JP2004198559A - Encoding method and decoding method for time-series signal - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an encoding method and a decoding method for a reversibly compressed time-series signal which sufficiently compress the time-series signal wherein patterns of similar amplitudes repeatedly appear. <P>SOLUTION: When the time-series signal consisting of sample sequences of time series is constituted of a plurality of channels, differences between channels are calculated, and operation results of parts having small differences are separated as inter-channel difference data (S1). Then a plurality of frames having a prescribed length are set in the sample sequence of each channel, and differences between the frames are calculated, and frames having small signal differences are separated as inter-frame difference data (S2). Parts having small signal variance in each sample sequence are separated as signal flat part data (S3). Higher-order bits and lower-order bits of each sample sequence are separated (S4) after these data are separated, and the higher-order bit data are converted to a variable bit length data (S6) after a predictive error is calculated for higher-order bits (S5). <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
[Industrial applications]
The present invention relates to the field of music production, such as music production, storage of audio data materials, and relay of location materials, particularly to the field of producing high-definition audio with higher quality than CDs, and audio recording using digital recording media such as CDs and DVDs. The present invention relates to a reversible data compression technique suitable in a field where reproduction of data, a transmission of a biological signal in telemedicine, and the like where data modification is disliked, and the like are performed.
[0002]
[Prior art]
Conventionally, various methods have been used for compressing an acoustic signal. MP3 (MPEG-1 / Layer3), AAC (MPEG-2 / Layer3) and the like have been put to practical use as a technique for compressing and encoding an audio signal. With such a compression encoding method, it is possible to treat an audio signal as small data, which contributes to the efficiency of data recording and transmission.
[0003]
[Problems to be solved by the invention]
MP3, AAC, and the like as described above are all called lossy coding schemes, and can be efficiently compressed. However, decoding involves a considerable deterioration in quality and completely reproduces the original signal. It is not possible. For this reason, in the music production field such as music production, material storage, and location material relay, these encoding methods cannot be applied, and although inefficient, non-compressed storage / transmission methods are used. In particular, recently, the number of productions that handle high-definition audio has increased, the material capacity has become enormous, and this has become a problem in managing work disks.
[0004]
In order to solve the above-described problems, the present applicant performs a difference operation between channels and between frames on a sample sequence of a time-series signal, reduces the value of each sample, and uses predictive coding. We propose a method for compressing data. (See Patent Document 1).
[0005]
[Patent Document 1]
Japanese Patent Application No. 2002-231150
[0006]
However, in the method proposed in the above-mentioned application, when the time series signal is composed of a plurality of channels like a stereo sound signal, if the difference processing is performed over the entire section of the time series, the signal amplitude may increase. May occur, and the subsequent encoding using the prediction error may rather increase the data amount. Further, when there is a correlation between frames which are a unit of samples at different times of the same channel, the signal amplitude is reduced by performing a difference operation between the frames, but the ratio of unpredictable noise components is increased. However, the subsequent encoding using the prediction error may rather increase the data amount. This is particularly likely to occur when a signal having a similar amplitude pattern (so-called signal waveform) is repeatedly generated. In addition, a portion having the same signal level is separated as a signal flat portion so as to reduce the data amount. However, since there are not so many acoustic materials in which portions having the same signal level continue, the compression effect is small.
[0007]
Therefore, in order to solve these problems, the present invention can perform sufficient compression on a time-series signal in which a pattern having a similar amplitude is repeatedly generated. An object of the present invention is to provide an encoding method and a decoding method of a time-series signal of a reversible compression scheme that can be completely decoded.
[0008]
[Means for Solving the Problems]
In order to solve the above problem, the present invention provides a coding method for compressing the amount of information so that all the sample sequences can be reproduced with respect to a time-series signal composed of a time-series sample sequence. , A plurality of frames composed of a predetermined number of sample sequences are extracted, and a correlation operation is performed between the extracted frames.When the correlation between the frames is high, the sample sequence of one frame is extracted. As an inter-frame difference data expressed by a small number of bits, an inter-frame calculation step of separating from the sample sequence, the separated inter-frame difference data, and a sample sequence remaining after the separation of the inter-frame difference data are performed in a predetermined format. The recording step is executed.
