JP4483811B2 - Data compression method, data compression circuit, and data expansion circuit - Google Patents

Description

  The present invention mainly relates to a data compression method for compressing digital audio data, and more particularly to a data compression method, a data compression circuit, and a data decompression circuit that have a large compression ratio and can be reproduced in a loop from an arbitrary position.

  As is well known, as a method for compressing digital audio data, a method using subband coding such as MPEG (Moving Picture Experts Group) Audio Layer 2 is known. This compression method using subband coding has the advantage of a high compression ratio, but encoding / decoding is performed in units of frames consisting of a plurality of samples (for example, 1152 samples in the case of MPEG Audio Layer 2). In the case of performing loop reproduction (repetitive reproduction), there is a problem that reproduction from the middle of a frame is difficult.

If you use a compression method based on subband coding and loop playback from the middle of a frame, you can decode the entire frame including the playback point of the loop playback and discard the decoded data before the playback point. Loop playback from the middle of the frame is possible. However, in this case, the decoding time of the next frame is shortened, and there is a problem that the reproduced sound is interrupted when the decoding of the next frame cannot catch up with the reproduction.
Further, by separately preparing compressed data having the same content starting from a playback point at the time of looping and reproducing the data after the loop, loop playback can be performed in units shorter than one frame. However, in this case, there is a disadvantage that data from the playback point for the loop to the last frame is required separately, and the size of all the compressed data becomes large.

If the decoded sample from the playback point at the time of the loop to the next frame boundary is stored in the temporary storage memory, the data is played back to the frame boundary after the loop and then the next frame is played back. Loop playback can be performed in units smaller than one frame. However, in this case, there is a problem that a temporary storage memory for one frame is required for the loop, and the capacity of the temporary storage memory is increased.
Note that a technique described in Japanese Patent Application Laid-Open No. 2004-228867 is known as a conventional technique that uses a compression method based on subband coding and enables loop reproduction from the middle of a frame.
JP 2004-348055 A

  The present invention has been made in view of the above circumstances, and its purpose is to use a compression method based on subband coding and to increase the capacity of a temporary storage memory without greatly increasing the amount of compressed data. Another object of the present invention is to provide a data compression method, a data compression circuit, and a data expansion circuit that enable loop reproduction from the middle of a frame.

  The present invention has been made to solve the above-described problems, and the invention according to claim 1 divides a data aggregate composed of a plurality of data before compression into first to nth frames, and each frame is divided into frames. A first step of dividing the uncompressed data belonging to a plurality of subband signals and quantizing the subband signals based on psychoacoustic analysis to generate first to nth compressed frames; A second step of generating data for loop by compressing data before compression to final data before compression of a frame to which the loop reproduction point belongs by a compression method that does not cause decoding delay; and the first to nth compressions And a third step of generating a bit stream of compressed data by adding the loop data to a completed frame.

According to a second aspect of the present invention, a data aggregate composed of a plurality of uncompressed data is divided into first to nth frames, and uncompressed data belonging to each frame is divided into a plurality of subband signals. A first step of quantizing the band signal based on psychoacoustic analysis to generate first to n-th compressed frames, and the first to n-th compressed frames from the data before compression at a loop reproduction point; And a second step of generating a bit stream of compressed data by adding loop data consisting of uncompressed data up to the final pre-compression data of the frame to which the loop reproduction point belongs. It is.

  According to a third aspect of the present invention, a data aggregate composed of a plurality of data before compression is divided into first to nth frames, and a kth frame (k is a positive integer smaller than n) to which a loop reproduction point belongs is obtained. The uncompressed data belonging to each frame to be removed is divided into a plurality of subband signals, and the subband signals are quantized based on psychoacoustic analysis to generate first to nth compressed frames excluding the kth frame. A first step, a second step of compressing the data of the k-th frame before compression by a compression method in which a decoding delay does not occur, and the second step into the compressed frame generated by the first step And a third step of generating a bitstream by adding the compressed frames generated by the data compression method.

  According to a fourth aspect of the present invention, a data aggregate composed of a plurality of uncompressed data is divided into the first to nth frames, and the uncompressed data belonging to each frame is divided into a plurality of subband signals. A first step of quantizing the band signal based on psychoacoustic analysis to generate first to n-th compressed frames, from the pre-compression head data of the frame to which the loop end point belongs to the pre-compression data of the loop end point A second step of generating a loop data by compressing the data by a compression method that does not cause a decoding delay, and generating a bit stream of the compressed data by adding the loop data to the first to nth compressed frames And a third step of compressing the data.

