M65762FP QM-Coder REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Description The M65762FP is a compression and decompression LSI conforming to the high efficiency encoding system (QMCoder) in the International Standard, the JBIG/JPEG (ITU-T Recommendations T.81 and T.82) for coding still images. It also conforms to the International Standard (ITU-T Recommendation T.85) for facsimile. The QM-Coder is an information dependent type which is capable of completely restoring original image data, and is equipped with the learning function to always optimize parameters according to the statistical characteristics of images. The QM-Coder is therefore superior in compression ratio compared with the existing binary coding system (MH/MR/MMR) and can greatly improve the half toning image (dithered half toning image) whose compression ratio is especially poor. Features • Completely conforms to the International Standard (ITU-T T.85) for facsimile. • Achieves encoding/decoding with the arithmetic coder (QM-Coder) conforming to the recommendation of the International Standard JBIG/JPEG. • Is expected to conform to the International Standard for color facsimile (T. Palette-colour). • High speed processing that puts into effect coding and decoding at 40 million pixels per second maximum. • Is possible data-through processing without coding and decoding. • Can select context Provides 10 pixel template model for minimum resolution conforming to JBIG and can select 2-line or 3-line template model. • Built-in typical prediction function Capable of coding and decoding by using the typical prediction. Since use of the typical prediction does not require the processing of the line (TP line) which is matched the previous line's data, is capable of reducing data and processing time. • Built-in adaptive template (AT) function Is capable of setting AT pixels before 127 pixels on the coding line. Since it is possible to change the position of AT pixel in a specified line, is capable of improving compression characteristics even when image characteristic is changed in the middle of the screen. • Supporting multi-stripe When a page consists of more than one stripe, is capable of repeating encoding/decoding process in stripes. • Built-in load/store function of line memory → Supporting multiple planes and multi-stripe function Is capable of loading image data for reference line from outside to line memory of the LSI and storing image data from line memory to outside. • Number of processing lines Is capable of issuing the start of processing (temporary stop command) several times to encode/decode any lines more than or equal to 65535 lines. • Supporting 3-bus interface An 8-bit host bus corresponds to the MPU is available to load and store of context table RAM. For input/output of binary image data, is capable of performing 32-bit or 16-bit parallel or serial input/output. For input/output of coding data, is capable of selecting 32-bit /16-bit/8-bit bus to perform DMA transfer of coding data. • Is capable of making scale-down for coding and scale-up for decoding. • Is capable of setting marker code for coding and detecting marker code for decoding • Built-in RAM for 4096 bytes for line memory, built-in context table RAM and built-in probability estimation table ROM of 113 status • +5 V single power supply REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 1 of 34 M65762FP Application • OA equipment including facsimile, copier and printer • Digital and amusement equipment for the purpose of reducing memory Block Diagram 55 PXCK* 56 PXCKO* 59 SVID* 57 RVID* Pixel data 58 Code data I/F Context table RAM Context generation 108 CDRQ 109 CDAK* 110 CDRD* 111 CDWR* 132 RESET* 135 HCS* HA0 to 3 Host bus I/F 54 PTIM* Probability estimation table ROM PRDY* Typical prediction 51 Line memory PDWR* 50 Image data I/F PDRD* 49 Parallel I/F PDAK* 48 Serial I/F PDRQ Encoding/decoding CD0 to 31 PD0 to 31 134 HWR* 133 HRD* HD0 to 7 112 INTR 129 MCLK (Asterisk "*" indicates negative logic.) Description on Block Functions (1) Host bus I/F block This bus is used to set command parameters and load the status between the MPU and this block. It is 8-bit bus. This block is also available to load and store of context table RAM via the host bus. (2) Code data I/F block Bus for input/output of coding data. For the bus width, 32 bits, 16 bits or 8 bits can be selected. Image data can also be transferred (in through mode) between the image data I/F and this block via built-in line memory. FIFO buffer for 16 bytes are provided in the code data I/F block. (3) Image data I/F block The Image data I/F is used for input/output of binary image data. The 32-/16-bit parallel I/F or serial I/F can be selected. Selection of the serial I/F transfers data in units of 1 pixel in synchronization with the line, using the handshake signal (PRDY*, PTIM*). Selection of parallel I/F uses an external DMA controller for DMA transfer (in units of stripe). The image data I/F provides a function for scale-down of length and breadth by 1/2 in coding and a function for scale-up of length and breadth by twice in decoding. (4) Line memory block 4 K-byte memory. This block can be set to a maximum of 8192 pixels/line for 3-line template and can be set to a maximum of 10240 pixels/line for 2-line template. A line is used for input/output processing of image data to/from outside and the other lines (2 or 3 lines) are used for encoding/decoding processing. These two processes can be independently carried out in synchronization with each line. The contents of line memory can be loaded or stored via the image data I/F or coding data I/F. (5) Typical prediction block In the typical prediction mode, compares the encoding/decoding process line agree with the immediately preceding line and generates pseudo-pixel (SLNTP). REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 2 of 34 M65762FP (6) Context generator By using the 10-pixel template of 2-lines or 3-lines, (including AT pixel) the standard context minimum of JBIG is generated with the resolution. (7) Context table RAM block Corresponds to the 10-bit standard context. This block can initialize, load and store the context table RAM. (8) Coding/decoding block This block performs arithmetic coding and decoding. It contains a ROM which contains a table capable of estimating 113 states and is capable of byte stuffing function ('OO' byte insertion/rejection) and is capable of end marker code control (Marker insertion/detection). REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 3 of 34 M65762FP 73 GND 74 CD8 75 CD9 76 CD10 77 CD11 78 CD12 79 VDD 80 GND 81 CD13 82 CD14 83 CD15 84 CD16 85 CD17 86 VDD 87 GND 88 CD18 89 CD19 90 CD20 91 CD21 92 CD22 93 VDD 94 GND 95 CD23 96 CD24 97 CD25 98 CD26 99 CD27 100 VDD 101 GND 102 CD28 103 CD29 104 CD30 105 CD31 106 VDD 107 GND 108 CDRQ Pin Arrangement CDAK 109 72 VDD CDRD 110 71 CD7 CDWR 111 70 CD6 INTR 112 69 CD5 VDD 113 68 CD4 GND 114 67 GND HD0 115 66 VDD HD1 116 65 CD3 HD2 117 64 CD2 HD3 118 63 CD1 HD4 119 62 CD0 VDD 120 61 GND GND 121 60 VDD HD5 122 59 PXCKO HD6 123 58 RVID HD7 124 57 SVID TEST0 125 56 PXCK 55 PTIM VDD 127 54 PRDY GND 128 53 GND MCLK 129 52 VDD VDD 130 51 PDWR GND 131 50 PDRD RESET 132 49 PDAK HRD 133 48 PDRQ HWR 134 47 GND HCS 135 46 VDD VDD 136 45 PD31 GND 137 44 PD30 HA0 138 43 PD29 HA1 139 42 GND HA2 140 41 VDD HA3 141 40 PD28 TOUT1 142 39 PD27 TOUT2 143 38 PD26 VDD 144 37 PD25 M65762FP 23 24 25 26 27 28 29 30 31 32 33 34 35 36 PD15 PD16 PD17 PD18 PD19 VDD GND PD20 PD21 PD22 PD23 PD24 VDD GND 15 GND 22 14 VDD GND 13 PD9 21 12 PD8 VDD 11 PD7 20 10 PD6 PD14 9 PD5 19 8 GND PD13 7 VDD 18 6 PD4 PD12 5 PD3 17 4 PD2 PD11 3 PD1 16 2 PD0 PD10 1 GND TEST1 126 (Top view) Outline: PLQP0144KA-A (144P6Q-A) REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 4 of 34 M65762FP Pin Description Pin No. I/O 1 Power supply Pin Name GND Pin No. I/O 51 I Pin Name PDWR Pin No. I/O 101 Power supply Pin Name GND 2 I/O PD0 52 Power supply VDD 102 I/O CD28 3 I/O PD1 53 Power supply GND 103 I/O CD29 4 I/O PD2 54 O PRDY 104 I/O CD30 5 I/O PD3 55 I PTIM 105 I/O 6 I/O PD4 56 I PXCK 106 Power supply 7 Power supply VDD 57 I SVID 107 Power supply 8 Power supply GND 58 O RVID 108 O CDRQ CD31 VDD GND 9 I/O PD5 59 O PXCKO 109 I CDAK 10 I/O PD6 60 Power supply VDD 110 I CDRD 11 I/O PD7 61 Power supply GND 111 I CDWR 12 I/O PD8 62 I/O CD0 112 O 13 I/O PD9 63 I/O CD1 113 Power supply VDD 14 Power supply VDD 64 I/O CD2 114 Power supply GND 15 Power supply GND 65 I/O CD3 115 I/O HD0 16 I/O PD10 66 Power supply VDD 116 I/O HD1 17 I/O PD11 67 Power supply GND 117 I/O HD2 18 I/O PD12 68 I/O CD4 118 I/O HD3 19 I/O PD13 69 I/O CD5 119 I/O HD4 20 I/O PD14 70 I/O CD6 120 Power supply VDD GND INTR 21 Power supply VDD 71 I/O CD7 121 Power supply 22 Power supply GND 72 Power supply VDD 122 I/O HD5 23 I/O PD15 73 Power supply GND 123 I/O HD6 24 I/O PD16 74 I/O CD8 124 I/O HD7 25 I/O PD17 75 I/O CD9 125 I TEST0 26 I/O PD18 76 I/O CD10 126 I TEST1 27 I/O PD19 77 I/O CD11 127 Power supply VDD 28 Power supply VDD 78 I/O CD12 128 Power supply GND 29 Power supply GND 79 Power supply VDD 129 I 30 I/O PD20 80 Power supply GND 130 Power supply VDD 31 I/O PD21 81 I/O CD13 131 Power supply GND 32 I/O PD22 82 I/O CD14 132 I RESET 33 I/O PD23 83 I/O CD15 133 I HRD 34 I/O PD24 84 I/O CD16 134 I HWR 35 Power supply VDD 85 I/O CD17 135 I HCS 36 Power supply GND 86 Power supply VDD 136 Power supply VDD 37 I/O PD25 87 Power supply GND 137 Power supply GND 38 I/O PD26 88 I/O CD18 138 I HA0 39 I/O PD27 89 I/O CD19 139 I HA1 40 I/O PD28 90 I/O CD20 140 I HA2 41 Power supply VDD 91 I/O CD21 141 I HA3 42 Power supply GND 92 I/O CD22 142 O TOUT1 43 I/O PD29 93 Power supply VDD 143 O 44 I/O PD30 94 Power supply GND 144 Power supply CD23 45 I/O PD31 95 I/O 46 Power supply VDD 96 I/O CD24 47 Power supply GND 97 I/O CD25 48 O PDRQ 98 I/O CD26 49 I PDAK 99 I/O 50 I PDRD 100 Power supply MCLK TOUT2 VDD CD27 VDD Notes: 1. Directly connect the input pin having pull-up (see "Description on Pin Functions") to VCC when the pin is not used. 2. Directly connect the input pin having pull-down (see "Description on Pin Functions") to GND when the pin is not used. 3. Connect test input pin TEST 0/1 to GND. 4. Leave test output pin TOUT 1/2 open. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 5 of 34 M65762FP Description on Pin Functions (Asterisk "*" in signal name indicates negative logic.) I/F Host bus I/F Pin Name RESET* HCS* HA0 to 3 HWR* HRD* HD0 to 7 INTR I/O I I I I I I O BUF S CD0 to 31 I/O UR8 CDRQ CDAK* CDRD* CDWR O I I I 4 US US US Coding data input/output bus signal (CD0 to 15 is used in 16-bit bus and CD0 to 7 is used in 8-bit bus.) DMA request signal for coding data (image data) DMA acknowledge signal for coding data (image data) Read strobe signal for coding data (image data) Write strobe signal for coding data (image data) Parallel PD0 to 31 PDRQ PDAK* PDRD* PDWR* I/O O I I I UR8 4 US US US Parallel image data input/output bus (PD0 to 15 is used in 16-bit bus.) DMA request signal for image data DMA acknowledge signal for image data Read strobe signal for image data Strobe signal for image data Serial PRDY* PTIM* PXCK* PXCKO* O I I O 4 US US 4 SVID* RVID* MCLK TEST0, 1 I O I I U 4 1-line input/output start ready signal for image data 1-line transfer sector signal for image data Transfer clock signal for image data Transfer clock signal for image data (LSI internal loop back output signal of PXCK*) Image data input signal Image data output signal DS Master clock input signal Test input signal 0/1 (Should be connected to GND when used normally.) VDD GND — — — — Power supply (+5 V) Ground Code data I/F Image data I/F Others Note: S S R8 4 Function H/W reset signal Chip select signal Address select signal of internal register Write strobe signal Read strobe signal Input/output data bus signal Interrupt request signal Input buffer for the input pins ("I" and "IO") are set at the TTL level and the options are as follows. (U: Having pull-up resistance, D: Having pull-down resistance, S: Schmitt trigger, R: Through rate control) Numbers (4, 8) in the BUF column for the output pins ("O" and "IO") indicate Io (= 4 or 8 mA). Specifications (1) Package Plastic QFP 144 pins (20 mm × 20 mm) (2) Power consumption 5 V, 120 mA (600 mW) (3) Maximum clock frequency 40 MHz REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 6 of 34 M65762FP Specifications of Coding Functions (1) Coding algorithm • QM-Coder (JBIG standard arithmetic coding system) (2) Context a) Template model • 2- or 3-line of 10-pixel template (see figure 1) (Conforming to the template for JBIG minimum resolution) Note: The coding efficiency of the 3-line template is better than that of 2-line template by several %. b) Adaptive template (AT) • It is possible to move up to 127 pixels on the coding line. (AT position is indicated by MPU.) Note: AT is available to improve the coding efficiency for dither image. • Even in the middle of coding/decoding, the position of AT line can be changed for a line (AT move) Note: When the position the AT pixel of is changed, the template model cannot be changed concurrently. X X X X X X X X X X ? X X A X X X X X ? X A Figure 1 Template (X, A) (Upper: 3 lines, Lower: 2 lines) X X A X X X X X ? X X MAX127 A X X X X X X X X ? X MAX127 Figure 2 Adaptive Template (A) (3) Typical prediction • Agreement with the typical prediction of the minimum resolution of JBIG. The pseudo-pixel (SLNTP) is generated by the symbol LNTP which shows whether the coding/decoding process lines agree with the immediately preceding line. If they agree, the pseudo-pixel only is coded. This makes it possible to shorten the time of process and rejection of the code data. SLNTPy = ! (LNTPy ⊕ LNTPy – 1) (Where: y indicates a lien No., y = 1 indicates that lines do not match each other, and initial value LNTP for head line is given with y – 1 = 1) (4) Coding data format • The stripe data entity. (SDE = stripe coded data with byte stuffing (PSCD) + end marker (SDNORM/SDRST)) Performs coding and decoding of one stripe (see "Appendix A.1") In the case of multi-striped (multi-stripes), can be supported by activation for each stripe. (5) Marker code • Supports the SDE end marker (During coding, the marker code previously set in the register is outputted. During decoding, the marker code byte detected by requesting on interrupt to MPU when the maker is detected is read out of the register.) REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 7 of 34 M65762FP (6) Estimation of coding/decoding speed Figure 3 compares the estimation of coding/decoding speed between the M65762FP and the existing product type (M65760/1FP). Polygonal lines in the diagram are processing speeds of images theoretically generated assuming the unmatched estimation ratio as a parameter. In addition, { ∆ indicate processing speeds of real image (without TP function). As shown in this diagram, the M65762FP has been largely improved in the processing speed compared with existing product types. If the compression ratio is reduced, the reduction ratio of processing speed is moderated. When a theoretical image is used to compare processing speeds in the worst case, the processing speed of existing product type is about 9.4 M pixels/s (1 / compression ratio ≈ is about 1), while the processing speed of the M65762FP is about 27.5 M pixels/s (1 / compression ratio ≈ 0.9) for coding and is about 31.2 M pixels/s (1 / compression ratio ≈ 0.75) for decoding. 20 Coding/decoding of existing product type (M65760/1FP) 15 Cafeteria, error diffusion image 25 Coding of M65762FP Baud rate, error diffusion image Processing Speed (M pixels/s) 30 Decoding of M65762FP Cafeteria and dither images 35 Baud rate and dither images Average of test charts 1 to 8 of former CCITT 40 10 (Legend) Theoretical image Actual image Decoding of M65762FP 5 Coding of M65762FP Coding/decoding of existing product type 0 0 0.25 0.5 0.75 1 / Compression Ratio Figure 3 Estimated Processing Speed REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 8 of 34 1.0 1.2 M65762FP Register Configuration 1. List of Registers Table 1 Address List of Registers Register Name R/W 0 System setting W/R 1 Parameter setting W/R 2 Command 2 Status 3 Interrupt enable setting W R W/R Content • • • • • • • • • LSI H/W reset Selects bit width of code data bus (32 bits/16 bits/8 bits) Selects coding (image) data byte swap on code data bus Selects coding (image) data bit swap on code data bus Selects image data bit swap on image data bus Selects image data I/F (parallel I/F and serial I/F) Selects bit width of image data bus (32 bits/16 bits) Template selection (3-line template/2-line template) Sets up the AT pixel position (127 max) (When set to 0, selects non-AT (default position)) • Context table RAM initializing processing command • Start/stop command • • • • • • • • • • (Coding/decoding, image data through, load/store of the line memory) Start/stop command of load/store of context table RAM Selects temporary stop/termination end mode Processing status (in process/end of process) Ready for reading/writing coding (image) data on code data bus Detects marker code (SDNORM, SDRST, ABORT, etc) Interrupt request status SC counter overflow error Processing mode (temporary stop/end of termination) Interrupt enable setting corresponding to each bit position of status register Indicates pause/restart with marker code detected (at time of decoding) Sets the number of pixels per line (a maximum of 10240 pixels with 2-line template selected) 4, 5 Setting number of pixels W/R • 6, 7 Setting number of lines W/R • Sets the number of lines to be coded/decoded 8, 9 Number of processing lines R A Load/store buffer (1 line or more, a maximum of 65535 lines) B Operation mode setting • Number of setting the coded/decoded lines (a maximum of 65535 lines) W/R W/R • Buffer register that loads/stores context table RAM data from the • • C C Marker code setting Marker code reading W R MPU (RAM address is automatically incremented each time data is written/read.) Sets the operation mode (Coding/decoding, image data through, and load/store of line memory) Selects read-through of head coding data in decoding (0 to 3 bytes) Selects the typical prediction function Selects prohibition of line memory initialization • • • Sets the terminal marker code in encoding (SDNORM/SDRST) • Reads a marker code in decoding (SDNORM, SDRST, ABORT, others) D Scale-up/scale-down setting W/R • Scale down in coding (1/2 scale-down of horizontal and vertical, horizontal OR processing) • Scale-up at time of decoding (scale-up of horizontal and vertical by twice) REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 9 of 34 M65762FP 2. Description on Register (1) System setting register (W/R) (Address: 0) SYS_REG: d0 (HR): d7(MSB) PB PI BX BS DS d0 (LSB) HR CB H/W reset (0: Active status, 1: Reset status) To reset H/W, set this bit to 1 then to 0. The entire LSI including register group and line memory is initialized by writing in this reset. However, context table RAM is not initialized. d1 to 2 (CB): Selects the bit width of code data bus (d2 = 0, d1 = 0: 8-bit bus (CD0 to 7), d2 = 0, d1 = 1: 16-bit bus (CD0 to 15), d2 = 1, d1 = 0: 32-bit bus (CD0 to 31)) d3 (DS): Notes: 1. Prohibition of setting for d2 = 1, d1 = 1 2. For encoding in 16-/32-bit bus, the last encoding data is output followed by bit byte of "00" (3 bytes maximum) for word alignment of encoding data at the end. Selects data bit swap of image data bus (0: MSB first, 1: LSB first) → see table 3. d4 (BS): Selection of data bit swap of code data bus (0: MSB first, 1: LSB first) → see table 2. d5 (BX): Selection of data byte swap of code data bus (0: low order byte first, 1: high order byte first) → see table 2. Note: BX is effective only when the host bus selects 16-bit/32-bit bus. Selection of image data input/output I/F (0: serial I/F, 1: parallel I/F) d6 (PI): d7 (PB): Selection of bit width of image data bus (0: 32-bit bus (PD0 to 31), 1: 16-bit bus (PD0 to 15) → see table 3. Note: PB and DS are effective only when PI = 1. Table 2 Line Up of Coded Data/Image Data in Code Data Bus Bus Width (CB) Swap (BX, BS) d2 d1 1 0 (32 bits) d5 0 0 1 d4 0 1 0 CD31 b24 b31 b0 ••• ••• ••• ••• CD24 b31 b24 b7 CD23 b16 b23 b8 ••• ••• ••• ••• CD16 b23 b16 b15 CD15 b8 b15 b16 ••• ••• ••• ••• CD8 b15 b8 b23 CD7 b0 b7 b24 ••• ••• ••• ••• CD0 b7 b0 b31 1 0 0 1 1 0 1 0 b7 ••• — — — b0 b15 ••• — — — b8 b23 b8 b15 b0 ••• ••• ••• ••• b16 b15 b8 b7 b31 b0 b7 b8 ••• ••• ••• ••• b24 b7 b0 b15 1 — — 1 0 1 b7 ••• — — b0 b15 b0 b7 ••• ••• ••• b8 b7 b0 0 1 (16 bits) 0 0 (8 bits) Note: — — — — — — b0 is image data, given in time series, on the left side of the first encoding data/screen. b31 is image data, given in time series, on the right side of the last encoding data/screen. Table 3 Bit Width Order of Image Data on Image Data Parallel Bus Swap PB = 0 DS = 0 DS = 1 PB = 1 DS = 0 DS = 1 Note: Order of Data in Code Data Bus (CD) PD31 ••••• PD16 PD15 ••••• PD0 p0 p31 ••••• ••••• p15 p16 p16 p15 ••••• ••••• p31 p0 p0 p15 ••••• ••••• p15 p0 — — p0 is image data on the left side of the screen. p31 is image data on the right side of the screen. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 10 of 34 M65762FP (2) Parameter setting register (W/R) (Address: 1) PARA_REG: d7 d6 d5 TM AT d4 d0 AT d0 to 4 (AT<0> to AT<4>): Low order 5 bits of AT pixel position (see figure 2) d5 (TM): Selection of template (0: 3-line template, 1: 2-line template) d6 to 7 (AT<5> to AT<6>): High-order 2 bits of AT pixel position (6th/7th bit) (Example) 3-line template, AT = 4 2-line template, AT = 48 Note: d7 d4 d0 0 0 0 0 0 1 0 0 0 1 1 1 0 0 0 0 AT pixel position is set (0 to 127) with AT<6:0>. At the default position (AT pixel is not used), set AT = 0. The 2line template, prohibits AT = 1 to 4 from being set. The 3-line template prohibits AT = 1 to 2 from being set. (3) Command register (W) (Address: 2) CMD_REG: d0 (IC): d7 0 d3 JP RC JC d0 IC Context table RAM initialization start command (1: Start initialization) Setting this bit to 1 starts to initialize context table RAM. When the initialization is completed automatically returns this bit to 0. d1 (JC): Processing (coding/decoding/through) start/end command (1: Start of processing, 0: End of processing) Setting this bit to 1 starts processing (coding/decoding, image data through and lead/store of line memory). Before the issuance of this command, concrete operation mode must be set in the operation mode setup register. When the processing for the number of setup lines ends with the end of termination selected this bit automatically returns to 0. d2 (RC): Note: When this JC bit is set to 0 during the coding process (is in progress,) and input of image data is stopped, the coding is stopped (flashed) even if the set lines are not filled. When this bit is set to 0 during decoding process, and input of encoding data ceases, processing for the number of setup lines is carried out assuming coding data "00" to have been input. In the case of multistripe coding, however, process must not be stopped by setting this bit to 0 except for the final stripe. Load/store start/end command of context table RAM (1: Start of load/store, 0: End of load/store) Setting this bit to 1 can load context data into context table RAM from outside via a buffer register or can store context data in outside. (see the section for buffer register.) When load/store processing is completed, this bit must be set to 0. d3 (JP): Temporary stop mode of processing (coding/decoding/through)/termination end mode selection (1: Selection of temporary stop, 0: Selection of termination end) Issuance of processing start command d1 (JC) with this JP bit set to 1 temporarily stops performing the process operation at the completion of processing for the number of setup lines. After that, reissuance of processing start command d1 (JC) restarts processing. (see " Register Setting Sequence" (3)) REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 11 of 34 M65762FP (4) Status register (R) (Address: 2) STAT_REG: d0 (JS): d7 0 d5 PS SC IS MS DS d0 JS Processing (initialization/coding/decoding/through) status (0: Processing in progress (temporary stop or initial), 1: Completion of processing) This JS bit is set to 1 in the following cases: when the initialization is complete with the RAM initialization command issued (IC = 1), when all coding data is read completely at time of coding with the start command of termination end processing issued (JC = 1, JP = 0), and when all image data is read completely at time of image data through and at time of decoding. When the temporary stop processing start command is issued (JC = 1, JP = 1), this JS bit remains to be 0, even if the process for the number of setup lines ends. (However, an interruption occurs at time of temporary stop.) d1 (DS): Ready for reading/writing coding data (image data case of the through mode) on the code data bus (1: Ready, 0: Read/write disabled) When this bit is set to 1, data can be read/written on the code data bus. (This bit is equivalent to the CDRQ pin.) d2 (MS): Detects marker code at time of decoding (0: Not detected, 1: Detected) This bit is set to 1 when some marker code is detected at time of decoding. d3 (IS): Status of interrupt request (INTR pin) (0: Not requested, 1: Requested) d4 (SC): SC count-over error at time of coding (0: Normal, 1: Occurrence of SC counter overflow) Note: The SC counter is a counter for consecutive "FF" data bytes generated in the coding process. Though coding process continues if the SC counter overflows, normal coding data is not output (encoding error). d5 (PS): Processing (temporary stop/termination end) mode (1: Temporary stop processing mode, 0: Termination end processing mode) This PS bit corresponds to the selection of process temporary stop/termination end of the d3 (JP) bit of command register. (5) Interrupt enable register (W/R) (Address: 3) IENB_REG: d0 (JE): d7 MP 0 d3 SE ME DE d0 JE Processing (initialization/coding/decoding/through) Temporary stop/termination end interrupt (0: Interrupt mask, 1: Interrupt enable) d1 (DE): Coding data (image data) read/write ready interrupt (0: Interrupt mask, 1: Interrupt enable) d2 (ME): Marker code detection interrupt at time of decoding (0: Interrupt mask, 1: Interrupt enable) d3 (SE): SC count-over error interrupt at time of coding (0: Interrupt mask, 1: Interrupt enable) Note: Bits d0 to d3 are interrupt enable of bits d0 to d2 and d4 corresponding to the status register. When one of the status bits set to interrupt enable is set to 1, the interrupt request signal (INTR) is asserted (for d0 (JE), an interrupt occurs even at the time of temporary stop). When the status is set to 0 by H/W reset etc., or when interrupt factor is eliminated by interruption masking, INTR is negated. The status register is not cleared by occurrence of interruption or by R/W of interruption enable register. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 12 of 34 M65762FP d7 (MP): Indication of pause at time of marker code detection (0: Indication of continuation/restart, 1: Indication of temporary pause) If this MP bit is in advance set to 1 in decoding, the decoding temporarily pauses at the time of marker code detected. (When the ME bit is set to 1, an interruption occurs when marker code is detected.) When decoding process is not completed at time of temporary pause of marker detection, the register for setting the number of lines can be respecified (see item (7)). Afterwards, setting this MP bit to 0 restarts the decoding process (the decoding process is carried out for the number of set lines). (6) Register for setting the number of pixels (W/R) (Address: 4) PEL_REG_L: d7 d0 PEL_L (Address: 5) PEL_REG_H: 0 d7 PEL_H d5 d0 d0 to 7 (PEL_L): Sets the number of pixels in a line. (Low byte) d0 to 5 (PEL_H): Sets the number of pixels in a line. (Upper byte) A maximum of 8192 pixels can be set at the 3-line template. A maximum of 10240 pixels can be set at the 2-line template. Set the number of pixels to be actually coded (decoded) at time of scale-up (scale-down). When the image data bus is 16 bits (32 bits) with the parallel I/F selected, set the number of pixels to multiples of 16 (multiples of 32). With the serial I/F selected, set the number of pixels to multiples of 8. (7) Register for setting the number of lines (W/R) (Address: 6) d7 d0 LSET_REG_L: (Address: 7) LSET_L LSET_REG_H: LSET_H d0 to 7 (LSET_L): Sets the number of lines to be processed. (Low order byte) (1 to 65535: 0 line is not allowed.) d0 to 7 (LSET_H): Sets the number of lines to be processed. (High order byte) At time of scale-down (scale-up), set the number of lines to be actually coded (decoded). Set the number of lines (number of relative lines) ranging from the processing start command to be issued next to the temporary stop/termination end just after. This register must be set to a specific value before the issuance of the process start command. As far as the following conditions are satisfied, this register can be rewritten in the course of processing. • When the maximum value (65535) is set before issuance of the processing start command, an arbitrary value can be set once in the course of processing. • When a value except for the maximum value (65535) is set before issuance of the processing start command, and the value requires to be respecified in the course, respecify the maximum value (65535) once and then respecify a desired value. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 13 of 34 M65762FP (8) Processing line count register (R) (Address: 8) LIN_REG_L: d7 d0 LINE_L (Address: 9) LIN_REG_H: LINE_H d0 to 7 (LINE_L): Read out the number of lines actually processed (Low byte) (0 to 65535) d0 to 7 (LINE_H): Read out the number of lines actually processed (Upper byte) The number of processed lines number of set lines, coding/decoding/through processing stop temporary/end of processing. Note: The number of lines in this process is cleared to 0 with the processing start command issued. (9) Buffer register (W/R) (Address: A) d7 d0 DWR_BUF: d0 to 7 (DWR): DWR Data for loading/storing context table RAM This register is a buffer for loading data into the context table RAM via the host bus or for storing data outside. After issuance of load/store start command of the context table RAM (command register d3 = 1), this register is available to start loading or storing data. Prediction value (MPS) and prediction unmatched probability (LSZ) can be stored in context table RAM for a unit of 1024 contexts in total. Figure 4 and table 4 provide the address assignment of context table RAM and the data bit array. Since context table RAM is 2-byte data, access is gained alternately in order from low byte to upper byte. Each time two-byte access is gained, the RAM address is automatically incremented (sequential access from address 0). Notes: 1. Data is not allowed to be loaded and stored at a time. Random access to RAM is not allowed. 2. Only 133 types specified by the JBIG international standard (see " Appendix A.2") are allowed to be specified for the LSZ value. (For example, load '5a1d' for initialization.) 8 7 6 5 4 3 2 9 1 0 ? 8 5 4 3 2 9 7 6 1 0 ? 3-line template 2-line template Figure 4 Address Assignment of Context Table RAM (Number for Address Bit (LSB: 0, MSB: 9), MSB: 9 for AT Pixel) Table 4 Data Bit Array of Context Table RAM High Order Byte d15 MPS Note: d14 L14 ••••• ••••• Low Order Byte d8 L8 d7 L7 MPS: Prediction value MPS (0/1) L14 to 0: Low 15 bits of prediction unmatched probability LSZ ('0001' to '5b12') REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 14 of 34 ••••• ••••• d0 L0 M65762FP (10) Operation mode setting register (W/R) (Address: B) MOD_REG: d7 d0 TP LI OB LIO MOD This register is used to set the LSI operation mode and requires to be set before issuance of the processing start command (command register d1 (JC) = 1). d0, 1 (MOD): Operation mode setting (d1 = 0, d0 = 0: Coding, d1 = 1, d0 = 0: Image data through (image data I/F → Code data I/F) load/store, d1 = 0, d0 = 1: Decoding, d1 = 1, d0 = 1: Image data through (Code data I/F → Image data I/F) load/store) d2, 3 (LIO): Load/store selection of image data of line memory (d2 = selection of load, d3 = selection of store) In the case of multi-stripe, this LIO bit is set according to the following table, to load image data for reference line from outside into line memory before coding/decoding of stripes or to store image data stored in line memory into outside after encoding/decoding of stripes. This LIO bit is effective only in the image data through mode (d1 = 1). Notes: 1. LIO (d3, d2) = (1, 1) not allowed being set. 2. When selection of load/store of image data of line memory, temporary stop (d3 (JP) = 1 of command register) is not allowed to be set. 3. When load/store mode of image data is selected, the number of lines to be transferred must be set in the register setting the number of lines. 4. The number of lines for image data load to line memory must be 2-line either case of 2line template or 3-line template. (This is because typical prediction (LNTP) cannot be judged correctly with only a line.) Table 5 Operation Mode List Operation Mode (d1, d0) Load/Store LIO (d3, d2) Operation Mode Remarks 0 0 0 1 X X X X Coding mode Normal coding mode Decoding mode Normal decoding mode 1 0 0 0 0 1 Image data through (image data I/F → code data I/F) For inter-I/F transfer of image data Image data load to line memory (input from image data I/F) For loading of reference line to LSI 1 0 0 0 Image data store of line memory (output to code data I/F) For storing line memory to outside Image data through (code data I/F → image data I/F) For inter-I/F transfer of image data 0 1 1 0 Image data load to line memory (input from code data I/F) For loading of reference line to LSI Image data store of line memory (output to image data I/F) For storing line memory to outside 1 1 d4, 5 (OB): Sets head of the coding data read-through at time of decoding (0 to 3: Sets the number of read-through bytes. For example, with d4 = 0 and d5 = 1, read-through of 2 bytes) When OB is set to 1 to 3 at time of decoding, and the first stripe decoding processing start command is issued, the head data for the number of set bytes is to be read through (not used for decoding process). With OB set to 0, no data is read through (normal decoding process). For example, if the code data bus is 32/16 bits, and the head of coding data does not contact the word boundary, this function is used. Note: When the code data bus is 8 bits, this function is effective. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 15 of 34 M65762FP d6 (LI): Prohibition of line memory initialization (0: Indication of initialization, 1: Prohibition of initialization) When first stripe coding/decoding process start command is issued, and LI = 1, initialization of built-in line memory is prohibited. (The final image data, coded/decoded just before, that is left in line memory is used as the reference line data at the head of next coding/decoding operation). With LI = 0, built-in line memory is initialized. (Full white (0) data is used as the reference line data at the head of next coding/decoding operation.) When the previous stripe is terminated at the SDNORM marker with coding/decoding of the multistripe configuration, this bit is set to initialization prohibition (1) to make the data of previous stripe left in line memory available as the coding reference line data of the next stripe. (For details, see " Register Setting Sequence " (6) sequence.) d7 (TP): Note: With LI =1, this LI bit is cleared (to 0) by H/W reset writing to an external reset pin or system setup register. At the same time, built-in line memory is also initialized. Selection of typical prediction at time of coding/decoding (0: Sets typical prediction function to OFF, 1: Sets typical prediction function to ON.) This bit is set to 1 when encoding/decoding process is carried out using the typical prediction function. (11) Marker code set up register (W) (Address: C) d7 MSET_REG: d0 to 7 (MSET): d0 MSET The end marker code used during coding is set (SDNORM = 02h, SDRST = 03h, etc.) The byte set to this register is output attached to coding data as the end marker during coding. (12) Marker code read out register (R) (Address: C) MDET_REG: d7 d0 d0 to 7 (MDET): Reads out the marker code detected during decoding (SDNORM = 02h, SDRST = 03h, ABORT = 04h, etc.) MDET Marker code bytes detected at time of decoding can be read directly. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 16 of 34 M65762FP (13) Scale-up/scale-down set register (W/R) (Address: D) CONV_REG: d7 0 d4 HO d0 HR VR HE d0 (VE): Selection of scale-up in vertical direction during decoding (0: Equal size, Scale-up by twice) d1 (HE): Selection of scale-up in horizontal direction during decoding (0: Equal size, Scale-up by twice) VE Note: Scale-up function is effective only in decoding (Scale-up enabled) d2 (VR): Selection of scale-down in vertical direction (0: Equal size, Scale-down by 1/2) d3 (HR): Selection of scale-down in horizontal direction (0: Equal size, Scale-down by 1/2) d4 (HO): Selection of thinned-out processing in horizontal direction (0: Simple thinned-out, 1: OR processing) Note: Scale-down function is effective only in encoding (Scale-down enabled) Notes: 1. During coding, simple thinned-out is applied to 1/2 scale-down in vertical direction (Odd lines are skipped in reading.) 2. With VR = 1 during coding, the number of lines on input image data must be larger by twice than the set value of line count setup register. 3. With VE = 1 during decoding, the number of lines on output image data must be larger by twice than the set value of line count setup register. 3. Register Initial Value Registers are initialized as provided in the following table by writing H/W reset into the external reset pin or system setup register. Table 6 Initial Values of Registers Register System setting Parameter setting Command Status Interrupt enable Pixel setting Line count setting Note: Initial Value 00h (Note) 00h 00h 00h 00h 00h 00h Register Number of processed lines Buffer register Operation mode setting Marker code setting Marker code reading Scale-up/scale-down setting Initial Value 00h Indefinite 00h 00h 00h 00h When H/W reset is written into the system setting register, written value is set in the system setting register. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 17 of 34 M65762FP 4. Register Setting Sequence (1) Initialization sequence of built-in line memory and context table RAM This sequence is used to carry out initialization sequence (0 clear) of context table RAM after the initialization (Note) of the built-in line memory by H/W reset. When the initialization is unnecessary (the contents of the current status table are directly used), this sequence is unnecessary. 1 d7 H/W Reset Context mode set up d0 SYS_REG: 0 0 0 0 0 0 0 1 ; H/W reset bit ON SYS_REG: 0 0 0 0 0 0 0 0 ; H/W reset bit OFF * Period of H/W reset bit set to ON (time from when d0 = "1" is written until d0 = "0" is written) requires 100 ns or more. Issue context table RAM initialization command CMD_REG: 0 0 0 0 0 0 0 1 ; Initializes context table RAM Set interrupt enable IENB_REG: 0 0 0 0 0 0 0 1 ; Process end interrupt enable Context table RAM is initialized (0 clear) in this period. The number of clocks required for initialization is as follows: 1024 + a [Clock] (Occurrence of interrupt) d7 d0 Set interrupt disable IENB_REG: 0 0 0 0 0 0 0 0 ; Interrupt disable Read out status register (check the end of processing) STAT_REG: − − − − − − − j ; j = End of processing CMD_REG: 0 0 0 0 0 0 0 0 ; End of initialization N j=1? (Error) Y End of initialization command 2 Note: To 2) Line memory is initialized by H/W reset to prepare the all white (0) data as a reference line to provide for the start of coding/decoding process and to initialize LNTP bit (LNTP = 1) for typical prediction. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 18 of 34 M65762FP (2) Stripe coding/decoding (without change in AT pixel position)/image data through processing sequence 2 d7 Set system (set LSI mode) SYS_REG: Set operation mode MOD_REG: ; Cb, Cb = Bit width of code data bus ; Bs, Bx = Code data bus bit, byte swap ; Pb, Pi = Bit width of image data bus, I/F selection ; mm = Operation mode (coding/decoding/through) ; Ob, Ob = Selection of head byte read-through during decoding (0 to 3) ; Li = Selection of inhibition of line memory initialization Note ; Tp = Typical prediction function ON/OFF ; aa, aaaaa = AT pixel position ; t = Template selection d0 Pb Pi Bx Bs 0 Cb Cb 0 Tp Li Ob Ob 0 0 m m Note: Set Li = 0 for the head stripe of single stripe of multi stripe. Set parameter (template, context) PARA_REG: a a Set the number of pixels PEL_REG_L: PEL_REG_H: 0 0 t a a a a pel_l ; pel_l, pel_h = Number of pixels per 1 line pel_h LSET_REG_L: LSET_REG_H: lset_h Set marker code (Note: Required cording only) MSET_REG: mset Set scale-up/scale-down Set the number of lines a lset_l ; lset_l, lset_h = Number of processing lines ; mset = sets marker code byte (SDNORM = 02h, SDRST = 03h) CONV_REG: 0 0 0 Ho Hr Vr He Ve Processing start command (coding/decoding/through) CMD_REG: 0 0 0 0 0 0 1 0 Set interrupt enable IENB_REG: 0 0 0 0 0 0 0 1 ; Ve, He = Selection of scale-up during decoding ; Vr, Hr, Ho = Selection of scale-down at time of coding ; Termination end processing (coding/decoding/through) Start command ; Process end interrupt enable [Performs coding/decoding processing during this period.]−−−Inputs/outputs image data and coding data (Coding/decoding/through processing for a stripe.) (Occurrence of interrupt) d7 d0 Set interrupt disable IENB_REG: 0 0 0 0 0 0 0 0 ; Interrupt disable Read out status register (check process for end) STAT_REG: − − − s − m − j ; j = End of processing ; m = Marker detection ; s = SC counter over error N j=1? (Error) Y N (Coding) Decoding ? Y (Decoding) N (Marker not detected) (Error) End (Error) (Marker detection) Read marker code (Note: At time of decoding only) N (SC counter over) Y m=1? Y s=0? MDET_REG: End REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 19 of 34 mdet ; mdet = Read marker code M65762FP (3) Stripe encoding/decoding (with change in AT pixel position) processing sequence 2 ; Cb, Cb = Bit width of code data bus ; Bs, Bx = Code data bus bit, byte swap ; Pb, Pi = Bit width of image data bus, I/F Pb Pi Bx 0 Cb Cb 0 Bs SYS_REG: selection ; m = Operation mode (encoding/decoding) ; Ob, Ob = Selection of head byte read-through MOD_REG: Tp Li Ob Ob 0 0 0 m during decoding (0 to 3) ; Li = Selection of inhibition of line memory Note: Set Li = 0 for single stripe or the head initialization Note ; Tp = Typical prediction function ON/OFF stripe of multi-stripe. ; aa,aaaaa = AT pixel position PARA_REG: a a t a a a a a ; t = Template selection d7 Set system (set LSI mode) Set operation mode Set parameter (template, context) d0 Set the number of pixels PEL_REG_L: PEL_REG_H: Set the number of lines LSET_REG_L: LSET_REG_H: lset_h Set marker code (Note: Required cording only) MSET_REG: mset Set scale-up/scale-down CONV_REG: 0 0 0 Ho Hr Vr He Ve Processing start command (temporary stop processing) CMD_REG: 0 0 0 0 1 0 1 0 Set interrupt enable IENB_REG: 0 0 0 0 0 0 0 1 pel_l 0 0 ; pel_l, pel_h = Number of pixels per 1 line pel_h ; lset_l, lset_h = Number of processing lines lset_l Note: Set the number of processing lines to position change of AT pixel. ; mset = Sets marker code byte (SDNORM = 02h, SDRST = 03h) ; Ve, He = Selection of scale-up at time of decoding ; Vr, Hr, Ho = Selection of scale-down at time of coding ; Temporary stop processing (coding/decoding) Start command ; Process end interrupt enable [Performs coding/decoding processing during this period.]−−−Input/output first image data and coding data. Repeat this routine (for the number of AT moves − 1) Note: At time of coding in the first processing, (number of lines of input image data) = (value set in the line count set register) + 1. During decoding, (number of lines in output image data) = (value set in the (Occurrence of interruption) line set register) − 1 d7 d0 Set interrupt disable IENB_REG: 0 0 0 0 0 0 0 0 ; Interrupt disable Read status register STAT_REG: − − p − − − − j ; Status check ; j = 0, p = 1, temporary stop status a' a' t' a' a' a' a' a' Set final AT Final set 3 Set in the course Set AT pixel position Set the number of lines PARA_REG: LSET_REG_L: LSET_REG_H: ; Set change of AT pixel (a'a', a'a'a'a'a') Note: Template is not allowed to be changed ; lset_l, lset_h = Number of processing lines lset_l Note: Set the number of processing lines ranging from processing restart to change of AT pixel position lset_h Processing start command (temporary stop processing) CMD_REG: 0 0 0 0 1 0 1 0 Set interrupt enable IENB_REG: 0 0 0 0 0 0 0 1 ; Temporary stop processing (encoding/decoding) Start command ; Process stop interrupt enable [Performs encoding/decoding process during this period.]