FXY File Format: Difference between revisions

From Bohemia Interactive Community
Jump to navigation Jump to search
(8 intermediate revisions by 4 users not shown)
Line 52: Line 52:
   ushort  CharCode;      // '''0x21''' = 'A' <<< see note
   ushort  CharCode;      // '''0x21''' = 'A' <<< see note
   ushort  PaaFileNumber; // 01
   ushort  PaaFileNumber; // 01
   ushort  X,Y            // Top-left Corner of char on the texture in pixels
   ushort  X,Y            // Offset (in pixels) in the paa file, to the Top-left of this glyph.
   ushort  Width,Height;  // Size of the char on the texture in pixels
   ushort  Width,Height;  // area (number of) pixels used from the paa file for this glyph.
   ushort  KerningWidth;  // ARMA only. Width used to determine this character spacing
   ushort  KerningWidth;  // ARMA only. Width used to determine this character spacing
  };
  };


Apart from the ARMA header (in Arma fxy files only), the fxy is a contiguous series of 12 byte (ofp) or 14 byte (Arma) entries.  
Apart from the ARMA header (in Arma fxy files only), the fxy is a contiguous series of 12 byte (ofp) or 14 byte (Arma) entries.


====CharCode====
====CharCode====
OFP uses a USASCII character set. Arma, uses Unicode.
OFP uses a USASCII character set. Arma uses Unicode.


*For OFP, there are 224 entries. This because the first 32 indexes of (most) Ascii codepages are unprintable, which leaves 256-32= 224.  
The first 256 unicodes ARE us-ascii. For this reason, the first 32 (x20) glyphs in EITHER character set, are control characters, never intended for display, and not included in the paa file(s) nor the fxy header. For that reason, in all circumstances, *the* displayed character is charcode+x20. Thus


*ArmA fxy's have 464 entries, as each font supports both Western European, Eastern European and Cyrillic character sets.
'Actual Character' = charcode+x20;


:The Arma Unicode Codepage is broken up as follows
Charcode x21 = index-entry =x21 = character x20+x21 = 0x41 = Glyph 'A'.


*  0...  FF 'International' Codepage equivalent to Windows(tm) ANSI 1252
This is a constant, irrespective of the number of fxy tables (fonts) now or in the future. Charcode x21 can only ever be glyph 'A' (rendered in whatever style and size the font wants to render it)
* 100... 15F Alt Language 1
* 3E0... 3FF Alt Language 2
* 400... 43F Alt Language 3
*1FF0...1FFF Special chars


Note that while the Charcode is of technical interest, it is redundant. There is a direct correspondence between the index value (eg the entry number) and the character it represents. Thus for Ascii, index entry 0x41-x20 always represents the glyph character 'A'
*For OFP, there are 224 entries in any fxy table. (which therefore covers the 256 (224+32) possible us ascii 'characters'.
Ascii = index + 0x20; // and
index = Charcode; // (ignoring Unicode alt languages)


Technically, the engines probably accommodate lack of characters (eg no lower case supplied), and even alternate mapping (index 0x21 != Charcode 0x21) but neither of these arrangements exist in any BI fxy file.
*ArmA fxy's have 448 character entries + 16 specials, as each font supports both Western European, Eastern European and Cyrillic character sets.
 
: Bearing in mind that basic 16 bit Unicode can display 65,000 glyphs, standard Arma fonts only use (and are only capable of displaying) a small selection:
 
*  x20...17F Us Ascii + (most) European Languages
* 400...45F Cyrillic
*2010...201F Special chars. unknown how they work.
 
You could extend this range in custom fonts of course to whatever you need from 16bit Unicode.
 
Note that index entries happen to be linear sequential. The nth index entry is also CharCode n. Which way, the engine actually accesses this table for character 'n' is up to the engine.


====PaaFileNumber====  
====PaaFileNumber====  
Line 88: Line 92:
  First texture file is "CourierNewB64-01.paa",  
  First texture file is "CourierNewB64-01.paa",  
  second is "CourierNewB64-02.paa".
  second is "CourierNewB64-02.paa".
*Note that you cannot assume the glyph character 'A' (nor any other glyph) is ordered the same way in any given paa font-set, nor, in fact that the glyph will 'always be' in fontfile 1 (eg)
Order of glyphs is entirely random in the sense that binarising these files uses a 'best fit' method to squeeze as much content into that file. Not perfect, but not badly done either.