[0009]
According to the present invention, for a time-series sample sequence, a portion having a high correlation in the sample sequence is separated as difference data, and the separated portion is represented by a small number of bits, so that a pattern having a similar amplitude is obtained. Sufficient compression can be performed on time series signals that occur repeatedly.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
(Data structure)
First, a time-series signal to be encoded in the time-series signal encoding method according to the present invention will be described. In the present embodiment, a case where an acoustic signal is applied as a time-series signal will be described as an example. FIG. 1A is a diagram schematically illustrating an acoustic signal handled in the present invention. In FIG. 1, the left-right direction is a time-series direction, and the time advances toward the right side. That is, the left end is the start time, and the right end is the end time. The acoustic signal shown in FIG. 1 is data of two channels, Ch1 and Ch2, in which an L (left) signal is recorded in Ch1 and an R (right) signal is recorded in Ch2.
[0011]
In order to obtain a digital audio signal as shown in FIG. 1A, first, an analog audio signal which is a time-series signal is digitized. This can be done by using a conventional general PCM technique, sampling this analog audio signal at a predetermined sampling frequency, and converting the amplitude into digital data using a predetermined number of quantization bits. In the present embodiment, a description will be given below on the assumption that positive and negative signs are recorded at a sampling frequency of 44.1 KHz and a quantization bit number of 16 bits. Sampling at a sampling frequency of 44.1 KHz results in a sample sequence composed of 44100 samples per second. Here, since the audio signal is composed of a plurality of channels, digitization is performed for each channel.
[0012]
(Encoding method)
Next, an outline of a time-series signal encoding method according to the present invention will be described. FIG. 2 is a flowchart showing an outline of a time-series signal encoding method according to the present invention. First, a difference operation between channels is performed on the sample sequence as the digital audio signal shown in FIG. 1 (step S1). Specifically, first, the difference calculation of the sample data at the same time between the entire section of the channel Ch1 and the entire section of the channel Ch2 is performed. As a result, the difference data of the section on the channel Ch2 where the difference is equal to or less than the predetermined threshold is separated from the sample sequence of the channel Ch2 and separately recorded as inter-channel difference data. In the present embodiment, the predetermined threshold value is set within the lower 4 bits. The expression “within the lower 4 bits” is a value from −8 to +7 as a decimal number when expressed with positive and negative signs. As a result of the difference calculation, a section in which the value of the sample sequence takes a value of -8 to +7 is recorded as inter-channel difference data. When the same value continues for a plurality of samples in the inter-channel difference data, the data amount is further reduced by recording the first sample number of the continuous portion, the sample value, and the number of continuous samples. In this case, the last sample number may be recorded instead of the number of consecutive samples.
[0013]
In the channel Ch2 from which the inter-channel difference data has been separated, the total number of samples is reduced by packing the sample sequence after the separated section forward. For example, when samples whose difference values are equal to or less than the threshold value are continuous in the section shown in FIG. 1B, this section is separated, and the sample sequence of the channel Ch2 and the inter-channel difference data are respectively shown in FIG. ). The number of samples of the channel Ch1 (represented by the length on the left and right in the figure) does not change, but the number of samples of the channel Ch2 decreases by the amount of the separated difference data. In addition, in the example of FIG. 1, only one section of the inter-channel difference data is separated, but in reality, many sections are separated. As a method of compressing data between channels, there is also a method of performing a difference operation over the entire section and recording the difference value as a sample value of one channel. However, in the case of such a method, if the original sound signal is recorded equally to the left and right, such as a vocal sound, the compression efficiency is high, but the sound is recorded around one of the channels like a musical instrument sound. However, the signal amplitude becomes large, and as a result of using the predictive coding performed later, the data amount may be rather increased. Therefore, in the present embodiment, a method is employed in which only sections having a small difference are separated as difference data.