  According to the fifth aspect of the present invention, a data aggregate composed of a plurality of data before compression is divided into first to nth frames, data before compression belonging to each frame is divided into a plurality of subband signals, A first step of quantizing the band signal based on psychoacoustic analysis to generate first to n-th compressed frames, and a pre-compression final frame of the frame to which the loop playback point belongs from the pre-compression data at the loop playback point A second step of generating loop data by compressing the data and the pre-compression head data of the frame to which the loop end point belongs to the pre-compression data of the loop end point by a compression method that does not cause a decoding delay; And a third step of generating a bit stream of compressed data by adding the loop data to the first to n-th compressed frames. It is a data compression method according to claim.

According to a sixth aspect of the present invention, a data aggregate composed of a plurality of data before compression is divided into first to nth frames, data before compression belonging to each frame is divided into a plurality of subband signals, A first step of quantizing the band signal based on psychoacoustic analysis to generate first to nth compressed frames, and compression of a frame to which a loop end point belongs to the first to nth compressed frames in the data compression method characterized by having a second step of generating a bit stream of the compressed data in addition to the loop data consisting of uncompressed data from the previous first data to the compression before the data of the loop end point is there.

According to a seventh aspect of the present invention, a data aggregate composed of a plurality of uncompressed data is divided into first to nth frames, and uncompressed data belonging to each frame is divided into a plurality of subband signals. A first step of quantizing the band signal based on psychoacoustic analysis to generate first to n-th compressed frames, and the first to n-th compressed frames from the data before compression at a loop reproduction point; , to the compression before the final data of the frame to which the loop playback point belongs, and in addition a loop for data consisting of uncompressed data from the compressed before beginning data frame loop end point belongs to the compression before the data of the loop end point And a second step of generating a bit stream of compressed data.

  According to an eighth aspect of the present invention, a data aggregate composed of a plurality of data before compression is divided into first to nth frames, and a kth frame (k is a positive integer smaller than n) to which a loop end point belongs. The uncompressed data belonging to each frame to be removed is divided into a plurality of subband signals, and the subband signals are quantized based on psychoacoustic analysis to generate first to nth compressed frames excluding the kth frame. A first step, a second step of compressing the data of the k-th frame before compression by a compression method in which a decoding delay does not occur, and the second step into the compressed frame generated by the first step And a third step of generating a bitstream by adding the compressed frames generated by the data compression method.

  The invention according to claim 9 divides a data aggregate composed of a plurality of data before compression into first to nth frames, and a kth (k is a positive integer smaller than n) frame to which a loop reproduction point belongs, And the uncompressed data belonging to each frame excluding the mth frame (m is a positive integer smaller than n) to which the loop end point belongs is divided into a plurality of subband signals, and the subband signals are quantized based on psychoacoustic analysis. A first step of generating first to nth compressed frames excluding the kth frame and mth frame, and compression without causing a decoding delay in the data of the kth frame and mth frame before compression A second step of compressing by the method, and adding the compressed frame generated by the second step to the compressed frame generated by the first step Is a data compression method which is characterized in that a third step of generating Tsu bets stream.

  According to a tenth aspect of the present invention, there is provided a dividing unit that divides a data aggregate composed of a plurality of data before compression into first to nth frames, and a plurality of pieces of data in each frame generated by the dividing unit. A first compression unit that divides the subband signal and quantizes the subband signal based on psychoacoustic analysis to generate compressed frames of the first to nth frames; Second compression means for generating data for loop by compressing the pre-compression final data of the frame to which the loop reproduction point belongs by a compression method that does not cause decoding delay; and the first to nth compressed frames A data compression circuit comprising bit stream generation means for adding a loop data to generate a bit stream of compressed data.

  According to an eleventh aspect of the present invention, there is provided a dividing unit that divides a data aggregate composed of a plurality of data before compression into first to nth frames, and a plurality of pieces of data in each frame generated by the dividing unit. A first compression unit that divides the subband signal and quantizes the subband signal based on psychoacoustic analysis to generate compressed frames of the first to nth frames; and before compression of the frame to which the loop end point belongs A second compression means for generating data for loop by compressing from the first data to pre-compression data at the loop end point by a compression method that does not cause a decoding delay; and for the loop in the first to nth compressed frames A data compression circuit comprising bitstream generation means for adding data and generating a bitstream of compressed data.