==Usage==
==Usage==
Line 161: Line 169:


===Arma2===
===Arma2===
  TBD...
 
[[Image:ArmA2OAFonts CO.png|thumb|280px|right|ArmA2's fonts]]
LucidaConsoleB
Zeppelin33
Zeppelin33Italic
  Zeppelin32
EtelkaNarrowMediumPro
Bitstream
TahomaB
EtelkaMonospaceProBold
 
===Arma3===
 
[[Image:Arma3Fonts.png|thumb|280px|right|ArmA3's fonts]]
EtelkaMonospacePro
EtelkaMonospaceProBold
EtelkaNarrowMediumPro
LucidaConsoleB
PuristaBold
PuristaLight
PuristaMedium
PuristaSemiBold
RobotoCondensed
RobotoCondensedBold
RobotoCondensedLight
TahomaB
 


[[category:Operation Flashpoint: Addons]]
[[category:Operation Flashpoint: Addons]]

Revision as of 18:43, 8 June 2017

Template:unsupported-doc

Introduction

PAA layout of a font

The OFP and ArmA engines use a limited set of bit mapped fonts developed by Bohemia Interactive for the engine. They are bitmaps and each glyph within is treated as a texture.

A character set (such as Ascii) is a series of 'glyphs' (characters) where any one of these characters is always identified by the same index value. For example the letter A is always hex 41 and so on. Regardless of the font set employed, the character set within it, is set-in-concrete. Hex 41 always renders a glyph that 'looks like' the letter A. Be it bold, italic, or fills the entire screen, it is the letter 'A'.

The character set for all BI fonts in OFP is different for each language version. For Western languages it approximates US-Ascii codepage 437. In ArmA Unicode is used for fonts.

A font set is a collection of displayable characters (termed glyphs) that have a common appearance. Eg all glyphs in THIS set are bold, or all the glyphs are italic, all the glyphs are Times Roman (seriffed).

Individual font sets exist for Tahoma, Garamond, and so on.

The glyphs within these sets, are contained in .paa files. Generally, the entire font set = one, paa file.

And, because these are bitmapped fonts, eg fixed dimensions, non-scaleable, there are unique font sets, for, unique sizes.

For example, in Resistance engine, the popular Tahoma font is in three different font sets. 24 point, 36 point, 48 point.

A 'point' approximates 1/72th of an inch. Thus the Tahoma fonts available (for Resistance) are approximately a third, a half, and two thirds of an inch tall when displayed.

Description

An fxy file is the 'character set' index. It is, the codepage.

An fxy file is implicitly associated with one or more paa files within the same folder. The fxy file is the header entry for these files, it describes the position in the paa file and the size to be rendered on screen, for a 'glyph'.

Association of files between fxy and paa is implicit. The CourierNewB64.fxy is implicitly associated with the CourierNewB64-xx.paa file(s). xx indicating one of several CourierNewB64 paa's.

Legend

see Generic FileFormat Data Types

Struct

Fxy
{
  bytes     ArmaSignature[8]; // ARMA ONLY
  FxyEntry  Entry1;
         ....
  FxyEntry  EntryLast;
}

ArmaSignature

ArmaSignature
{
  char Magic[4]; //"BIFo"
  ulong Version; // 0x101
}

FxyEntry

FxyEntry 
{
  ushort  CharCode;      // 0x21 = 'A' <<< see note
  ushort  PaaFileNumber; // 01
  ushort  X,Y            // Offset (in pixels) in the paa file, to the Top-left of this glyph.
  ushort  Width,Height;  // area (number of) pixels used from the paa file for this glyph.
  ushort  KerningWidth;  // ARMA only. Width used to determine this character spacing
};

Apart from the ARMA header (in Arma fxy files only), the fxy is a contiguous series of 12 byte (ofp) or 14 byte (Arma) entries.

CharCode

OFP uses a USASCII character set. Arma uses Unicode.

The first 256 unicodes ARE us-ascii. For this reason, the first 32 (x20) glyphs in EITHER character set, are control characters, never intended for display, and not included in the paa file(s) nor the fxy header. For that reason, in all circumstances, *the* displayed character is charcode+x20. Thus 

'Actual Character' = charcode+x20;

Charcode x21 = index-entry =x21 = character x20+x21 = 0x41 = Glyph 'A'.