[0014]
Subsequently, a frame having a predetermined section length is set for the sample sequence of each channel on which the difference calculation process between channels has been performed, and the calculation between the set frames is performed (step S2). First, the similarity of the sample sequence constituting each frame is determined, and similar frames are selected. In the present embodiment, the frame length is fixed over the entire section from the start time to the end time of the sample sequence. Specifically, one frame has 256 samples. 256 samples are extracted from the beginning of the channel data (sample sequence constituting the channel) each as one frame, and the similarity of each frame is obtained. Since the similarity between frames means that the correlation between both signals is obtained, various methods for performing the correlation calculation can be used. In the present embodiment, among the 256 samples in each frame, the other frames are used. The maximum value of the absolute value of the difference value from the corresponding sample in is extracted, and a frame pair whose maximum value falls within a predetermined value is selected as one similar frame. This process is performed over the entire section of the sample sequence. FIGS. 3A to 3C show how the sample sequence changes due to the inter-frame arithmetic processing. Although only one channel is shown in FIG. 3 unlike FIG. 1, the other channels are processed in the same manner. First, as shown in FIG. 3A, the sample sequence framed into a fixed length is divided into frames A1, A2, A3,.
[0015]
Subsequently, a difference is calculated for each frame. In this embodiment, 256 difference values are obtained for each sample time. If the maximum absolute value of the obtained difference values is within a predetermined value, the sample sequence after the difference processing of the frame is separated from the sample sequence of each channel and recorded as difference data. For example, as shown in FIG. 3B, when processing is performed on the frames A1 and A2, the preceding frame A1 remains as it is, but the maximum absolute value of the difference value between the frames A1 and A2 is maintained. Since the value is within the predetermined value, as shown in FIG. 3 (c), the frame A2 is separated from the sample sequence of the channel, and another frame is stuffed. Thus, when one frame is separated, 256 samples are reduced from the sample sequence. The separated frame A2 is not recorded with the value as it is, but the difference data between the frame A1 and the frame A2 is recorded with the minimum number of bits capable of expressing the maximum value. Although the information of the frame A2 is deleted, in order to restore the information of the frame A2 at the time of decoding, a difference value (expressed as “A2-A1” in FIG. 3) between the samples of the frame A1 and the frame A2 is determined by an inter-frame difference. Separated as data.
[0016]
On the other hand, if the maximum absolute value of the difference between the frame A1 and the frame A2 does not fall within the predetermined value, the original sample sequence of the frame A2 is left as it is. Similarly, the same processing is performed between the subsequent frames, such as frame A1 and frame A3, frame A2 and frame A3, frame A1 and frame A4, frame A2 and frame A4, and frame A3 and frame A4. Is At this time, the original sample sequence is also referred to in the subsequent inter-frame difference processing for the frame A2 deleted as a frame similar to the frame A1. Further, in order to reduce the load of the difference calculation processing, the distance between the frames to be referred to is limited to 100 frames or less. That is, the frame on which the difference calculation process is performed with the frame A1 is up to the frame A100, and the frames after the frame A101 are excluded from the similar frame determination.
[0017]
The above-mentioned inter-frame difference data also records two frame numbers subjected to the difference processing. If the same value continues for a plurality of samples in the inter-frame difference data separated in step S2, as in the case of the inter-channel difference data, the first sample number of the continuous part, the sample value, and the continuous The amount of data is further reduced by recording the number of samples. In this case, the last sample number may be recorded instead of the number of consecutive samples. As for the sample sequence in which the frame is separated as the difference data, the total number of samples is reduced by packing the sample sequence after the separated frame in front.
