JP4906097B2 - Memory interface unit and memory unit - Google Patents

Description

  The present invention relates to a memory suitable for storing image data and the like.

  Patent Document 1 discloses a system called a multifunction printer (MFP) or a scanner printer that can easily use three functions of a scanner, a printer, and a copier, and other functions including a fax machine in one device. Is disclosed. In the MFP, a scanner and a printer are integrated, and the MFP operates alone, so that a copy operation is easy. Further, since it can be used as a scanner or a printer, the area efficiency is good and it is economical.

In Patent Document 2, in order to reduce the memory capacity, the rasterized data is subjected to variable-length reversible processing in a predetermined block unit, and the compressed data is stored in the compression memory. It is described that the data of the block is stored in an uncompressed memory without being compressed. Furthermore, the compression process is performed by one of the run-length compression method and the differential compression method, and if the compression is impossible, the other is performed. If compression is possible by either method, the compressed data is compressed into the compression memory. It is described that a flag indicating that is stored.
JP 2004-88275 A Japanese Patent Laid-Open No. 9-270919

  In a system that provides multiple functions at low cost, such as an MFP, a large-capacity memory is required at low cost. By storing the data in the memory by performing variable-length reversible compression processing such as a run-length compression method, the memory capacity for storing the data can be reduced, and as a result, the memory capacity can be increased. However, if lossless compression processing such as a run length compression method is performed on image data or the like, the amount of data that is intended to be compressed may increase rather than the original data amount.

  For this reason, in order to efficiently reduce the memory capacity for recording data and increase the effective capacity of the memory, it is desirable to select and record compressed data and non-compressed data according to the result after compression. . Further, by recording variable length compressed data in a memory with a variable length, the memory capacity for recording the data can be efficiently reduced, and the effective capacity of the memory can be increased. In Patent Document 2, the compressed data is stored in the compressed memory, and the non-compressed data is stored in the non-compressed memory, so that an advantage is obtained by selecting and recording the uncompressed data and the compressed data. . Patent Document 2 requires a management 32-bit fixed-length white map table (WMT), a 32-bit fixed-length compressed data table (CDT), and a 64-bit fixed-length uncompressed data table (UDT). The merit of variable length is not utilized. In addition, since it is necessary to input / output data to / from these three tables, the resources (including circuits, control microcode, etc.) and processing time required for data input / output have reached a level that can be adopted by the MFP. Absent.

One embodiment of the present invention is an input unit that writes image data including a plurality of fixed-length data blocks to a memory, and each fixed-length data block is a plurality of multi-value pixel data that forms one line of an image. include, adds a header data at the beginning of each fixed length data block, the memory includes an input unit for writing following the data written prior to the memory, a plurality of data blocks of a fixed length an output unit for outputting the image data, reads data from the memory, a memory interface unit and an output unit for outputting the read data following the header data as output data. The input unit performs variable-length lossless compression processing by run-length encoding of the data block, and generates a variable-length data block including a plurality of compressed data sets each having a predetermined number of bits composed of a combination of pixel data and the number of continuous pixels. a compression unit for, when the variable length data block is shorter than the fixed length, including information about the number of bits unit of compressed data sets included in the header data to the variable-length data blocks, instead of the data block, the variable-length data block And an input selection unit for selecting as data for writing to the memory. When the header unit contains information related to compression, the output unit uses the data set in units of the number of bits specified by the header data as a compressed data set , and decompresses the read data with the decompression unit. An output selection unit that outputs the read data as output data instead of the read data, and a data set following the read data when the read data or decompressed data following the header data reaches a fixed length. And a determination unit for determining the header data of the next data block .

  The input unit of the memory interface unit continuously writes a fixed-length input data block and a compressed variable-length data block with header data at the head in the memory. Therefore, the compressed data and the fixed-length data before being compressed can be selected and recorded according to the result after compression. Furthermore, it is possible to record a variable-length data block in a memory in a state where it is continuously mixed with a fixed-length data block (although header data is sandwiched) while maintaining a variable length. For this reason, the memory capacity for recording data can be efficiently reduced, and the effective capacity of the memory can be increased.

