Technical Documentation - Dungeon Master II Graphics.dat: How to read and decompress images from the PC version
Author of this documentation: Kentaro.k-21
This is a binary dump of DM2 PC specific image to be described.
ADDRESS 00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F 0123456789ABCDEF
------------------------------------------------------------------------
0000000 0C 00 07 7C 00 00 03 00 BD 0A 22 00 27 2A 2A 27 ...|......".'**'
0000010 44 01 27 EF 2A 2D 2D 2A 44 01 2A 3B 3F CF 3F 3B D.'.*--*D.*;?.?;
0000020 2A 27 41 01 C0 02 2D 3B 9F 3F 2D 27 00 00 01 03 *'A...-;.?-'....
0000030 A0 01 2D 10 41 06 C3 01 A5 07 81 03 27 00 09 ..-.A.......'..
It is 1620th binary (0 based index) in DM2 PC GRAPHICS.DAT, and is the compressed image of Blue gem.
The DM2 PC specific compressed image has 8 bytes for prefix and sequent bytes for body of compressed image. It has no suffix.
| Part | Heading hexa binary |
|---|---|
| Prefix | 0C 00 07 7C 00 00 03 00 |
| Body | BD 0A 22 00 27 2A 2A 27 44 01 27 EF 2A 2D 2D 2A 44... |
The 4th byte in Prefix is highlighted. It may be used to know that this image is formatted in DM2 PC specific image.
Majority images has 4th byte fixed in 0x7C.
Body binary format
There are command control byte and sequent 8 primitives in a set.
| Command control byte |
| If 1st bit is set in command control byte, here is a pixel(1 byte), otherwise here is repetition command(2 bytes) |
| If 2st bit is set in command control byte, here is a pixel(1 byte), otherwise here is repetition command(2 bytes) |
| If 3st bit is set in command control byte, here is a pixel(1 byte), otherwise here is repetition command(2 bytes) |
| If 4st bit is set in command control byte, here is a pixel(1 byte), otherwise here is repetition command(2 bytes) |
| If 5st bit is set in command control byte, here is a pixel(1 byte), otherwise here is repetition command(2 bytes) |
| If 6st bit is set in command control byte, here is a pixel(1 byte), otherwise here is repetition command(2 bytes) |
| If 7st bit is set in command control byte, here is a pixel(1 byte), otherwise here is repetition command(2 bytes) |
| If 8st bit is set in command control byte, here is a pixel(1 byte), otherwise here is repetition command(2 bytes) |
| Command control byte |
| ... |
* 1st bit means LSB in a byte. 8th bit means MSB in a byte.
Format of Pixel
A pixel is represented in 1 byte and is a 8-bit indexed palette in DM2.
It renders only 1 pixel per pixel command.
Format of Repetition Command
This command is to copy pixels from previous rendered pixels. It has 2 following scalar factors
- How far do you go back to fetch pixels (As negative-offset)
- How length do you copy (As length)
It is always consisted with 2 bytes.
First byte is variant against each image prefix. They will be grouped in 'Type'.
| Type | Format of 1st byte (from MSB to LSB) |
|---|---|
| Type 2 | XXXXyyyy (leading 4 bits for X, sequent 4 bits for Y) |
| Type 3 | XXXyyyyy (leading 3 bits for X, sequent 5 bits for Y) |
X and Y are defined for expression.
Second byte is a unsigned byte value and defined as B for expression.
negative_offset and copy_length are figured out as next table shows.
| Type | Expression |
|---|---|
| Type 2 | negative_offset = X + (16 * B) copy_length = Y + 3 |
| Type 3 | negative_offset = X + (8 * B) copy_length = Y + 3 |
* It is possible that source pixel range and destination pixel range between a repetition command are overlapped. It says (negative_offset < copy_length) is actually possible.
Decision of Type
This is prefix part of the image: 0C 00 07 7C 00 00 03 00
The 7th byte, the highlighted part shows the Type of decompression factor.
If it is 02, it is Type2 factor applied compressed image.