[0018]
Next, the signal flat portion is processed (step S3). The signal flat portion is a portion where the same signal level is originally continuous. In particular, it often occurs in a silent part where the signal level is “0” and in a saturated part where the absolute value of the signal level is the maximum. Silence occurs when the sound is actually silent or when the sound is not recorded very small, while saturation occurs during the signal recording and A / D conversion. Regardless of whether a silent portion, a saturated portion, or the other same signal level continues, the same signal level is continuously recorded for a predetermined time (a predetermined number of samples) in the signal flat portion. For this reason, this part is data that can be easily compressed. In the present embodiment, it is assumed that the signal flat portion includes not only a portion where the signal level has the same value but also a portion where the signal level does not change much. That is, in step S3, a portion in which samples whose difference from the previous sample value is equal to or smaller than the predetermined value is continuous is extracted as a signal flat portion and separated from the original sample sequence. As for the portion where the same value continues for a plurality of samples (original signal flat portion) in the separated signal flat portion, the head sample number, sample value, Further, the data amount is further reduced by recording the number of consecutive samples. In this case, the last sample number may be recorded instead of the number of consecutive samples. As for the sample sequence from which the signal flat portion is separated, the number of samples is reduced by packing the sample sequence after the separated signal flat portion forward. For example, when it is determined that a portion shaded in FIG. 4A is a signal flat portion, as shown in FIG. 4B, a sample sequence corresponding to the signal flat portion is extracted from the sample sequence. Will be deleted and padded before. Information on the signal flat portion deleted from the sample sequence is separated as signal flat portion data.
[0019]
As described above, after the difference between channels is calculated, the difference between each frame in the sample sequence of each channel is calculated, and the sample sequence is reduced by separating the signal flat portion, each sample constituting the remaining sample sequence is performed. The upper bits and the lower bits of the data are separated (step S4). For example, when digitizing an audio signal by PCM, when sampling is performed with a quantization bit number of 16, each sample is represented by 16 bits. In this case, in this embodiment, the upper bits are separated into 12 bits and the lower bits into 4 bits. This separation is basically performed to separate the thermal noise of a circuit used for digitizing an acoustic signal, such as an A / D converter. Therefore, lower bits considered to be thermal noise are separated. The degree to which the lower bits are separated depends on the characteristics of the sound source and the circuit used, but it is usually desirable to set the number of quantization bits to about 1/4. Therefore, in this case, 4 bits corresponding to 1/4 of 16 bits are separated as lower bits.
[0020]
Here, the state of data separation of upper bits and lower bits is schematically shown in FIG. In FIG. 5, H indicates upper bits or upper sample data, and L indicates lower bits or lower sample data. FIG. 5A shows sample data before separation. By the processing in step S4, the sample data is separated into the upper sample data shown in FIG. 5B and the lower sample data shown in FIG. 5C. Note that the sign bit included in the upper bits is directly included in the upper sample data and separated. The sample data separated in this way will be separately processed thereafter.
[0021]
(Encoding of upper sample)
For the upper sample data, first, a prediction value and a prediction error of each sample are calculated based on the two immediately preceding samples (step S5). Here, a calculation method of the prediction error will be described with reference to FIG. For example, consider a case where the sample values are in a state as shown in FIG. In FIG. 6A, the horizontal axis is the time (sample number), and the vertical axis is the upper sample value x (t). The line segment at each time indicates the value of the upper sample x (t) at each time. In this state, when calculating the prediction error e (t) in the sample at time t, the upper sample value x (t-1) at the immediately preceding time t-1 and the upper sample value x (t-1) at the immediately preceding time t-2. It is calculated by the following [Equation 1] using the value x (t-2).
[0022]
[Formula 1]
e (t) = x (t) −2 × x (t−1) + x (t−2) −e (t−1) / 2
[0023]
In the above [Equation 1], “2 × x (t−1) −x (t−2)” is a linear prediction component based on two past samples. That is, the prediction error e (t) at the time t is calculated using the calculated linear prediction component and the prediction error “e (t−1) / 2” (error feedback component) calculated in the immediately preceding sample. Will be. The prediction error is calculated for all the samples, and the prediction error is recorded instead of the sample value.
[0024]
This will be described based on the upper sample shown in FIG. First, each prediction error eo (t) is calculated without adding an error feedback component. As shown in FIG. 6B, when calculating the prediction error eo (t) at the time t, the upper sample value x (t-1) at the immediately preceding time t-1 and at the immediately preceding time t-2. The prediction error eo (t) is based on the difference between the value taken by the prediction line connecting the upper sample values x (t−2) at time t and the upper sample values x (t) at time t (indicated by the thick dotted line in the figure). Is calculated. The same operation is performed after time t + 1 to calculate the prediction error eo (t + 1). The calculated prediction error eo (t) is as shown in FIG. As can be seen from a comparison between FIG. 6A and FIG. 6C, the range in which the value fluctuates is greatly narrowed, and data compression becomes more convenient.