  The output unit of the memory interface unit determines that the data set following the read data is header data when the read data or the decompressed data reaches a fixed length by the determination unit. Therefore, it is possible to find the header data recorded in the memory sandwiched between the variable-length data block and the fixed-length data block. For this reason, if the fixed-length variable-length data block recorded in the memory mixed with the fixed-length data block is known, the output unit finds the header data independently of the input unit. The data can be decompressed and output as a fixed-length data block.

  Another aspect of the present invention is a memory unit having the memory interface unit described above and a memory body whose input / output is controlled by the memory interface unit. In this memory unit, fixed-length data is recorded in the memory body (memory) in a mixed state, with a fixed-length data block and a compressed variable-length data block sandwiched between header data. can do. Also, fixed-length data can be continuously output from fixed-length data blocks and variable-length data blocks mixed in the memory. Therefore, the effective recording capacity of the memory main body can be increased, and resources (including circuits, control microcode, etc.) required for data input / output to the memory main body and processing time can be reduced. Therefore, it is possible to provide a memory unit that is economical and has a large substantial recording capacity.

The variable length lossless compression process is run-length encoding. The compression unit generates a variable-length data block by run length encoding, and the expansion unit expands the data read by run length encoding. At this time, the determination unit can determine that the header data is a fixed length and a specified data set unit, and it is not necessary to determine the minimum input / output unit for the memory, for example, byte unit, and the processing time required for the determination Can be reduced. Run-length encoding, the data blocks, compresses the data block including a plurality of pixel data of the multi-level constituting one line of the image.

  Still another embodiment of the present invention includes the above-described memory unit, a scanner unit for acquiring image data including a plurality of pixel data, a printer unit capable of printing an image using the image data, and a memory unit. And a control unit that can output image data to a printer unit or an external device. This system includes an MFP.

  FIG. 1 shows a schematic configuration of an MFP according to an embodiment of the present invention. The MFP 10 is connected to a personal computer 3 as a host machine via a computer network such as the LAN 2 or the Internet, or a public telephone line. The number of connectable host PCs 3 is not limited to one. The MFP 10 includes a scanner unit 11 that scans a document (including images such as character information and photographs) to acquire image data, and a printer unit 12 that can print the image data on paper or the like. The scanner unit 11 and the printer unit 12 are connected to the control section 14 via the internal bus 13. The control section 14 includes a CPU 15, a ROM 16 containing a program for the CPU, a working RAM 17, and an interface unit 18 that can communicate with the host PC 3 via the network 2. The function as the control section 14 can be provided in other forms such as an ASIC or a system LSI. The MFP 10 also includes a memory unit 20 that functions as an internal buffer for providing an image read by the scanner unit 11 in units of lines to other functions.

  The MFP 10 uses these resources to provide various functions. One of them is a scanner, which provides image data acquired by the scanner unit 11 to the host PC 3. One of the other functions is a printer, and the printer unit 12 prints out image data provided from the host PC 3. Another one of the other functions is copying, in which the image data acquired by the scanner unit 11 is printed by the printer unit 12. Furthermore, various functions such as a fax and a photo printer can be included or added using the function of the host PC 3.

  A schematic configuration of the memory unit 20 is shown in FIG. The memory unit 20 functions as a buffer for supplying image data 19 from the scanner unit 11 to other functions of the MFP 10. The memory unit 20 is useful for enabling the scanner unit 11 to acquire the image data 19 at a unique timing and speed, and for enabling parallel operation with other functions included in the MFP 10. . In order to store a large amount of image data, it is desirable that the memory capacity of the memory unit 20 is large. However, since it is occupied by the scanner unit 11, it is difficult to adopt the memory unit 20 having a large memory capacity from an economical viewpoint.

  As shown in FIG. 2, the memory unit 20 includes a memory (memory main body) 21 and a memory interface unit 22 that controls input / output with respect to the memory 21. The memory interface unit 22 includes a memory controller 23, an input unit 24 for writing data to the memory 21 via the memory controller 23, and an output unit 25 for reading data from the memory 21 via the memory controller 23. ing. An example of the memory 21 is SDRAM or DDRSDRAM. The memory controller 23 includes read / write control, read / write arbitration, address control, command control, and an input / output buffer.