  According to the twelfth aspect of the present invention, a dividing unit that divides a data aggregate composed of a plurality of data before compression into first to nth frames, and a plurality of pieces of data in each frame generated by the dividing unit A first compression unit that divides the subband signal and quantizes the subband signal based on psychoacoustic analysis to generate compressed frames of the first to nth frames; The loop is compressed by a compression method that does not cause a decoding delay until the pre-compression final data of the frame to which the loop reproduction point belongs and the pre-compression data of the frame to which the loop end point belongs to the pre-compression data at the loop end point. Second compression means for generating data for use, and a bit stream of compressed data by adding the loop data to the first to nth compressed frames. A data compression circuit, characterized by comprising a bit stream generating means for generating a beam.

According to a thirteenth aspect of the present invention, a dividing unit that divides a data aggregate including a plurality of uncompressed data into first to nth frames, and uncompressed data belonging to each frame is divided into a plurality of subband signals. Compression means for quantizing the subband signal based on psychoacoustic analysis to generate first to n-th compressed frames, and the first to n-th compressed frames before being compressed at a loop reproduction point. And a bit stream generating means for generating a bit stream of compressed data by adding loop data consisting of uncompressed data from the data to the final pre-compression data of the frame to which the loop reproduction point belongs. A data compression circuit.

  According to the fourteenth aspect of the present invention, there is provided dividing means for dividing a data aggregate composed of a plurality of uncompressed data into first to nth frames, and a kth (k is a positive integer smaller than n) to which a loop reproduction point belongs. ) First to nth compressed frames excluding the kth frame by dividing the uncompressed data belonging to each frame except the frame into a plurality of subband signals, quantizing the subband signals based on psychoacoustic analysis The first compression means for generating the data, the second compression means for compressing the data of the k-th frame before compression by a compression method that does not cause decoding delay, and the compressed frame generated by the first step. And a bit stream generating means for generating a bit stream by adding the compressed frames generated in the second step. It is.

  A fifteenth aspect of the present invention is a data decompression circuit for decompressing a bitstream generated by the data compression circuit according to any one of the tenth to twelfth aspects, wherein the first to first belonging to the bitstream. A first decoding means for decoding the nth frame and returning it to the pre-compression data; and a second decoding means for decoding the loop data belonging to the bitstream and returning it to the pre-compression data and writing it to the buffer memory A data expansion circuit comprising: decoding means; and reading means for reading the first to nth frames and the loop data from the bitstream and outputting the data to the first and second decoding means. It is.

  In a sixteenth aspect of the present invention, in the data decompression circuit for decompressing the bitstream generated by the data compression circuit according to the thirteenth aspect, the first to nth frames belonging to the bitstream are decoded and compressed. Decoding means for returning to the previous data and writing to the buffer memory, reading the first to nth frames from the bit stream, outputting to the decoding means, and reading the loop data from the bit stream to the buffer memory A data decompression circuit comprising a reading means for writing.

  According to the present invention, there can be obtained an advantage that compression can be performed at a high compression rate by subband coding and loop reproduction can be performed from an arbitrary position in the middle of the frame. Further, according to the present invention, there is an advantage that it is not necessary to greatly increase the amount of compressed data for loop reproduction, and it is not necessary to increase the capacity of the temporary storage memory. Furthermore, according to the present invention, even when the end point of loop playback is in the middle of a frame, loop playback can be performed without greatly increasing the amount of compressed data for loop playback.

  Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of a data compression circuit according to a first embodiment of the present invention. This data compression circuit is a digital audio data (PCM) for loop reproduction before compression stored in a memory 1. Data). FIG. 2 is a diagram showing the compressed data (bit stream) generated by the data compression circuit. This bit stream is composed of a header H, compressed frames 1 to 4 and a final frame, and loop data attached to the rear part of the final frame. Here, the frame 1 to the final frame are compressed by, for example, MPEG Audio Layer 2. Point B is a loop playback point. During playback of this bit stream, after playback of frame 1 to the last frame, loop playback is performed by returning to loop playback point B. Point C is a boundary point between frame 2 including loop reproduction point B and the next frame 3. The loop data is data obtained by compressing data before compression corresponding to data between points B and C by ADPCM (Adaptive Differential Pulse Code Modulation) that can be decoded without delay.