[0025]
Subsequently, 50% of the prediction error e (t-1) obtained by adding the correction at the immediately preceding time t-1 to the prediction error eo (t) based on [Equation 1] is subtracted, and the error feedback processing is performed. FIG. 6D shows the added result. The prediction error at times t + 1 and t + 2 is remarkably reduced as compared with FIG. Conversely, the prediction error increases at times t + 3 and t + 4, but the prediction error decreases on average, and the range in which the value fluctuates is further narrowed as compared with FIG. 6A, thereby improving the data compression effect.
[0026]
By the process in step S5, the value of each upper sample is replaced with the prediction error value from the original value, but each bit configuration remains fixed-length 12 bits. Next, the fixed-length upper sample sequence is converted into a variable-length bit configuration (step S6). For this purpose, first, sign-inverted data is inserted. Specifically, sign-inverted data is inserted in a portion where the sample value changes from a positive value to a negative value, and the negative sample value is replaced with its absolute value. An appropriate bit array is assigned as the sign-inverted data. Since the sign-inverted data is converted into a different bit array in the subsequent processing, at this point, any bit array that can be distinguished from other sample strings may be used. However, it is configured to have 12 bits in accordance with the number of bits of the other sample strings.
[0027]
Next, in order to represent the higher-order sample data recorded with the prediction error value with a smaller data amount, conversion of the bit configuration is performed (step S7). First, a look-up table used for converting the bit configuration is created. Specifically, first, a histogram of each sample value is calculated over the entire time. Since all the sample values have been converted into absolute values in the process of step S6, the histogram is calculated without distinguishing between positive and negative. As a result, when the type of the sample absolute value is 640 or more, the separator bit is set to the fixed value of 2 bits “00”. When the type of the sample absolute value is 639 or less, the separator bit is set to the fixed value of 1 bit “0”. ". Furthermore, a bit pattern with a smaller number of bits is assigned in order from the sample absolute value having the highest appearance frequency. At this time, there is a rule for the bit pattern to be allocated. The most significant bit is always set to “1”, and when the separator bit is 2 bits “00”, the bit pattern including the bit pattern of “001” is prohibited, and When one bit is “0”, a bit pattern including a bit pattern of “01” is prohibited. FIG. 7 shows an example of a look-up table when the separator bits are 1 bit “0” and 2 bits “00”.
[0028]
Using the lookup table created as described above, continuous high-order sample data having a fixed length of 12 bits is converted into a variable-length bit pattern. Since the length is variable, it is necessary to distinguish the breaks of each data after conversion. Therefore, in the present embodiment, the above-described 1-bit or 2-bit separator bit is inserted between each data. When the separator bit is one bit “0”, the bit arrangement and the number of bits for expressing the data of each rank are as shown in FIG. 7A. In FIG. 7A, the 0th place is represented by 1 bit “1” having the smallest number of bits. In FIG. 7A, the bit string before conversion is omitted, but the sign-inverted data that appears most frequently is actually represented by “1”. Also, since a separator is always added to each variable-length bit, two bits are required to represent the data of the 0th rank. In the example of FIG. 7, since the separator bit is one bit “0”, the bit pattern “01” is not assigned. However, the bit pattern of “1000” shown as the sixth place is exceptionally changed to the bit pattern of “101” when the immediately preceding bit is “0” (separator bit) during conversion to variable-length bits. be able to. At this time, an array of “0101” appears with the immediately preceding separator bit and bit pattern. This bit array “0101” can be considered as a case where two bit arrays “1” of the 0th rank are consecutive with the separator bit interposed therebetween. However, the bit array “1” of the 0th rank is assigned with sign-inverted data, and it is impossible that two sign-inverted data are consecutive. Therefore, the decoding system uses “101” bit array data. It can be determined that there is. Thus, the bit pattern of the sixth place can be reduced from 5 bits to 4 bits by combining the separator bits.