−−−Input/outputs image data and coding data in the course. Note: During encoding in the course of processing, (number of lines of input image data) = (value set in the line count set register). At time of decoding, (number of lines in output image data) = (value set in the line count set register). REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 20 of 34 M65762FP 3 Set AT pixel position PARA_REG: a" a" t a" a" a" a" a" ; Set change in final AT pixel. (a"a", a"a"a"a"a") Note: Template is not allowed to be changed LSET_REG_L: LSET_REG_H: Set number of lines ; lset_l, lset_h = Number of processing lines lset_l Note: Enter the number of processing lines ranging from restart of processing to the final line. lset_h Processing start command (termination end processing) CMD_REG: 0 0 0 0 0 0 1 0 Set interrupt enable IENB_REG: 0 0 0 0 0 0 0 1 ; Termination end processing (coding/decoding) Start command ; Process stop interrupt enable [Performs coding/decoding processing during this period.]−−−Inputs/outputs final image data and coding data. Note: During coding in the final processing, (number of lines in input image data) = (value set in the line count set register) − 1. During decoding, (number of lines in output image data) = (value set in the line count (Occurrence of interrupt) set register) + 1. d7 d0 Set interrupt disable IENB_REG: 0 0 0 0 0 0 0 0 Read out status register (check end of processing) STAT_REG: − − − s − m − j N ; Interrupt disable ; j = End of processing ; m = Marker detection ; s = SC counter over error j=1? (Error) Y N (Encoding) Decoding ? Y N (SC counter over) s=0? (Decoding) (Error) N (Marker not detected) Y m=1? End (Error) Y (Marker detection) Read out marker code (Note: Decoding only) MDET_REG: ; mdet = Read marker code mdet End (4) Load/store processing sequence of the context table RAM This sequence is used to load or store context table RAM. d7 RAM load/store start command CMD_REG: 0 d0 0 0 0 0 0 0 0 ; Starts to load/store context table RAM [Stores (loads) the context table RAM during this period.] Context RAM data is stored (loaded) via buffer register. Reading (writing) 2 bytes automatically increments the RAM address. Note: Reading (storing) operation and writing (loading) operation are not allowed to be done at a time. End of RAM load/store command CMD_REG: REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 21 of 34 0 0 0 0 0 1 0 0 ; End of loading/storing RAM Since the operation does not automatically stop, be sure to write the load/store end command. M65762FP (5) Load/store processing sequence of line memory image data 2 d7 SYS_REG: Set system (set LSI mode) d0 Pb Pi Bx Bs Ds Cb Cb 0 Set operation mode MOD_REG: Set parameter* (selection of template) PARA_REG: − − Set number of pixels* PEL_REG_L: PEL_REG_H: 0 0 Set number of lines (= 2) Tp Li 0 0 Lio Lio 1 t − − − − pel_l ; m = Operation mode (selection of through mode) ; Lio, Lio = 01 or 10 (selection of load or store) ; Li = 1 (selection of prohibition of line memory initialization) ; Tp = Typical prediction function ON/OFF Note1 ; t = Selection of template ; pel_l, pel_h = Number of pixels per line pel_h LSET_REG_L: LSET_REG_H: Set scale-up/scale-down* − m ; Cb,Cb = Bit width of code data bus ; Ds = Bit swap of image data bus ; Bs, Bx = Code data bus bit, byte swap ; Pb, Pi = Bit width of image data bus, I/F selection lset_l ; lset_l, lset_h = 2 (Number of processed lines)Note2 lset_h CONV_REG: 0 0 0 Ho Hr Vr He Ve Processing start command (load/store into line memory) CMD_REG: 0 0 0 0 0 0 1 0 Set interrupt enable IENB_REG: 0 0 0 0 0 0 0 1 ; Ve, He = Selection of scale-up during decoding ; Vr, Hr, Ho = Selection of scale-down during coding ; Load/store processing start command of image data ; Process end interrupt enable * Settings of template selection, number of pixels per line, selection of scale-up/scale-down and typical prediction function must meet the settings at time of stripe coding/decoding to be carried out after this. [Performs loading/storing process during this period.]−−−Inputs (outputs) image data. (Transfer processing of image data for 2 lines) (Occurrence of interrupt) d7 d0 Set interruption disable IENB_REG: 0 0 0 0 0 0 0 0 ; Interrupt disable Read out status register (check end of processing) STAT_REG: − − − − − − − j ; j = End of processing N j=1? (Error) Y End Stripe Head line Notes: 1. For ON/OFF bit of TP function in the image data processing, the ON/OFF bit of the TP function just before coding/decoding shall be kept. 2. In the image data load/store processing, be sure to set the number of transfer lines to "2". (The 1st line is data on the line (final line – 1) of the stripe. The 2nd line is data on the last line of stripe.) When a line stripe is adopted for the first stripe of the page in the image data store processing, and read out line of the first line is outside data of stripe, the all white data must used for replacement or the image data load function must be used in advance to clear line memory. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 22 of 34 • • • (Final line − 1) line 1st line Final line 2nd line M65762FP (6) Total sequence of multi-stripe coding/decoding For an image with a page consisting of more than one stripe or plane, coding or decoding process must be carried out in units of stripe after initialization. Multi-stripe coding/decoding Initialization of built-in memory and context table RAM [Process (1)] 1st stripe coding/decoding processing [Process (2) or (3)] End of processing of all stripes? Y End of page Repetition of this routine (for the number of stripes − 1) N Is previous stripe SDNORM? N Y (SDNORM) Y (Same plane) (SDRST) [case 1] Does the same plane stripe continue? N (Difference plane) [case 3] Initialization of line memory and context table RAM [Process (1)] Loading of image data of line memory, and loading of context table RAM [Processes (4) and (5)] [case 2] Stripe coding/decoding processing (Indication of line memory initialization: Li = 0, AT pixel = Default position (0)) [Process (2) or (3)] Stripe coding/decoding processing (Prohibition of line memory initialization: Li = 1, AT pixel = Previous stripe taken over) [Process (2) or (3)] Stripe coding/decoding processing (Prohibition of line memory initialization: Li = 1, AT pixel = Respecify) [Process (2) or (3)] Storing of image data of line memory and storing of context table RAM [Processes (4) and (5)] Notes: 1. 2. 3. Since use of the host bus with 32-/16-bit bus during coding adopts word boundary, the end marker code may be followed by the pad bytes ("00") of 1 to 3-byte. These pad bytes must be removed outside. (see "Description on Register" (7)) When decoding of stripes starts at time of decoding, the head coding data of SDE (stripe data entity) must be first entered. Read-through of head byte is indicated, if necessary. (At time of end of decoding stripes, the head block of coding data may be entered into LSI (FIFO) or may not be arranged in the word boundary. Management is therefore required outside.) The process of inter-stripe marker codes (ATMOVE, NEWLEN, etc.) (insert at time of coding and detection/removal at time of decoding) must be carried out outside. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 23 of 34 M65762FP Description If the end marker of the previous stripe is SDRST, the status must be initialized for coding/decoding the next stripe. Start to carry out the process of next stripe by returning the AT pixel position to the default position after the initialization of built-in line memory and context table RAM. [case 1] If the termination marker of the previous stripe is SDNORM, the status of the previous stripe must be taken over for coding/decoding the next stripe. If the stripe of the same plane is continuously coded/decoded, the AT pixel position takes over the final value of the previous stripe and the process of the next stripe is to start without initializing line memory and context table RAM to use the status of line memory and context table RAM at the end of previous stripe for the next stripe. [case 2] On the other hand, since the status at the end of pre-stripe status of the same plane must be respecified for the status of line memory and context table RAM, line memory and context table RAM are to be loaded into LSI to respecify the AT pixel position and to start processing the next stripe when alternately coding/decoding stripes of different planes. After coding/decoding of stripe, save line memory and context table RAM for next stripe. [case 3] (Example) • Single plane, multi-stripe Plane 1 2 • Multiple planes and multi-stripe 3 Plane 1 1 4 7 10 Plane 2 2 5 8 11 Stripe Plane 3 (Processes in numeric order) 3 6 9 12 4 REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 24 of 34 M65762FP Timing Chart 1. Host Bus I/F CS* RD* WR* A0 to 3 D0 to 7 Read access Write access 2. Code Data I/F (a) For 8-bit bus CDRQ CDAK* CDRD*/CDWR* CD0 to 7 (b) For 16-bit bus CDRQ CDAK* CDRD*/CDWR* CD0 to 15 Note: For 16-bit bus, only the word access (CD0 to 15) is allowed. (c) For 32-bit bus CDRQ CDAK* CDRD*/CDWR* CD0 to 31 Note: For 32-bit