Hereinafter, the data compression circuit of FIG. 1 will be described in detail.
In the figure, reference numeral 2 denotes a reading circuit, which reads data from the memory 1 in units of frames and outputs it to the MPEG encoder 3. Data corresponding to frame 2 is output to MPEG encoder 3 and data corresponding to between loop reproduction points B and C is output to ADPCM encoder 4.

  The MPEG encoder 3 is a circuit that compresses data input by, for example, a compression method based on MPEG Audio Layer 2, and divides the data before compression into a plurality of subband signals, and the subband signals are based on psychoacoustic analysis. The compressed data is generated by quantization and output to the bit stream generation circuit 5. The ADPCM encoder 4 compresses input data by the ADPCM method and outputs the compressed data to the bit stream generation circuit 5.

The bit stream generation circuit 5 adds the compressed data output from the ADPCM encoder 4 to the rear part of the compressed data of the frame 1 to the final frame output from the MPEG encoder 3, and adds a header H to the head of the data string. Generate a bitstream. In this case, data necessary for MPEG decoding is written in the header H, and the frame number (“5” in the example of FIG. 2) of the last frame and the frame to which the loop reproduction point B belongs are used as data for loop reproduction. A number (“2” in the example of FIG. 2) is written.
The bit stream generated by the bit stream generation circuit 5 is burned into the ROM 6.

Next, the data decompression circuit shown in FIG. 3 will be described.
In the figure, reference numeral 11 denotes a read circuit, which reads a bit stream in the ROM 6 and outputs the frame 1 to the final frame to the MPEG decoder 12 and the loop data to the ADPCM decoder 13. Details will be described later. Reference numeral 14 denotes a preset switch for the user to set the mode. That is, the data decompression circuit can reproduce data in the following three modes, and the user selects one of the three modes in advance by the preset switch 14.

(1) One-shot playback Play from frame 1 to the last frame once and finish.
(2) Head-loop playback Frame 1 to the last frame are played, then return to the beginning of frame 1 to play again from frame 1 to the last frame, and this operation is repeated.
(3) Loop playback on the way Play from frame 1 to the last frame, then return to the loop playback point B of frame 2 to play back the last frame, and repeat this operation.

The MPEG decoder 12 decodes each frame output from the readout circuit 11 to return it to PCM data before compression, and sequentially writes it into a FIFO (first-in / first-out) memory 15. The ADPCM decoder 13 decodes the loop data output from the readout circuit 11 to return it to the PCM data before compression, and sequentially writes it in the FIFO memory 15. The FIFO memory 15 sequentially stores the PCM data output from the MPEG decoder 12 and the ADPCM decoder 13 and sequentially outputs the internal data to the D / A (digital / analog) converter 16 at the timing of the sampling pulse fs. To do. The D / A converter 16 converts the PCM data output from the FIFO memory 15 into an analog tone signal and outputs it.
Next, the operation of the data decompression circuit described above will be described with reference to the operation flowchart of the read circuit 11 shown in FIG.

  At the start of reproduction, the reading circuit 11 first reads the bit stream header H from the ROM 6 and stores it internally (step S1), then reads the setting state of the preset switch 14, and sets the currently set mode number. Store it inside (step S2). Next, the internal data n is set to “1” (step S3). Next, compressed data of frame n (in this case, frame 1) is read from the ROM 6 (step S4), and the read data is output to the MPEG decoder 12. The MPEG decoder 12 decodes the frame 1 and outputs the decoded data to the FIFO memory 15. The FIFO memory 15 sequentially stores the data output from the MPEG decoder 12, and after the time when all the decoded data of the frame 1 is stored, the internal data is sequentially converted into the D / A converter 16 at the timing of the sampling pulse fs. Output to.

  After the output to the MPEG decoder 12 (step S5) is completed, the reading circuit 11 determines whether or not the output frame is the last frame from the header information stored therein in step S1 (step S6). If it is not the last frame, the internal data n is incremented (step S7), and the process proceeds to step S4. In step S4, since n = 2 in this case, frame 2 is read from ROM 6 and output to MPEG decoder 12 (step S5). If the decoding of frame 1 is not completed in the MPEG decoder 12, it is output after waiting for completion.

  The MPEG decoder 12 decodes the frame 2 and outputs the decoded data to the FIFO memory 15 as described above. The output data is read into the FIFO memory 15. The FIFO memory 15 sequentially outputs internal data in a first-in / first-out manner regardless of data reading.