[0029]
When the separator bits are two bits “00”, the bit arrangement and the number of bits for expressing the data of each rank are as shown in FIG. 7B. In FIG. 7B, the 0th place is represented by 1 bit “1” having the smallest number of bits. As described above, the sign-reversed data that appears most frequently is represented by “1”. In addition, since a separator is always added to each variable-length bit, three bits are required to represent the data of the 0th rank. In the example of FIG. 7B, since the separator bit is one bit “00”, the bit pattern “001” is not allocated. However, the bit pattern of “10000” shown as the 14th place can be exceptionally changed to the bit pattern of “1001” when the immediately preceding bit is “00” during conversion to variable-length bits. At this time, an array of “001001” appears with the immediately preceding separator bit and bit pattern. This bit array “001001” can also be considered as a case where two bit arrays “1” of the 0th rank are consecutive with the separator bit interposed therebetween. However, the bit array “1” of the 0th rank is assigned with sign-inverted data, and it is impossible that two sign-inverted data are consecutive. Therefore, the decoding system uses the data of “1001” bit array. It can be determined that there is. Thereby, the bit pattern of the 14th place can be reduced from 7 bits to 6 bits by combining the separator bits. FIGS. 8A and 8B schematically show how data is converted in step S6. 8 (a) and 8 (b) correspond to the upper part of the sample sequence, and FIG. 8 (a) shows a state in which upper samples of a fixed length are continuously recorded. The upper sample sequence as shown in FIG. 8A is converted as shown in FIG. 8B using the look-up tables shown in FIGS. 7A and 7B.
[0030]
(Encoding of lower sample)
On the other hand, the lower sample data is arranged continuously as it is. Specifically, the lower 4 bits of data separated in step S4 are arranged continuously.
[0031]
(Recording of code data)
The code data obtained as described above is as shown in FIG. That is, there are an upper variable length sample sequence, a lower fixed length sample sequence, a lookup table, signal flat portion data, inter-frame difference data, and inter-channel difference data. This data is recorded in a format suitable for the recording medium to be recorded.
[0032]
(Decryption method)
Next, a method of decoding the encoded data encoded by the encoding method will be described. The decryption is executed by a computer and a dedicated software program mounted on the computer. An outline of the decoding method is shown in a flowchart of FIG.
[0033]
First, a recording medium on which code data as shown in FIG. 9 is recorded is read into an apparatus (computer) for decoding. Subsequently, the fixed-length upper fixed-length sample sequence, that is, the linear prediction error e (t) is restored from the higher-order variable-length sample sequence by referring to the lookup table in the read data (step S11). . As a result, a fixed-length sample sequence as shown in FIG. 8A is restored. Next, 12-bit fixed-length high-order sample data x (t) is sequentially restored based on the equation obtained by exchanging the left-hand term and the right-hand first term of [Equation 1] (step S12). In step S12, each sample sequence remains at a fixed length of 12 bits, but its value changes. Subsequently, the restored upper fixed length sample sequence and lower fixed length sample sequence are integrated (step S13). Specifically, a process of extracting 12 bits from the upper fixed-length sample sequence, extracting 4 bits from the lower fixed-length sample sequence, and sequentially integrating them is performed. As a result, a sample sequence in which each sample has 16 bits is restored.
[0034]
Next, flat portion data is inserted into such a 16-bit fixed-length sample sequence (step S14). The flat part data is inserted into the sample sequence based on the first sample number of the flat part data. Thus, the sample sequence as shown in FIG. 4A is restored.
[0035]
Further, the original frame data is restored using the inter-frame difference data, and inserted into the sample sequence (step S15). Since the inter-frame difference data also has information on the head sample number and the frame on which the difference calculation is to be performed, the original frame is restored using this information. Further, the restored frame is inserted into a predetermined position of the original sample sequence. For example, in the example of FIG. 3, since the inter-frame difference data “A2-A1” has information indicating that it is the difference from the frame A1, the frame A2 is restored using the sample sequence of the frame A1. I do. Subsequently, the data is inserted into the sample sequence using the head sample number held by the inter-frame difference data “A2-A1”. Thus, the sample sequence as shown in FIG. 3B is restored.
