Format of a GRF File
A GRF file is a collection of "sprites", defining pretty much everything
that you can see in the game. The sprites can be graphics, but also define
many non-visual properties of the game entities like vehicles.
There are two different file formats defined. The older format 1 simply
consists of one sprite following another until the end of the file,
with no header or other meta-information, except for a checksum at the
very end of the file.
Format 2 consists of a header, followed by a data section containing all
non-image sprites, followed by a sprite section containing the actual
graphics.
All numbers in a GRF file are stored in little-endian byte order.
File header (format 2 only)
A format 2 GRF starts with a header with the following bytes:
00 00 47 52 46 82 0D 0A 1A 0A
DWORD sprite_offs
Number of bytes to skip after this DWORD to reach the start of
the sprite section.
BYTE data_compr
Compression format of the data section.
00 No compression
others Not defined
Data section (formats 1 and 2)
The data section consists of sprites with the following format:
WORD size (format 1)
DWORD size (format 2)
The size of the current sprite, not including itself. The meaning
depends on the info byte that follows.
BYTE info
Bitcoded value that determines what type this sprite is.
FF Pseudo sprite that contains non-image data, see the
NewGraphicsSpecs for details. For a pseudo sprite, the
size does *not* include the info byte.
FD (Only valid in format 2) Reference to sprite section. Size does
*not* include the info byte.
others (Only valid in format 1) Normal sprite with image data, see the
section "Normal sprites" for details. If bit 2 is set, size
is the length of the sprite in the file including the info byte.
If it is not set, you have to decode the image data to find out
where the next sprite starts.
DATA data
The length of the data depends on size and info.
If the info value is FD, this is a single DWORD containing an ID
that is looked up in the sprite section and processed instead of
this sprite.
The data sections is termindated by a WORD (format 1) / DWORD (format 2)
of value 0.
For format 1 GRFs, a four byte checksum follows. The algorithm to calculate
it is unknown, but it is never processed by the game anyway.
Sprite section (format 2 only)
The sprite section consists of entries with the following format:
DWORD id
ID used to reference the entry from the data section.
DWORD size
Length of the following data in the file.
BYTE info
Bitcoded value that determines what type this sprite is.
FF Non-image data, for example sounds. See the NewGraphicsSpecs
for details.
others Image data, see the section "Normal sprites" for details.
DATA data
size-1 bytes of data.
Multiple entries with the same ID are allowed. IDs must be ordered
ascending.
The sprite section is terminated by a DWORD of value 0, which means
that 0 is never a valid ID.
Normal sprites
Normal sprites contain the visible graphics of the game. The exact format
depends on the file format. For format 1 files, normal sprites occur in
the data section, while for format 2 files they are found in the sprite
section.
For format 1 GRFs, the info byte has the following meaning:
Bit Val Meaning
0 1 Color index 0 is transparent (should always be set).
1 2 Size is compressed size if set.
If this bit is set, the given size is simply the size in the
file. If it is unset, you *must* decompress it to find out
how large it is in the file.
3 8 Has transparency (i.e. is a tile), see below.
6 40h The exact size of this sprite is significant. If not set,
grfcodec may attempt to remove extraneous transparent pixels.
others Undefined.
For format 2 GRFs, the info byte has the following meaning:
Bit Val Meaning
0 1 Pixel format contains RGB components.
1 2 Pixel format contains alpha component.
2 4 Pixel format contains mask/palette component.
3 8 Has transparency (i.e. is a tile), see below.
6 40h The exact size of this sprite is significant. If not set,
grfcodec may attempt to remove extraneous transparent pixels.
others Undefined.
Pixel components are ordered R, G, B, A, M, with not present components
skipped without padding. For file format 1, pixel format is implicitly
defined to be only the palette component.
BYTE zoom_level (format 2 only)
Zoom level of the current sprite, the following values are defined:
00 normal zoom
01 4x zoom-in
02 2x zoom-in
03 2x zoom-out
04 4x zoom-out
05 8x zoom-out
BYTE ydim (format 1)
WORD ydim (format 2)
How many lines there are in the sprite (y dimension)
WORD xdim
How many columns there are (x dimension)
WORD xrel
Horizontal offset. The offset is counted from the base
coordinate for each sprite.
WORD yrel
Vertical offset.
DWORD uncomp_size (format 2, only if info bit 3 is set)
Uncompressed size of the tile encoded image data.
After this follows the actual compressed data. If info bit 3 is not
set, the data is simply a stream of bytes from left to right, and
from top to bottom, making up xdim*ydim pixels.
Tile sprites
If info bit 3 is set, the sprite is a tile and has some special
transparency information that is encoded like follows. Each line is
encoded separately and split into "chunks". Each chunk contains
pixels, but the chunks may skip a few pixels which are then
transparent.
The sprite data first starts off with a list of WORD offsets for
format 1 GRFs and format 2 GRFs with an uncompressed size below 65536
bytes. If the size is bigger in a format 2 GRF, the sprite starts
with a list of DWORD offsets. There is one offset for each line. These
determine at which offset each line starts, counted from the first data
byte.
Then follow the chunks for the lines:
BYTE cinfo (format 1)
WORD cinfo (format 2 if width > 256)
The high bit is set if this is the last chunk in the line. The
line need not be filled entirely, any remaining pixels are
simply transparent. The lower seven/fifteen bits give the length of
this chunk in pixels.
BYTE cofs (format 1)
WORD cofs (format 2 if width > 256)
x offset at which this chunk starts. The pixels between this
chunk and the last one will be transparent.
After this follow (cinfo & 0x7f or cinfo & 0x7fff, respectively) bytes
of pixels.
Compression algorithm
The compression used is a variation on the LZ77 algorithm which
detects redundancy and losslessly reduces the size of the data. Here's
how the compressed data looks in a GRF file.
The compressed stream contains either a pointer to an earlier location
and a length, which means that these bytes are copied over from the
given location, or it contains a length and a verbatim chunk which is
copied to the output stream.
BYTE code
The high bit of the code shows whether this is a verbatim chunk
(not set) or a repetition of earlier data (set).
The meaning of the following bytes depends on whether the high bit of
code is set.
If the high bit is not set, what follows is code&0x7f bytes of
verbatim data.
If the high bit is set, the code has a slightly different meaning.
Bits 3 to 7 are now three bits to a length value, stating how much
data should be copied from the earlier location. Bits 0 to 2 are the
high bits of an offset, with the low bits being in the next byte.
BYTE lofs
Low bits of the offset
Use this to extract length and offset:
unsigned long length = -(code >> 3);
unsigned long offset = ( (code & 7) << 8 ) | lofs;
It's important that the variables are unsigned and at least two bytes
large.
The offset is counted backwards from the current location. So you
subtract the offset from your position in the output stream and copy
the given number of bytes.
And that's pretty much all you need to know about a GRF file!
______________________________________________________________________
Copyright © 1999-2012 by Josef Drexler.