  After executing step S5, the reading circuit 11 determines whether or not the frame output to the MPEG decoder 12 is the last frame (step S6). If the determination result is “NO”, the reading circuit 11 performs the step via step S7. Returning to S4, thereafter, steps S4 to S7 are repeatedly executed. As a result, the data from the frame 3 to the final frame are sequentially output to the MPEG decoder 12, sequentially decoded by the MPEG decoder 12, and the decoded data is written into the FIFO memory 15.

  When the data of the last frame is output from the readout circuit 11 to the MPEG decoder 12, the determination result in step S6 is “YES”, and the processing of the readout circuit 11 proceeds to step S8. In step S8, the reading circuit 11 checks the mode number stored therein. If the check result is mode (1) / one-shot reproduction, the processing of the readout circuit 11 is terminated. In this case, all the data in the FIFO memory 15 is output from the FIFO memory 15 and converted into an analog tone signal by the D / A converter 16, and the entire processing of the data decompression circuit in FIG. 3 is completed.

If the mode number is mode (2) / head loop reproduction, the processing of the readout circuit 11 proceeds to step S9. In step S9, the reading circuit 11 sets the data n to “1”, and the process proceeds to step S4. Thereafter, the frame 1 to the last frame are repeatedly decoded in the same manner as described above.
If the mode number is mode (3) and loop playback on the way, the processing of the reading circuit 11 proceeds to step S10. In step S10, the reading circuit 11 reads loop data from the ROM 6 and outputs it to the ADPCM decoder 13 (step S11). The ADPCM decoder 13 decodes the data and outputs it to the FIFO memory 15. The output data is written into the FIFO memory 15. As a result, the decoded loop data is written next to the decoded data of the final frame. As a result, after the final frame is converted into a musical tone signal, the loop data is converted into a musical tone signal. A musical tone signal corresponding to data between the loop reproduction points B and C shown in FIG. 2 is generated.

  After the process of step S11, the readout circuit 11 sets the number “3” next to the frame number to which the loop reproduction point B read from the header H in step S1 belongs to data n (step S12), and returns to step S4. . Thereafter, the frame 3 to the final frame are sequentially decoded, and then the loop data is decoded, and this operation is repeated.

  Thus, in the above embodiment, loop data is attached to the bitstream in advance, and the loop data is reproduced at the time of loop reproduction. On the other hand, after decoding the last frame, frame 3 is decoded instead of frame 2 at loop reproduction point B. That is, when the decoded data of the last frame and the decoded data of the loop data are sequentially output from the FIFO memory 15, the frame 3 is decoded and written into the FIFO memory. As a result, even when loop playback is performed from the middle of the frame, the playback sound is not likely to be interrupted due to a lack of decoding time.

  Next explained is the second embodiment of the invention. FIG. 5 is a diagram showing a bit stream used in the second embodiment. As shown in the figure, in this embodiment, (a) and (b) 2 format bit streams are used. In the bit stream (a), the LSP (loop start point) is the beginning of the frame (in the example shown, the beginning of frame 2), and the LEP (loop end point) is in the middle of the frame (in this example, the middle of frame 4). ing. In the bit stream (b), both LSP and LEP are in the middle of the frame (in the example shown, the middle of frame 1 and the middle of frame 4).

  In addition, at the end of the bit stream (a), uncompressed musical sound data corresponding to the period from the beginning of the loop end frame (frame 4 in the example in the figure) to LEP is compressed by ADPCM, and is used as loop data LD1. It has been added. Also, at the end of the bit stream (b), the uncompressed musical sound data corresponding to the period from the beginning of the loop end frame (frame 4 in the example in the figure) to LEP is compressed by ADPCM, and is used as loop data LD1. Subsequently, the uncompressed musical sound data corresponding to the portion from the LSP to the end of the loop start frame (frame 1 in the example in the figure) is compressed by ADPCM and added as loop data LD2.

Next, the operation of the data decompression circuit for decompressing the above-described bit frame will be described with reference to the flowchart shown in FIG. The block configuration of the data decompression circuit is the same as in FIG. Further, in FIG. 6, the same steps as those in FIG.
When the reproduction process is started, the reading circuit 11 first performs the above-described steps S1 to S5 sequentially. When the process of step S5 is completed, the MPEG decoder 12 decodes the frame 1 and outputs the decoded data to the FIFO memory 15. The FIFO memory 15 sequentially stores the data output from the MPEG decoder 12, and after the time when all the decoded data of the frame 1 is stored, the internal data is sequentially converted into the D / A converter 16 at the timing of the sampling pulse fs. Output to.