[0036]
Subsequently, the channel difference data is inserted into the sample sequence (step S16). The inter-channel difference data includes the first sample number, the last sample number, the channel number of the original sample sequence (Ch2 in the above example), and the channel number of the referenced channel (Ch1 in the above example). Therefore, the sample value of the original channel is restored using the sample value of the reference channel and the sample value of the difference, and then inserted into the sample sequence of the original channel. Thus, the sample sequence as shown in FIGS. 1A and 1B is restored. As a result, the digital audio signal in which the analog signal is converted into the PCM is restored without data loss.
[0037]
(Specific configuration for realization)
Although the encoding method and the decoding method according to the present invention have been described above, the encoding method is actually executed by an arithmetic processing device such as a computer. Specifically, a program for executing the steps shown in the flowchart of FIG. 2 in the above procedure is installed in a computer. Then, after time-series signals such as acoustic signals are digitized by the PCM method or the like, the signals are captured by a computer, and the processes of steps S1 to S6 are performed. Code data is output as digital data from the computer and recorded on a recording medium. I do. The output code data is decoded according to a decoding method. Specifically, a program for executing the steps shown in the flowchart of FIG. 10 in the above procedure is installed in a computer. Then, the code data recorded on the recording medium is taken into a computer, and after performing the processes of steps S11 to S16, a time-series signal such as a digital acoustic signal is restored and output.
[0038]
As described above, the preferred embodiments of the present invention have been described, but the present invention is not limited to the above embodiments, and various modifications are possible. For example, in the above-described embodiment, when performing an inter-frame operation, the frame length is set to a fixed length and sequentially determined from the beginning. However, the frame length is set to be variable based on the characteristics of the time-series signal. May be.
[0039]
【The invention's effect】
As described above, according to the present invention, a time-series signal composed of a time-series sample sequence is used to compress an information amount so that all sample sequences can be reproduced. A plurality of frames composed of the sample sequence of the number are extracted, a correlation operation is performed between the extracted frames, and when the correlation between the frames is high, the sample sequence of one frame is reduced with a smaller number of bits for each sample. Separated from the sample sequence as the expressed inter-frame difference data, and the separated inter-frame difference data, and the sample sequence remaining after the separation of the inter-frame difference data are recorded in a predetermined format. By expressing with a small number of bits, sufficient compression is performed on time-series signals in which patterns with similar amplitudes occur repeatedly. There is an effect that it is possible.
[Brief description of the drawings]
FIG. 1 is a diagram showing how differential data is separated by calculation between channels.
FIG. 2 is a flowchart showing an outline of a time-series signal encoding method according to the present invention.
FIG. 3 is a diagram showing how differential data is separated by calculation between frames.
FIG. 4 is a diagram illustrating a state of flat part data separation.
FIG. 5 is a diagram showing how upper bits and lower bits of sample data are separated.
FIG. 6 is a diagram illustrating a state of a prediction error calculation process in step S5.
FIG. 7 is a diagram showing a lookup table used for bit length conversion.
FIG. 8 is a diagram schematically showing conversion to a bit length of an upper sample.
FIG. 9 is a diagram showing code data obtained by the time-series signal coding apparatus according to the present invention.
FIG. 10 is a flowchart showing an outline of a time-series signal decoding method according to the present invention.