  The target of input / output of the memory unit 20 is a data block including a plurality of pixel data of multiple values, for example, 256 gradations (8 bits) constituting one line of the image data 19 of the document read by the scanner unit 11. . The memory unit 20 receives a fixed-length data block (input data block) 31 corresponding to one line of an image. The input data block 31 includes a predetermined number N (integer) of pixel data sets (one pixel 8 bits). Actually, the scanner unit 11 supplies a bit stream 32 read in units of lines, vertical and horizontal synchronization signals 33, and data 34 indicating the number of pixels (N) in units of lines designated by resolution conversion. Then, the fixed-length input data block 31 is cut out from the bit stream 32. The input unit 24 adds header data 41 to the head of the fixed-length input data block 31, supplies it to the memory controller 23 together with the write request 42, and writes data to the memory 21.

  When receiving the memory request (signal) 49 from the control section 14, the output unit 25 outputs a read request (signal) 43 to the memory controller 23. From the data supplied from the memory controller 23 to the read request 43, the output unit 25 outputs data following the header data 41 as output data 48.

  FIG. 3 shows a more detailed configuration of the input unit 24. The basic function of the input unit 24 is to add header data 41 to the head of the fixed-length input data block 31 and write data to the memory 21 following the data written prior to the memory 21. is there. Therefore, a continuous pixel number counter 51 that extracts fixed-length line data (input data block) 31 from continuous input pixel data 32 using vertical and horizontal synchronization signals 33, and a buffer that stores the input data block 31 for one line ( A delay buffer) 52a and a buffer 52h for adding a header. FIG. 4 shows a function of cutting out the fixed-length input data block 31 by the continuous pixel number counter 51. The read / write control function 53 for the buffer 52a is controlled by the continuous pixel number counter 51, and a fixed-length input data block 31 including a predetermined number N of valid pixel data is stored in the buffer 52a. In the buffer 52h, data indicating a fixed length is set in advance. The header data 41 is added to the head of the fixed length input data block 31 and output to the memory controller 23.

  The continuous pixel number counter 51 further has a function of compressing the fixed-length input data block 31 and generating the variable-length data block 35 by run-length encoding, which is one of the variable-length lossless compression processes. . That is, the continuous pixel number counter 51 is a data set of 8 bits × 3 (24 bits) or 8 bits × 2 (16 bits) consisting of pixel data and the number of continuous pixels. And the variable length data block 35 is generated. Whether the compressed data set is 24 bits or 16 bits can be selected in line units (data block units), and is written in the header data 41 of the header buffer 54h. Since the header data 41 is recorded in the memory 21 together with the variable length data block 35, the output unit 25 can reversibly expand the variable length data block 35 based on the header data 41.

  More specifically, the continuous pixel number counter 51 reads / write-controls certain pixel data and the result of continuing counting the number of pixels until the pixel data of the subsequent pixels change into the compressed data buffer 54a as one data set. Write via function 55. As a result, a variable-length data block (compressed data block) 35 including a plurality of data sets including pixel data and the number of pixels is generated in the compressed data buffer 54a.

  When the continuous pixel counter 51 processes one line of data (N pixel data), the compression efficiency determination circuit 56 determines the fixed-length input data block 31 based on the number of bytes of data included in the variable-length data block 35. It is determined whether or not there is a compression effect compared with the number of bytes (N bytes). If the data amount (number of bytes) of the variable length data block 35 is smaller than the data amount of the fixed length input data block 31, the variable length data block 35 is selected by the selector 57 and recorded in the memory 21. That is, when the variable-length data block 35 is shorter than the fixed length N, the input unit 24 includes information on compression of the variable-length data block 35 in the header data 41, and instead of the input data block 31, the variable-length data block An input selection unit 57 that selects 35 as data to be written in the memory 21 is included. When the variable length data block 35 and the fixed length data block 31 have the same data amount, either one may be selected in terms of writing to the memory 21. However, in the output unit 25, the compressed variable length data block 35 needs to be expanded, and continuous or burst read is difficult. Therefore, when the data amount is the same, it is preferable to record the fixed length data block 31.