bus, only the long word access (CD0 to 31) is allowed. Description CDRQ can be checked for being asserted (H) to assert (L) CDAK*. Asserting (L) CDAK* negates (L) CDRQ. Asserting (L) section of CDRD*/CDWR* must be included in the CDAK* asserting section (L). REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 25 of 34 M65762FP 3. Image Data I/F (1) Serial image data I/F PRDY* PTIM* PXCK* PXCKO* SVID*/RVID* 1 2 3 4 5 N Note: The above chart shows a timing for a line (N pixel/line). Description PRDY* can be checked for being asserted (L) to assert (L) PTIM*. Asserting (L) PTIM* negates (H) PRDY*. PXCKO* is an output of having gated PXCK* input with PTIM*. The image data (SVID*/RVID*) is input/output in synchronization with PXCK* or PXCKO*. (2) Parallel image data I/F (a) 16-bit bus PDRQ PDAK* PDRD* /PDWR* PD0 to 15 Note: For 16-bit bus, only the word access (PD0 to 15) is allowed. (b) 32-bit bus PDRQ PDAK* PDRD* /PDWR* PD0 to 31 Note: For 32-bit bus, only the long word access (PD0 to 31) is allowed. Description PDRQ can be checked for being asserted (H) to assert (L) PDAK*. Asserting (L) PDAK* negates (H) PDRQ. Asserting (L) section of PDRD*/PDWR* must be included in the asserting section (L) of PDAK*. REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 26 of 34 M65762FP Examples of System Configuration 1. Example of Application for Digital PPC and Multifunctional FAX Machine Image sensor Binaly image processing Printer No frame memory required by QM-Coder Reduction of memory size and increase of memory efficiency by QM-Coder QM-Coder (M65762FP) MPU QM-Coder (M65762FP) DMAC Signed memory Communications control Disk device High-speed communications by QM-Coder Figure 5 Example of Application for Digital PPC and Multifunctional FAX Machine 2. Example of Application for Printer High-speed communications by QM-signed data Signed memory QM encoder (hardware or software) Image data file QM decoder (M65762FP) Printer (LBP) PC/WS Reduction of memory size by real-time decoding Figure 6 Example of Application for Printer REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 27 of 34 Memory M65762FP [Appendix A.1] JBIG Data Structure B I E ; Bi-level B I H DL D P XD YD LD MX MY Order 1 1 1 1 4 4 4 1 1 1 HITOLO SEQ ILEAVE SMID Options 1 Image Entity ; Bi-level Image Header ; lowest resolution layer ; final resolution layer ; number of bit-planes ; dummy 0 ; horizontal dimension at highest resolution ; vertical dimension at highest resolution ; number of lines per stripe at lowest resolution ; maximum horizontal offsets allowed for AT pixel ; maximum vertical offsets allowed for AT pixel ; order byte b7-4 b3 b2 b1 b0 ; dummy 0 ; resolution-order distinction ; progressive-versus-seqential distinction ; interleaving of multiple bit-planes ; indexed over stripe is in middle ; option byte LRLTWO VLENGTH TPDON TPBON DPON DPPRIV DPLAST DPTABLE 0/1728 b7 ; dummy 0 b6 ; lowest resolution-layer two line template b5 ; NEWLEN (new vertical dimension) marker enable b4 ; differential-layer TP enable b3 ; lowest-resolution-layer TP enable b2 ; DP enable b1 ; private DP table b0 ; DP table last is to be reused ; private DP table (It is present only if DPON = 1, DPPRIV = 1, DPLAST = 0) B I D ; bi-level Image Data (( 1 2 ) x N) 1 Floating Marker Segments ( a ~ c ) a AT move marker ESC ATMOVE YAT τX τY 1 1 4 1 1 ; FFh ; 06h ; line in which an AT switch is to be made ; horizontal offset of the AT pixel ; vertical offset of the AT pixel 1 1 4 ; FFh ; 05h ; new YD 1 1 4 LC ; FFh ; 07h ; length in bytes of private comment ; contents of comment b new-length marker ESC NEWLEN YD c comment marker ESC COMMENT LC comment 2 SDE ; Stripe Data Entry (Within the frame: LSI support range) PSCD ESC SDNORM/SDRST 1 1 abort BID marker ESC ABORT 1 1 ; FFh ; 04h reserved marker ESC RESERVE 1 1 ; FFh ; 01h REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 28 of 34 ; Protected Stripe Coded Data = byte stuffed SCD (Stripe Code Data) ; FFh ; normal terminate (02h) ; /reset "state" for next SDE (03h) M65762FP [Appendix A.2] JBIG Probability Estimation Table ST LSZ NLPS NMPS SWTCH ST LSZ NLPS NMPS 0 0x5ald 1 1 1 57 0x01a4 55 58 SWTCH 0 1 0x2586 14 2 0 58 0x0160 56 59 0 2 0x1114 16 3 0 59 0x0125 57 60 0 3 0x080b 18 4 0 60 0x00f6 58 61 0 4 0x03d8 20 5 0 61 0x00cb 59 62 0 5 0x01da 23 6 0 62 0x00ab 61 63 0 6 0x00e5 25 7 0 63 0x008f 61 32 0 7 0x006f 28 8 0 64 0x5b12 65 65 1 8 0x0036 30 9 0 65 0x4d04 80 66 0 9 0x001a 33 10 0 66 0x412c 81 67 0 10 0x000d 35 11 0 67 0x37d8 82 68 0 11 0x0006 9 12 0 68 0x2fe8 83 69 0 0 12 0x0003 10 13 0 69 0x293c 84 70 13 0x0001 12 13 0 70 0x2379 86 71 0 14 0x5a7f 15 15 1 71 0x1edf 87 72 0 15 0x3f25 36 16 0 72 0x1aa9 87 73 0 16 0x2cf2 38 17 0 73 0x174e 72 74 0 17 0x207c 39 18 0 74 0x1424 72 75 0 0 18 0x17b9 40 19 0 75 0x119c 74 76 19 0x1182 42 20 0 76 0x0f6b 74 77 0 20 0x0cef 43 21 0 77 0x0d51 75 78 0 21 0x09a1 45 22 0 78 0x0bb6 77 79 0 22 0x072f 46 23 0 79 0x0a40 77 48 0 23 0x055c 48 24 0 80 0x5832 80 81 1 24 0x0406 49 25 0 81 0x4d1c 88 82 0 25 0x0303 51 26 0 82 0x438e 89 83 0 26 0x0240 52 27 0 83 0x3bdd 90 84 0 27 0x01b1 54 28 0 84 0x34ee 91 85 0 28 0x0144 56 29 0 85 0x2eae 92 86 0 29 0x00f5 57 30 0 86 0x299a 93 87 0 30 0x00b7 59 31 0 87 0x2516 86 71 0 31 0x008a 60 32 0 88 0x5570 88 89 1 32 0x0068 62 33 0 89 0x4ca9 95 90 0 33 0x004e 63 34 0 90 0x44d9 96 91 0 34 0x003b 32 35 0 91 0x3e22 97 92 0 35 0x002c 33 9 0 92 0x3824 99 93 0 36 0x5ae1 37 37 1 93 0x32b4 99 94 0 37 0x484c 64 38 0 94 0x2e17 93 86 0 38 0x3a0d 65 39 0 95 0x56a8 95 96 1 39 0x2ef1 67 40 0 96 0x4f46 101 97 0 40 0x261f 68 41 0 97 0x47e5 102 98 0 41 0x1f33 69 42 0 98 0x41cf 103 99 0 42 0x19a8 70 43 0 99 0x3c3d 104 100 0 43 0x1518 72 44 0 100 0x375e 99 93 0 44 0x1177 73 45 0 101 0x5231 105 102 0 45 0x0e74 74 46 0 102 0x4c0f 106 103 0 46 0x0bfb 75 47 0 103 0x4639 107 104 0 47 0x09f8 77 48 0 104 0x415e 103 99 0 48 0x0861 78 49 0 105 0x5627 105 106 1 49 0x0706 79 50 0 106 0x50e7 108 107 0 50 0x05cd 48 51 0 107 0x4b85 109 103 0 51 0x04de 50 52 0 108 0x5597 110 109 0 52 0x040f 50 53 0 109 0x504f 111 107 0 53 0x0363 51 54 0 110 0x5a10 110 111 1 54 0x02d4 52 55 0 111 0x5522 112 109 0 55 0x025c 53 56 0 112 0x59eb 112 111 1 56 0x01f8 54 57 0 REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 29 of 34 M65762FP [Appendix B] Timing Characteristics (Conditions: VDD = 5 V ± 5%, C = 50 pF, Ta = 0 to 70°C) 1. Host Bus I/F RESET* t0 CS* t1 t2 t12 t11 A0 to 3 t4 t3 t5 t13 t15 RD* t14 WR* t6 t7 t16 Input Output D0 to 7 t17 2. Code Data I/F CDRQ t20 CDAK* t22 t21 CDRD* t32 t31 t24 t34 CDWR* t27 t26 CD0 to 31 REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 30 of 34 Output t37 t36 Input M65762FP Table B.1 Host Bus I/F Timing Characteristics (Unit: ns) Abbreviation Item Min Timing Conditions Typ Max t0 t1 t2 t3 t4 t5 t6 t7 RESET* assert time CS* setup time to RD* assert CS* hold time to RD* negate A0 to 3 setup time to RD* assert RD* assert time A0 to 3 hold time to RD* negate D0 to 7 output determination time to RD* assert D0 to 7 output hold time to RD* negate 100 15 15 15 20 15 0 0 — — — — — — — — — — — — — — 20 20 t11 t12 t13 t14 t15 t16 t17 CS* setup time to WR* assert CS* hold time to WR* negate A0 to 3 setup time to WR* assert WR* assert time A0 to 3 hold time to WR* negate D0 to 7 input setup time to WR* negate D0 to 7 input hold time to WR* negate 15 15 15 15 15 20 5 — — — — — — — — — — — — — — Table B.2 Timing Characteristics of Code Data Bus I/F Timing Conditions Abbreviation Item t20 t21 t22 t24 t26 t27 CDRQ negate time to CDAK* assert CDAK* setup time to CDRD* assert CDAK* hold time to CDRD* negate CDRD* assert time CD0 to 31 output determination time to CDRD* assert CD0 to 31 output hold time to CDRD* negate Min — 15 15 20 0 0 Typ — — — — — — Max 15 — — — 20 20 t31 t32 t34 t36 t37 CDAK* setup time to CDWR* assert CDAK* hold time to CDWR* negate CDWR* assert time CD0 to 31 input setup time to CDWR* negate CD0 to 31 input hold time to CDWR* negate 15 15 15 15 5 — — — — — — — — — — REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 31 of 34 M65762FP 3. Image Data I/F (1) Serial image data I/F PRDY* t40 t45 PTIM* t41 t44 t43 t42 PXCK* t47 t46 PXCKO* t49 t50 t51 RVID* 1 t56 SVID* t48 2 3 N t57 1 2 3 N (2) Parallel image data I/F PDRQ t60 PDAK* t61 t62 PDRD* t71 t72 t64 t74 PDWR* t66 Output PD0 to 31 Input 4. Master Clock Input Frequency (LSI Operating Frequency) Mx Mh MCLK Ml REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 32 of 34 t77 t76 t67 M65762FP Table B.3 Timing Characteristics of Image Data I/F (Unit: ns) Abbreviation Item Min Timing Conditions Typ Max t40 t41 t42 t43 t44 t45 t46 t47 t48 t49 PRDY* negate time to PTIM* assert PTIM* setup time to PXCK* fall PTIM* hold time to PXCK* rise PXCK* high time PXCK* low time PXCK* cycle RVID* output determination time to PXCK* fall RVID* output change time to PXCK* fall RVID* negate time to PTIM* negate PXCKO* delay time to PXCK* — 15 15 10 10 25 — — 0 — — — — — — — — — — — 20 — — — — — 20 20 — 10 t50 t51 RVID* output determination time to PXCKO* fall RVID* output change time to PXCKO* fall — — — — 12 12 t56 t57 SVID* setup time to PXCK* rise SVID* hold time to PXCK* rise 10 10 — — — — t60 t61 t62 t64 t66 t67 PDRQ negate time to PDAK* assert PDAK* setup time to PDRD* assert PDAK* hold time to PDRD* negate PDRD* assert time PD0 to 31 output determination time to PDRD* assert PD0 to 31 output hold time to PDRD* negate — 15 15 20 0 0 — — — — — — 15 — — — 20 20 t71 t72 t74 t76 t77 PDAK* setup time to PDWR* assert PDAK* hold time to PDWR* negate PDWR* assert time PD0 to 31 input setup time to PDWR* negate PD0 to 31 input hold time to PDWR* negate 15 15 15 15 5 — — — — — — — — — — Table B.4 Master Clock Frequencies (Unit: ns) Timing Conditions Item MCLK cycle (Mx) MCLK high level time (Mh) MCLK low level time (Ml) REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 33 of 34 Min 25 10 10 Typ — — — Max — — — Max Frequency 40 MHz M65762FP Package Dimensions JEITA Package Code P-LQFP144-20x20-0.50 RENESAS Code PLQP0144KA-A Previous Code 144P6Q-A / FP-144L / FP-144LV MASS[Typ.] 1.2g HD *1 D 108 73 109 NOTE) 1. DIMENSIONS "*1" AND "*2" DO NOT INCLUDE MOLD FLASH. 