Claims (12)

An encoding method for compressing the amount of information so that all the sample sequences can be reproduced for a time-series signal composed of a time-series sample sequence,
A plurality of frames composed of a predetermined number of sample sequences are extracted from the sample sequence, a correlation operation is performed between the extracted frames, and when the correlation between the frames is high, a sample sequence of one frame is extracted. An inter-frame calculation step of separating the sample from the sample sequence as inter-frame difference data expressed by a smaller number of bits;
Recording the separated inter-frame difference data, and a sample sequence remaining after separation of the inter-frame difference data in a predetermined format;
A coding method for a time-series signal, comprising: In claim 1,
After the inter-frame operation step,
In the sample sequence remaining after the separation of the inter-frame difference data, a section in which the value of a sample is continuously within a predetermined range is defined as signal flat portion data in which each sample is represented by a smaller number of bits. A method for encoding a time-series signal, comprising a signal flat part separation step of separating a signal from a column. In claim 1 or claim 2,
After the inter-frame calculation step or the signal flat part separation step,
A time-series signal of a time-series signal, comprising a sample encoding step of encoding each sample data of the remaining sample sequence using a prediction error from a temporally past sample sequence. Encoding method. In claim 3,
Before the sample encoding step,
Upper / lower separation that divides each bit data constituting the sample sequence at a predetermined bit position and separates into upper sample data composed of a sample sequence of upper bits and lower sample data composed of a sample sequence of lower bits Having a stage,
The encoding method of a time-series signal, wherein the sample encoding step encodes the higher-order sample data using a prediction error from a temporally past sample sequence. In claim 3 or claim 4,
The sample encoding step includes:
A prediction error calculating step of updating each sample value as a new value using the prediction error calculated using the immediately preceding two samples of each sample value;
A bit length conversion step of converting each fixed-length upper-order sample data recorded as a prediction error value into a variable-length sample data;
A time-series signal encoding method characterized by having: In claim 5,
The bit length conversion step includes:
A look-up table creating step of creating a look-up table described with a minimum bit length such that the most significant bit of the converted sample data becomes 1 based on the histogram of the target bit data;
Bit data conversion means for performing conversion using the lookup table for the target bit data,
A bit data arranging step of arranging the bit data so as to insert a predetermined number of bits of divisional bit data between the converted bit data;
A coding method for a time-series signal, comprising: In any one of claims 1 to 6,
The time-series signal is configured by a plurality of channels having a sample sequence,
Before the inter-frame operation step,
The method further comprises performing a predetermined operation on samples between channels, and separating a sample sequence of a portion having a high correlation between channels as inter-channel difference data from a sample sequence of one channel. Time-series signal encoding method. A recording medium which records a data group obtained by the method for encoding a time-series signal according to any one of claims 1 to 7 for a given time-series signal. A decoding method for decoding code data obtained by compression-encoding a time-series signal and reproducing all sample sequences of the time-series signal,
Obtaining a sample sequence restored to independent sample values for each time from the sample values recorded in the prediction error,
Integrating bit data constituting each sample of the restored sample sequence and lower bit data;
Signal flat portion insertion step of inserting signal flat portion data into each integrated sample sequence,
Restoring a sample sequence of the original frame based on the inter-frame difference data, and a frame data restoring step of inserting the sample sequence into the sample sequence;
A method for decoding a time-series signal, comprising: In claim 9,
The time-series signal is configured by a plurality of channels having a sample sequence,
After the frame data restoration step,
A method for decoding a time-series signal, further comprising restoring a sample sequence of an original channel based on the inter-channel difference data and inserting the sample sequence into the sample sequence. An encoding program for compressing the amount of information so that all of the sample sequences can be reproduced for a time-series signal composed of a time-series sample sequence,
On the computer,
A plurality of frames composed of a predetermined number of sample sequences are extracted from the sample sequences, and a correlation operation is performed between the frames. When the correlation between the frames is high, the sample sequence of one frame is compared with the inter-frame difference. As data, an inter-frame operation step of separating from the sample sequence,
In the sample sequence, a section in which the value of the sample is continuously within a predetermined range, as signal flat portion data, a signal flat portion separating step of separating from the sample sequence,
Upper / lower separation that divides each bit data constituting the sample sequence at a predetermined bit position and separates into upper sample data composed of a sample sequence of upper bits and lower sample data composed of a sample sequence of lower bits Stages,
An upper sample encoding step for encoding the upper sample data using a prediction error from a temporally past sample sequence,
The program to execute. A program that decodes encoded data obtained by compression-encoding a time-series signal and reproduces all sample sequences of the time-series signal,
On the computer,
Obtaining a sample sequence restored to independent sample values for each time from the sample values recorded in the prediction error,
Integrating the bit data constituting each of the restored samples and lower bit data;
Signal flat portion insertion step of inserting signal flat portion data into each integrated sample sequence;
Restoring a sample sequence of the original frame based on the inter-frame difference data, and restoring a frame data to be inserted into the sample sequence;
The program to execute.

Images