  After the output to the MPEG decoder 12 (step S5) is completed, the reading circuit 11 determines whether or not the output frame is a frame immediately before the loop end frame (frame 4 in the example of FIG. 5). The determination is made from the header information stored in S1 (step S6a). If the determination result is “NO”, the internal data n is incremented (step S7), and the process proceeds to step S4. In step S4, since n = 2 in this case, frame 2 is read from ROM 6 and output to MPEG decoder 12 (step S5). If the decoding of frame 1 is not completed in the MPEG decoder 12, it is output after waiting for completion.

  The MPEG decoder 12 decodes the frame 2 and outputs the decoded data to the FIFO memory 15 as described above. The output data is read into the FIFO memory 15. The FIFO memory 15 sequentially outputs internal data in a first-in / first-out manner regardless of data reading.

  After executing step S5, the reading circuit 11 performs the determination in step S6a again. Since the determination result is “NO”, the reading circuit 11 returns to step S4 via step S7. Next, steps S4 and S5 are executed, and the data of frame 3 is output to the MPEG decoder 12, and then the process proceeds to step S6a. In this case, the determination result of step S6a is “YES”, and the process proceeds to step S10a. In step S10a, the loop data LD1 is read from the ROM 6 (bit stream (a)), and if the loop data LD2 is added (bit stream (b)), the data LD2 is also read. Next, the read data is output to the ADPCM decoder 13 (step S11).

  The ADPCM decoder 13 decodes the data and outputs it to the FIFO memory 15. The output data is written into the FIFO memory 15. As a result, the decoded loop data LD1 (and LD2) is written next to the decoded data of frame 3, and as a result, after the frame 3 is converted into a musical sound signal, the loop data LD1 (and LD2) is converted. Is converted into a musical sound signal. Thereby, in the case of the bit stream (a), a musical sound signal corresponding to the data between the beginning of the frame 4 and the LEP shown in FIG. 5 is generated, and in the case of the bit stream (b), from the beginning of the frame 4. After the musical sound signal corresponding to the data up to LEP is generated, the musical sound signal corresponding to the data from the LSP of frame 1 to the last part of frame 1 is generated.

After the process of step S11, the read circuit 11 sets the loop start frame number (“2” in the example of FIG. 5) read from the header H in step S1 to the data n (step S12a), and proceeds to step S4. Return. Thereafter, the above-described operation is repeated.
As described above, according to the second embodiment, even when the LEP is in the middle of the frame, it is possible to perform loop reproduction without greatly increasing the amount of compressed data for loop reproduction.

  In the first and second embodiments, ADPCM is used as a compression method for loop data. However, this is not limited to ADPCM, and any compression method that does not require a delay time for decoding. Any method may be used, and uncompressed data may be used. When uncompressed data is used, the uncompressed data output from the reading circuit 11 is directly read into the FIFO memory 15 directly.

In the first and second embodiments, the loop data is the last part of the bit stream. However, this is not limited to the last part, and may be another position.
Instead of using the loop data, all the frames to which the LSP and LEP belong (frame 2 in FIG. 2 and frames 1 and 4 in FIG. 5) may all be compressed by a method such as ADPCM. In this case, the frame is decoded by a decoder different from other frames even in the case of one-shot reproduction or head loop reproduction.

  The present invention is used in the field of sound sources such as game machines and audio equipment.

It is a block diagram which shows the structure of the data compression circuit by 1st Embodiment of this invention. It is a figure which shows the bit stream produced in the data compression circuit. It is a block diagram which shows the structure of the data expansion circuit by embodiment of this invention. It is a flowchart for demonstrating operation | movement of the data expansion circuit. It is a figure which shows the bit stream produced by the data compression method by 2nd Embodiment of this invention. 6 is a flowchart for explaining the operation of a data decompression circuit for decompressing the bitstream shown in FIG. 5.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Memory, 2 ... Reading circuit, 3 ... MPEG encoder, 4 ... ADPCM encoder, 5 ... Bit stream generation circuit, 6 ... ROM, 11 ... Reading circuit, 12 ... MPEG decoder, 13 ... ADPCM decoder, 14 ... Preset switch, 15 ... FIFO memory, 16 ... D / A converter.