If it is 03, it is Type3 factor applied compressed image.
Decompression Example
Here is reading sample for the 1620th compressed image.
This is Type3 factor applied compressed image.
| Index in c.c.b. | Binary | Reading |
|---|---|---|
| BD | Command control byte.bit array is 1011 1101 | |
| 1st | 0A | Pixel 0A |
| 2nd | 22 00 | Repetition command. negative_offset is (1 + 8 * 0) = 1 copy_length is (3 + 2) = 5 Pixels are 0A 0A 0A 0A 0A |
| 3rd | 27 | Pixel 27 |
| 4th | 2A | Pixel 2A |
| 5th | 2A | Pixel 2A |
| 6th | 27 | Pixel 27 |
| 7th | 44 01 | Repetition command. negative_offset is (2 + 8 * 1) = 10 copy_length is (3 + 4) = 7 Pixels are 0A 0A 0A 0A 0A 0A 27 |
| 8th | 27 | Pixel 27 |
| EF | Command control byte.bit array is 1110 1111 | |
| 1st | 2A | Pixel 2A |
| 2nd | 2D | Pixel 2D |
| 3rd | 2D | Pixel 2D |
| 4th | 2A | Pixel 2A |
| 5th | 44 01 | Repetition command. negative_offset is (2 + 8 * 1) = 10 copy_length is (3 + 4) = 7 Pixels are 0A 0A 0A 0A 27 27 2A |
| 6th | 2A | Pixel 2A |
| 7th | 3B | Pixel 3B |
| 8th | 3F | Pixel 3F |
| CF | Command control byte.bit array is 1100 1111 | |
| 1st | 3F | Pixel 3F |
| 2nd | 3B | Pixel 3B |
| 3rd | 2A | Pixel 2A |
| 4th | 27 | Pixel 27 |
| 5th | 41 01 | Repetition command. negative_offset is (2 + 8 * 1) = 10 copy_length is (3 + 1) = 4 Pixels are 27 27 2A 2A |
| 6th | C0 02 | Repetition command. negative_offset is (6 + 8 * 2) = 22 copy_length is (3 + 0) = 3 Pixels are 2A 2D 2D |
| 7th | 2D | Pixel 2D |
| 8th | 3B | Pixel 3B |
| 9F | Command control byte.bit array is 1001 1111 | |
| 1st | 3F | Pixel 3F |
| 2nd | 2D | Pixel 2D |
| 3rd | 27 | Pixel 27 |
| 4th | 00 | Pixel 00 |
| 5th | 00 | Pixel 00 |
| 6th | 01 03 | Repetition command. negative_offset is (0 + 8 * 3) = 24 copy_length is (3 + 1) = 4 Pixels are 27 27 2A 2A |
| 7th | A0 01 | Repetition command. negative_offset is (5 + 8 * 1) = 13 copy_length is (3 + 0) = 3 Pixels are 2D 2D 2D |
| 8th | 2D | Pixel 2D |
| 10 | Command control byte.bit array is 0001 0000 | |
| 1st | 41 06 | Repetition command. negative_offset is (2 + 8 * 6) = 50 copy_length is (3 + 1) = 4 Pixels are 2A 27 0A 0A |
| 2nd | C3 01 | Repetition command. negative_offset is (6 + 8 * 1) = 14 copy_length is (3 + 3) = 6 Pixels are 00 00 27 27 2A 2A |
| 3rd | A5 07 | Repetition command. negative_offset is (5 + 8 * 7) = 61 copy_length is (3 + 5) = 8 Pixels are 2A 2A 27 0A 0A 0A 0A 0A |
| 4th | 81 03 | Repetition command. negative_offset is (4 + 8 * 3) = 28 copy_length is (3 + 1) = 4 Pixels are 00 00 27 27 |
| 5th | 27 | Pixel 27 |
| 6th | 00 09 | Repetition command. negative_offset is (0 + 8 * 9) = 72 copy_length is (3 + 0) = 3 Pixels are 27 0A 0A |