2. DIMENSION "*3" DOES NOT INCLUDE TRIM OFFSET. 72 bp c Reference Symbol *2 E HE c1 b1 36 A 1 ZD Index mark c 37 A2 144 ZE Terminal cross section A1 F L D E A2 HD HE A A1 bp b1 c c1 L1 *3 e y bp x REJ03F0235-0200 Rev.2.00 Sep 14, 2007 Page 34 of 34 Detail F e x y ZD ZE L L1 Dimension in Millimeters Min Nom Max 19.9 20.0 20.1 19.9 20.0 20.1 1.4 21.8 22.0 22.2 21.8 22.0 22.2 1.7 0.05 0.1 0.15 0.17 0.22 0.27 0.20 0.09 0.145 0.20 0.125 0° 8° 0.5 0.08 0.10 1.25 1.25 0.35 0.5 0.65 1.0 Sales Strategic Planning Div. Nippon Bldg., 2-6-2, Ohte-machi, Chiyoda-ku, Tokyo 100-0004, Japan Notes: 1. This document is provided for reference purposes only so that Renesas customers may select the appropriate Renesas products for their use. Renesas neither makes warranties or representations with respect to the accuracy or completeness of the information contained in this document nor grants any license to any intellectual property rights or any other rights of Renesas or any third party with respect to the information in this document. 2. Renesas shall have no liability for damages or infringement of any intellectual property or other rights arising out of the use of any information in this document, including, but not limited to, product data, diagrams, charts, programs, algorithms, and application circuit examples. 3. You should not use the products or the technology described in this document for the purpose of military applications such as the development of weapons of mass destruction or for the purpose of any other military use. When exporting the products or technology described herein, you should follow the applicable export control laws and regulations, and procedures required by such laws and regulations. 4. All information included in this document such as product data, diagrams, charts, programs, algorithms, and application circuit examples, is current as of the date this document is issued. Such information, however, is subject to change without any prior notice. Before purchasing or using any Renesas products listed in this document, please confirm the latest product information with a Renesas sales office. Also, please pay regular and careful attention to additional and different information to be disclosed by Renesas such as that disclosed through our website. (http://www.renesas.com ) 5. Renesas has used reasonable care in compiling the information included in this document, but Renesas assumes no liability whatsoever for any damages incurred as a result of errors or omissions in the information included in this document. 6. When using or otherwise relying on the information in this document, you should evaluate the information in light of the total system before deciding about the applicability of such information to the intended application. Renesas makes no representations, warranties or guaranties regarding the suitability of its products for any particular application and specifically disclaims any liability arising out of the application and use of the information in this document or Renesas products. 7. With the exception of products specified by Renesas as suitable for automobile applications, Renesas products are not designed, manufactured or tested for applications or otherwise in systems the failure or malfunction of which may cause a direct threat to human life or create a risk of human injury or which require especially high quality and reliability such as safety systems, or equipment or systems for transportation and traffic, healthcare, combustion control, aerospace and aeronautics, nuclear power, or undersea communication transmission. If you are considering the use of our products for such purposes, please contact a Renesas sales office beforehand. Renesas shall have no liability for damages arising out of the uses set forth above. 8. Notwithstanding the preceding paragraph, you should not use Renesas products for the purposes listed below: (1) artificial life support devices or systems (2) surgical implantations (3) healthcare intervention (e.g., excision, administration of medication, etc.) (4) any other purposes that pose a direct threat to human life Renesas shall have no liability for damages arising out of the uses set forth in the above and purchasers who elect to use Renesas products in any of the foregoing applications shall indemnify and hold harmless Renesas Technology Corp., its affiliated companies and their officers, directors, and employees against any and all damages arising out of such applications. 9. You should use the products described herein within the range specified by Renesas, especially with respect to the maximum rating, operating supply voltage range, movement power voltage range, heat radiation characteristics, installation and other product characteristics. Renesas shall have no liability for malfunctions or damages arising out of the use of Renesas products beyond such specified ranges. 10. Although Renesas endeavors to improve the quality and reliability of its products, IC products have specific characteristics such as the occurrence of failure at a certain rate and malfunctions under certain use conditions. Please be sure to implement safety measures to guard against the possibility of physical injury, and injury or damage caused by fire in the event of the failure of a Renesas product, such as safety design for hardware and software including but not limited to redundancy, fire control and malfunction prevention, appropriate treatment for aging degradation or any other applicable measures. Among others, since the evaluation of microcomputer software alone is very difficult, please evaluate the safety of the final products or system manufactured by you. 11. In case Renesas products listed in this document are detached from the products to which the Renesas products are attached or affixed, the risk of accident such as swallowing by infants and small children is very high. You should implement safety measures so that Renesas products may not be easily detached from your products. Renesas shall have no liability for damages arising out of such detachment. 12. This document may not be reproduced or duplicated, in any form, in whole or in part, without prior written approval from Renesas. 13. Please contact a Renesas sales office if you have any questions regarding the information contained in this document, Renesas semiconductor products, or if you have any other inquiries. http://www.renesas.com RENESAS SALES OFFICES Refer to "http://www.renesas.com/en/network" for the latest and detailed information. Renesas Technology America, Inc. 450 Holger Way, San Jose, CA 95134-1368, U.S.A Tel: <1> (408) 382-7500, Fax: <1> (408) 382-7501 Renesas Technology Europe Limited Dukes Meadow, Millboard Road, Bourne End, Buckinghamshire, SL8 5FH, U.K. Tel: <44> (1628) 585-100, Fax: <44> (1628) 585-900 Renesas Technology (Shanghai) Co., Ltd. Unit 204, 205, AZIACenter, No.1233 Lujiazui Ring Rd, Pudong District, Shanghai, China 200120 Tel: <86> (21) 5877-1818, Fax: <86> (21) 6887-7898 Renesas Technology Hong Kong Ltd. 7th Floor, North Tower, World Finance Centre, Harbour City, 1 Canton Road, Tsimshatsui, Kowloon, Hong Kong Tel: <852> 2265-6688, Fax: <852> 2730-6071 Renesas Technology Taiwan Co., Ltd. 10th Floor, No.99, Fushing North Road, Taipei, Taiwan Tel: <886> (2) 2715-2888, Fax: <886> (2) 2713-2999 Renesas Technology Singapore Pte. Ltd. 1 Harbour Front Avenue, #06-10, Keppel Bay Tower, Singapore 098632 Tel: <65> 6213-0200, Fax: <65> 6278-8001 Renesas Technology Korea Co., Ltd. Kukje Center Bldg. 18th Fl., 191, 2-ka, Hangang-ro, Yongsan-ku, Seoul 140-702, Korea Tel: <82> (2) 796-3115, Fax: <82> (2) 796-2145 Renesas Technology Malaysia Sdn. Bhd Unit 906, Block B, Menara Amcorp, Amcorp Trade Centre, No.18, Jalan Persiaran Barat, 46050 Petaling Jaya, Selangor Darul Ehsan, Malaysia Tel: <603> 7955-9390, Fax: <603> 7955-9510 © 2007. Renesas Technology Corp., All rights reserved. Printed in Japan. Colophon .7.0