling period Tc is given to the DA converter DAC ₂ . In step (4P), step (2P) ~
Waiting is performed so that the time required for one round (5P) matches the time length indicated by the sampling period Tc. 1/8000 = (Tc-1) (program loop time) +
{Fixed portion of step (2P), step (3P), step (5P), and step (4P)} At step (5P), determine whether reading of N sample values has been completed; If the process has not yet been completed, return to step (2P); if it has been completed, proceed to step (6P). In step (6P), check whether the frame counter has reached a full count or whether the stop button Br has been operated. If the frame counter has reached a full count or the stop button Bs has been operated, the process is over, or not. Return to step (2P). DA converter in the above steps (4P)
By supplying digital data to DAC ₁ and DAC ₂ , analog signals corresponding to successive sample values within one frame signal are input from the DA converter DAC ₁ to the variable passband type low-pass filter VLPF. It is supplied as a signal and also 1 from the DA converter DAC ₂ .
Sampling period Tc of sampled values in the frame signal
The corresponding analog signal is given to the variable passband type low-pass filter VLPF as its control signal. FIG. 5 is a block circuit diagram showing a configuration example of the DA converters DAC ₁ and DAC ₂ and a variable pass band type low pass filter VLPF. , the higher cut-off frequency in the variable range of the cut-off frequency is determined by the resistors R ₅ , R ₆ and the capacitors C ₁ , C ₂ , and the lower cut-off frequency in the variable range of the cut-off frequency is determined by the resistors R 5 , R 6 and the capacitors C 1 , C 2 . It is determined by R ₆ and R ₇ and capacitors C ₁ and C ₂ , and the intermediate frequency value of the cut-off frequency variable range described above is determined by the sampling period Tc data input to the DA converter DAC ₂ . The analog switch ASW, which is switched by the input, and the variable resistors R ₂ and R ₃ , which are connected in series, are varied to adjust each to a predetermined frequency value corresponding to the data of the sampling period Tc. be done. An example of the correspondence between the sampling period Tc, the sampling frequency s corresponding to the sampling period Tc, and the cutoff frequency c of the low-pass filter is shown in Table 1 below.

[Table] The variable passband type low-pass filter VLPF shown in Fig. 5 is selected by the analog switch ASW depending on the data value of the sampling period Tc given to the DA converter DAC ₂ . The resistor changes the gain of the amplifier A ₁ and accordingly the photocoupler PC ₁ ,
Photoresistor (e.g. CdS) due to changes in the amount of light emitted from photodiode Pd and Pd in PC ₂
The cutoff frequency changes as the resistance values of VR and VR change, and the passband is variably controlled. In addition, in FIG. 5, R ₁ , R ₄ , R ₅ to R ₆ are resistors, R ₂ and R ₃ are variable resistors, A ₁ and A ₂ are amplifiers,
C ₁ and C ₂ are capacitors, and PC ₁ and PC ₂ are photocouplers. As is clear from the above description, the signal encoding, storage and reproducing apparatus of the present invention includes a means for sampling and quantizing a signal at a predetermined sampling period, and a first means for storing a signal section of a predetermined time length T as a signal of one frame, and means for measuring the number of zero points Zc in the signal of one frame;
The quotient obtained by dividing the time length T of the signal of one frame by the value of {(count value Zc of the number of zero points) + 1} is further divided by a predetermined number to obtain the sampling period Tc. means for obtaining the number of samples N in one frame signal by dividing the time length T of the one frame signal by the sampling interval Tc; means for obtaining sample values; means for determining whether all sample values for each sampling period Tc within one frame signal are the same; and sampling period within one frame signal.
If the sample values for each Tc are all the same, means for obtaining data that combines one sample value and a code representing the repetition of the sample value, a sampling period Tc, and one frame. The number of samples in the signal N,
A sample value for each sampling period Tc in a signal of one frame or data that combines one sample value and a code representing the repetition of the sample value, and a frame number assigned to distinguish each frame. A means for creating a set of data for each successive frame and sequentially storing it in a second storage means, and performing DA conversion on the data for each successive frame read from the second storage means to reproduce a signal. In the signal encoding storage and reproducing apparatus of the present invention, when the signals of one frame all indicate the same sample value, the signal is encoded by one sample value and a code representing its repetition. To configure the value column data,
The amount of data can be highly compressed,
In particular, for audio signals with many silent periods, not only can the amount of data be significantly reduced, but also recording and reproduction can be performed satisfactorily using a recording device with a simple configuration. The conventional problems are well solved.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of a recording/reproducing apparatus including an encoding storage device of the present invention;
FIG. 2 is an explanatory waveform example diagram, FIGS. 3 and 4 are explanatory flowcharts, and FIG. 5 is a block circuit diagram of a variable passband type low-pass filter and its related parts. MIC...Microphone, LPF...Low pass filter,
ADC...DA converter, CG...clock pulse generator, CCT...control circuit including a microcomputer, OP...operation unit, Br...record button, Bp...play button, Bs...stop Button, _M1 ...first memory device (buffer memory), _M2 ...second memory device, _DAC1 , _DAC2 ...DA...converter,
VLPF...Variable pass band type low pass filter.

Claims (1)

[Claims] 1. A means for sampling and quantizing a signal at a predetermined sampling period, a means for storing the digital signal obtained by the above means in a first storage means, and a predetermined period of time. A signal interval of length Tf is taken as a signal of one frame, and means for measuring the number Zc of zero points in the signal of one frame, and time length Tf of the signal of one frame described above are {(measured value of the number of zero points) Zc)
+1} and further divides the obtained quotient by a predetermined number to obtain a sampling period Tc;
Divide the time length Tf of the signal of one frame by the sampling interval Tc to find the number of samples N in the signal of one frame.
means for obtaining sample values for each sampling period Tc within a signal of one frame; means for determining whether all sample values for each sampling period Tc within a signal of one frame are the same; Since the sampled values for each sampling period Tc in one frame signal were all the same, there was a problem in that the quality of the reproduced signal deteriorated due to quantization distortion. When the amount of data decreases significantly, the applicant's company does not determine the decrease in bits in the amplitude direction, but rather in the time axis direction, and compares the waveform obtained from the sample value sequence with the original. We proposed an approximation compression method for audio signals that can be expected to significantly reduce the amount of data by approximating the waveform so that the degree of difference from the signal waveform is less than a certain ratio (Japanese Patent Application Laid-Open No. 1983-1999).
(Refer to Publication No. 155998), and achieved some success. However, in this existing proposed method, the sampling period is determined based on the deviation of the signal waveform from the previous sample value, and the sampling period is determined based on the signal waveform itself. However, it was observed that the reproduction of fine waveforms in the signal was insufficient because it was not done in a way that was well defined. In order to solve the problems with the above-mentioned existing proposal by the applicant company, the applicant company has developed a digital means for causing the encoding device to obtain data combining one sample value and a code representing repetition of the sample value; a sampling period Tc; and a number N of samples in one frame of the signal; A sample value for each sampling period Tc in a signal of one frame or data that combines one sample value and a code representing the repetition of the sample value, and a frame number assigned to distinguish each frame. A means for creating a set of data for each successive frame and sequentially storing it in a second storage means, and performing DA conversion on the data for each successive frame read from the second storage means to reproduce a signal. A signal encoding storage/reproduction device comprising means.