Claims (16)

A data set composed of a plurality of data before compression is divided into first to nth frames, data before compression belonging to each frame is divided into a plurality of subband signals, and the subband signals are based on psychoacoustic analysis. A first step of quantizing to generate first to nth compressed frames;
A second step of generating data for loop by compressing data before compression at a loop reproduction point to final data before compression of a frame to which the loop reproduction point belongs by a compression method that does not cause a decoding delay;
A third step of generating a bit stream of compressed data by adding the loop data to the first to n-th compressed frames;
A data compression method comprising: A data set composed of a plurality of data before compression is divided into first to nth frames, data before compression belonging to each frame is divided into a plurality of subband signals, and the subband signals are based on psychoacoustic analysis. A first step of quantizing to generate first to nth compressed frames;
Said first to n compressed frames, the compressed data in addition to the loop data consisting of uncompressed data from the data prior to compression in the loop playback point to the compressed before the final data of the frame to which the loop playback point belongs A second step of generating a bitstream;
A data compression method comprising: The data aggregate composed of a plurality of data before compression is divided into the first to nth frames, and the uncompressed data belonging to each frame excluding the kth frame (k is a positive integer smaller than n) to which the loop reproduction point belongs. Dividing the subband signal into a plurality of subband signals and quantizing the subband signals based on psychoacoustic analysis to generate first to nth compressed frames excluding the kth frame;
A second step of compressing the data of the k-th frame before compression by a compression method in which no decoding delay occurs;
A third step of generating the bitstream by adding the compressed frame generated by the second step to the compressed frame generated by the first step;
A data compression method comprising: A data set composed of a plurality of data before compression is divided into first to nth frames, data before compression belonging to each frame is divided into a plurality of subband signals, and the subband signals are based on psychoacoustic analysis. A first step of quantizing to generate first to nth compressed frames;
A second step of generating data for the loop by compressing from the pre-compression head data of the frame to which the loop end point belongs to the pre-compression data of the loop end point by a compression method that does not cause decoding delay;
A third step of generating a bit stream of compressed data by adding the loop data to the first to n-th compressed frames;
A data compression method comprising: A data set composed of a plurality of data before compression is divided into first to nth frames, data before compression belonging to each frame is divided into a plurality of subband signals, and the subband signals are based on psychoacoustic analysis. A first step of quantizing to generate first to nth compressed frames;
Decoding delay from the pre-compression data at the loop playback point to the final data before compression of the frame to which the loop playback point belongs, and from the pre-compression data of the frame to which the loop end point belongs to the pre-compression data at the loop end point, respectively. A second step of generating data for loop by compression by a compression method that does not occur;
A third step of generating a bit stream of compressed data by adding the loop data to the first to n-th compressed frames;
A data compression method comprising: A data set composed of a plurality of data before compression is divided into first to nth frames, data before compression belonging to each frame is divided into a plurality of subband signals, and the subband signals are based on psychoacoustic analysis. A first step of quantizing to generate first to nth compressed frames;
Bits of compressed data by adding loop data consisting of uncompressed data from the first data before compression of the frame to which the loop end point belongs to the pre-compression data of the loop end point to the first to nth compressed frames A second step of generating a stream;
A data compression method comprising: A data set composed of a plurality of data before compression is divided into first to nth frames, data before compression belonging to each frame is divided into a plurality of subband signals, and the subband signals are based on psychoacoustic analysis. A first step of quantizing to generate first to nth compressed frames;
Loops from the pre-compression data at the loop playback point to the pre-compression final data of the frame to which the loop playback point belongs and the pre-compression head data of the frame to which the loop end point belongs to the first to nth compressed frames A second step of adding a loop data consisting of uncompressed data up to the pre-compression data at the end point to generate a bit stream of compressed data;
A data compression method comprising: A data aggregate composed of a plurality of data before compression is divided into first to nth frames, and uncompressed data belonging to each frame excluding the kth (k is a positive integer smaller than n) frame to which the loop end point belongs. Dividing the subband signal into a plurality of subband signals and quantizing the subband signals based on psychoacoustic analysis to generate first to nth compressed frames excluding the kth frame;
A second step of compressing the data of the k-th frame before compression by a compression method in which no decoding delay occurs;
A third step of generating the bitstream by adding the compressed frame generated by the second step to the compressed frame generated by the first step;