  FIG. 5 shows a more detailed configuration of the output unit 25. Data 37 read by the output unit 25 via the memory controller 23 includes header data 41, data constituting the fixed-length data block 31, and data constituting the variable-length data block 35. Therefore, the output unit 25 includes a selector 62 for supplying the read data 37 to the header decoder 61 as the header data 41. The output unit 25 further outputs the read data 37 as data (uncompressed data) constituting the fixed-length data block 31 as it is, and data (compressed) constituting the variable-length data block 35. A selector 65 that distributes data to the decompression unit 64 is provided. The output unit 25 further includes an output selection unit (selector) 66 that outputs the data 38 expanded by the expansion unit 64 as output data 48 instead of the read data 37.

  The output unit 25 also includes a determination unit 67 that determines that the data set following the read data 37 is the header data 41 when the read data 37 or the decompressed data 38 reaches a fixed length (N bytes). . The selector 62 selects the header data 41 according to the determination result of the determination unit 67 and supplies it to the header decoder 61. The selector 65 selects data constituting the fixed-length data block 31 or data constituting the variable-length data block 35 according to the decoding result of the header decoder 61. If the data constitutes the variable-length data block 35, Supply to the expansion unit 64. When the header decoder 61 acquires the header data 41, the header decoder 61 outputs a read request 43 for reading the next data, and the data following the header data 41 is the data constituting the variable length data block 35 or the fixed length data block 31. A signal indicating whether it is data to be configured is output.

  The decompression unit 64 includes a selector 71 that selects the read data 37 as pixel data 39a or pixel number data 39b. As described above, the data set 39 composed of the pixel data 39a and the number of pixels 39b when recorded as the variable length data block 35 is 24 bits or 16 bits, and the information is included in the header data 41. . Therefore, the selector 71 can select the data 39a and 39b based on the decoding result of the header decoder 61. The pixel data 39a is output as decompressed data 38 by the subsequent holding circuit 72 by the number of pixels included in the pixel number data 39b. The holding circuit 72 includes a self-holding selector 72a and a register (flip-flop) 72b.

  The number of pixels 39b is supplied to the continuous pixel number counter 73 on the output side. The continuous pixel number counter 73 includes a counter 73a that counts down the pixel number 39b one by one with a clock and a comparator 73b that compares the result of the counter 73a. The comparator 73b outputs a read request 43 when the count result of the counter 73a becomes 0, sets new pixel data 39a in the pixel data holding circuit 72, and sets a new pixel number 39b in the counter 73a.

  The determination unit 67 detects, at the time of data decompression, header information added to the head of compressed data (variable length data block) obtained by compressing image data including pixel data of 256 gradations in units of one line using a run length compression method. The detection method using a counter is adopted. The head of the data first read from the memory 21 by the read request 43 based on the memory request 49 from the control section 14 is header data. Therefore, it can be determined whether the data 37 read subsequently is compressed data or non-compressed data. If it is non-compressed data, data for one line is output as it is, and if it is compressed data, decompression processing for one line is performed. The determination unit 67 includes two types of counters 75 and 76 that count the output total value of the number of pixel data after detecting the header data 41. The counter 75 is a counter that simply counts the number of bytes of the read data 37 in the case of reading the data of the fixed-length data block 31. The counter 75 outputs a signal 79 a for selecting the header decoder 61 to the selector 62 when the number of read data 37 reaches the fixed length N bytes by down-counting or up-counting.

  The counter 76 is a counter for determining that the decompressed data 38 has reached the fixed length N bytes when reading the data of the variable length data block 35. Therefore, the counter 76 compares the subtraction result with a line counter 76a that adds the number of pixels 39b to calculate the total, and a circuit 76b that subtracts the total value (bytes) of the line counter 76a from the fixed length N bytes. And a comparator 76c. If the comparison result is 0, the comparator 76c determines that the decompressed data 38 has reached a fixed length of N bytes, outputs a signal 79b for selecting the header decoder 61 to the selector 62, and further outputs a line counter. 76a and the counter 75 are reset.