A data compression method comprising: A data aggregate composed of a plurality of data before compression is divided into first to nth frames, and the kth (k is a positive integer smaller than n) frame to which the loop reproduction point belongs and the mth (to which the loop end point belongs). m is a positive integer smaller than n) The uncompressed data belonging to each frame excluding the frame is divided into a plurality of subband signals, the subband signals are quantized based on psychoacoustic analysis, and the kth and mth frames are quantized. A first step of generating first to nth compressed frames excluding frames;
A second step of compressing the data of the k-th frame and the m-th frame before compression by a compression method in which no decoding delay occurs;
A third step of generating the bitstream by adding the compressed frame generated by the second step to the compressed frame generated by the first step;
A data compression method comprising: A dividing unit that divides a data aggregate composed of a plurality of data before compression into first to nth frames;
The data in each frame generated by the dividing means is divided into a plurality of subband signals, and the subband signals are quantized based on psychoacoustic analysis to generate compressed frames of the first to nth frames. 1 compression means;
Second compression means for generating data for loop by compressing data before compression at a loop reproduction point to final data before compression of a frame to which the loop reproduction point belongs by a compression method that does not cause a decoding delay;
Bitstream generating means for generating a bitstream of compressed data by adding the loop data to the first to nth compressed frames;
A data compression circuit comprising: A dividing unit that divides a data aggregate composed of a plurality of data before compression into first to nth frames;
The data in each frame generated by the dividing means is divided into a plurality of subband signals, and the subband signals are quantized based on psychoacoustic analysis to generate compressed frames of the first to nth frames. 1 compression means;
Second compression means for generating data for the loop by compressing the pre-compression data of the frame to which the loop end point belongs to the pre-compression data of the loop end point by a compression method that does not cause a decoding delay;
Bitstream generating means for generating a bitstream of compressed data by adding the loop data to the first to nth compressed frames;
A data compression circuit comprising: A dividing unit that divides a data aggregate composed of a plurality of data before compression into first to nth frames;
The data in each frame generated by the dividing means is divided into a plurality of subband signals, and the subband signals are quantized based on psychoacoustic analysis to generate compressed frames of the first to nth frames. 1 compression means;
Decoding delay from the pre-compression data at the loop playback point to the final data before compression of the frame to which the loop playback point belongs, and from the pre-compression data of the frame to which the loop end point belongs to the pre-compression data at the loop end point, respectively. Second compression means for generating data for loops by compression by a compression method that does not occur;
Bitstream generating means for generating a bitstream of compressed data by adding the loop data to the first to nth compressed frames;
A data compression circuit comprising: A dividing unit that divides a data aggregate composed of a plurality of data before compression into first to nth frames;
Compression means for dividing uncompressed data belonging to each frame into a plurality of subband signals, and quantizing the subband signals based on psychoacoustic analysis to generate first to nth compressed frames;
Said first to n compressed frames, the compressed data in addition to the loop data consisting of uncompressed data from the data prior to compression in the loop reproduction point to the compression before the last data frame to which the loop playback point belongs Bitstream generation means for generating a bitstream;
A data compression circuit comprising: Dividing means for dividing a data aggregate composed of a plurality of data before compression into first to nth frames;
The uncompressed data belonging to each frame except the kth frame (k is a positive integer smaller than n) to which the loop reproduction point belongs is divided into a plurality of subband signals, and the subband signals are quantized based on psychoacoustic analysis. First compression means for generating first to nth compressed frames excluding the kth frame,
Second compression means for compressing the data of the k-th frame before compression by a compression method in which no decoding delay occurs;
Bitstream generation means for generating a bitstream by adding the compressed frame generated by the second step to the compressed frame generated by the first step;
A data compression circuit comprising: A data decompression circuit for decompressing a bitstream generated by the data compression circuit according to any one of claims 10 to 12,
First decoding means for decoding the first to nth frames belonging to the bitstream, returning them to pre-compression data, and writing to the buffer memory;
Second decoding means for decoding the loop data belonging to the bitstream and returning it to the pre-compression data and writing it into the buffer memory;
Reading means for reading the first to n-th frames and the loop data from the bitstream and outputting them to the first and second decoding means;
A data decompression circuit comprising: A data decompression circuit for decompressing a bitstream generated by the data compression circuit according to claim 13,
Decoding means for decoding the first to nth frames belonging to the bitstream and returning to the pre-compression data, and writing to the buffer memory;
Reading means for reading the first to n-th frames from the bitstream and outputting them to the decoding means; and reading the loop data from the bitstream and writing them into the buffer memory;
A data decompression circuit comprising:

Images