  The determination unit 67 includes a selector 77 that selects a signal 79a generated by counting fixed-length data and a signal 79b generated by counting variable-length data. Depending on the result of decoding 41, either one is selected and the selector 62 is controlled. Thereby, even when the data for configuring the fixed-length data block 31 and the data for configuring the variable-length data block 35 are mixedly recorded in the memory 21, the header of the next data block is recorded. The data 41 can be found reliably, and the next data block can be reliably reproduced and output as the output data 48.

  With this method, the fixed-length N-byte numerical value recorded in the memory 21 in the input unit 24 is transmitted to the output unit 25 at a predetermined timing, so that the header data 41 can be reliably found and compressed and non-compressed. Reproduce and output compressed data. Therefore, even when the resolution of an image scanned in the middle changes and the numerical value of the fixed length N bytes changes, the image data can be reproduced. That is, the pixel number data of the data block (one line) 31 after a certain resolution conversion and the value of the line counter always coincide with each other at a certain point when the continuous output of arbitrary pixel data is finished. The matched next 8-bit data can be easily detected as the header of the next line.

  FIG. 6 schematically shows an example in which image data is recorded in the memory 21 of the memory unit 20. The data block 31 for one line of fixed length following the header data 41 and the data block 35 for one line of variable length following the header data 41 are recorded in the memory 21 in a mixed state. Can do. Therefore, only data effective for compression can be compressed and stored in the memory 21. Furthermore, fixed-length uncompressed data and variable-length compressed data are efficiently and wasted. So that it can be stored in the memory 21. For this reason, the substantial memory capacity of the memory 21 can be improved efficiently.

  Further, in the memory unit 20, the memory 21 is encapsulated by the memory interface unit 22, and the user of the memory unit 20 services the input / output of fixed-length data. Therefore, the user can use the memory unit 20 as a memory having an expected effective memory capacity larger than the memory capacity of the memory main body 21.

  In the above description, run-length encoding is adopted as a suitable example of variable-length lossless compression processing, but other variable-length lossless encoding methods such as a differential compression method can be used in combination. The memory unit is not limited to the MFP, and can be used as a memory unit for buffering continuous data including image data in other devices and applications such as a scanner and a printer.

1 is a diagram showing a schematic configuration of an MFP according to an embodiment of the present invention. The figure which shows schematic structure of a memory unit. The figure which shows schematic structure of an input unit. The figure which shows one function of the continuous pixel number counter. The figure which shows schematic structure of an output unit. The figure which shows an example of the storage state of a memory.

Explanation of symbols

10 MFP
20 memory units, 21 memory (memory body)
22 memory interface unit 23 memory controller, 24 input unit, 25 output unit

Claims (3)

An input unit for writing image data including a plurality of fixed-length data blocks into a memory, each fixed-length data block including a plurality of multi-valued pixel data constituting one line of an image, An input unit that adds header data to the beginning of a long data block and writes to the memory following the data written prior to the memory;
An output unit for outputting the image data including a plurality of data blocks of the fixed length reading data from said memory, and an output unit for outputting the read data following the header data as output data,
The input unit is
A compression unit that generates a variable-length data block including a plurality of compressed data sets each having a predetermined number of bits composed of a combination of pixel data and the number of continuous pixels by subjecting the data block to variable-length lossless compression processing by run-length encoding. When,
When the variable-length data block is shorter than the fixed length, the header data includes information on the number of bits of a compressed data set included in the variable-length data block , and the variable-length data is used instead of the data block. An input selection unit for selecting a block as data for writing to the memory,
The output unit is
When the header data includes information related to compression, a decompression unit that decompresses the read data using a data set in units of the number of bits specified by the header data as a compressed data set ;
An output selection unit that outputs the data expanded by the expansion unit as the output data instead of the read data;
A determination unit that determines a data set following the read data as the header data of the next data block when the read data or the decompressed data following the header data reaches the fixed length. , Memory interface unit. A memory interface unit according to claim 1 ;
A memory unit whose input / output is controlled by the memory interface unit. A memory unit according to claim 2 ;
A scanner unit for acquiring image data including the plurality of pixel data;
A printer unit capable of printing an image using the image data;
And a control unit capable of outputting the image data to the printer unit or an external device via the memory unit.

Images