This is a constant, irrespective of the number of fxy tables (fonts) now or in the future. Charcode x21 can only ever be glyph 'A' (rendered in whatever style and size the font wants to render it)

  • For OFP, there are 224 entries in any fxy table. (which therefore covers the 256 (224+32) possible us ascii 'characters'.
  • ArmA fxy's have 448 character entries + 16 specials, as each font supports both Western European, Eastern European and Cyrillic character sets.
Bearing in mind that basic 16 bit Unicode can display 65,000 glyphs, standard Arma fonts only use (and are only capable of displaying) a small selection:
  • x20...17F Us Ascii + (most) European Languages
  • 400...45F Cyrillic
  • 2010...201F Special chars. unknown how they work.

You could extend this range in custom fonts of course to whatever you need from 16bit Unicode.

Note that index entries happen to be linear sequential. The nth index entry is also CharCode n. Which way, the engine actually accesses this table for character 'n' is up to the engine.

PaaFileNumber

-number of texture file: This is simply a convenience to allow large fonts to be spread across several paa's. 

font name is "CourierNewB64" 
Fxy file is "CourierNewB64.fxy", 
First texture file is "CourierNewB64-01.paa", 
second is "CourierNewB64-02.paa".
  • Note that you cannot assume the glyph character 'A' (nor any other glyph) is ordered the same way in any given paa font-set, nor, in fact that the glyph will 'always be' in fontfile 1 (eg)

Order of glyphs is entirely random in the sense that binarising these files uses a 'best fit' method to squeeze as much content into that file. Not perfect, but not badly done either.

Usage

This example using the new font name in a ArmA description.ext. 

class RscText
{
       type = CT_STATIC;
       idc = -1;
       style = ST_LEFT;
       colorBackground[] = {0, 0, 0, 0};
       colorText[] = {1, 1, 1, 1};
       font = Bitstream;
       sizeEx = 0.04;
};

Paa Content

Resistance

.paa files used for the font must be in 8080 format: (Alpha luminosity) 

8bit - Brightness (all with 255 value - "white");
8bit - Alpha (255 for character, 0 for space);

ArmA

Arma Glyph files (paa) are formatted in DXT5 compressed textures. Using the TAGG labelling, they must additionally specify 

Avg Color: any
Max Color: always FFFFFFFF
AlphaFlag: 1

Available Fonts

Resistance and CWC

AudreysHand 48pt Italic
AudreysHand 48pt Bold
CourierNew Bold 64pt
Garamond 64
Steelfish bold 64 and 128 pt
SteelfishCE bold 64 
Tahoma 24/36/48 bold

Elite

Arial Bold 7,8,10,11,12 pt
Arial Italic 9,13,18,48
CourierNew 9pt
Hel67-CM 10,11 pt
HelvCondLight 8,9,10 pt
Helvetica-Narrow 11pt
Helvetica-Narrow Bold 8pt
Helvetica37-CondensedThin bold 10pt
Helvetica57-Condensed 10,13,16pt
Helvetica57-Condensed 9 ,13,14 pt Bold 
Helvetica67-Condensed 10,11,9pt Medium
HelveticaNarrow 10,11pt
HelveticaNarrow bold 10,13,18,22
HelveticaNeue 10,11,12,13,14,22
HelveticaNeue  Bold 22 B22-01.paa
LucidaConsole Bold 11,8
MSS Bold 9
System 10pt
System 10pt Bold
Tahoma 10,11,12,14,16,20,21pt
Tahoma 5,6,7,9,10,11,12,13,14,16,20,21 pt bold
Tahoma 12pt Bold Italic

ArmA

Arial Bold 7->24 pt
BitStream
Versans bold 7->20, 22, 28 pt
Zeppelin32 12->20,22,28 pt
Zeppelin33 12->20,22,28 pt
Zeppelin33 12->20,22,28 pt italic

Arma2

ArmA2's fonts
LucidaConsoleB
Zeppelin33
Zeppelin33Italic
Zeppelin32
EtelkaNarrowMediumPro
Bitstream
TahomaB
EtelkaMonospaceProBold

Arma3

ArmA3's fonts
EtelkaMonospacePro
EtelkaMonospaceProBold
EtelkaNarrowMediumPro
LucidaConsoleB
PuristaBold
PuristaLight
PuristaMedium
PuristaSemiBold
RobotoCondensed
RobotoCondensedBold 
RobotoCondensedLight
TahomaB
Retrieved from "https://community.bistudio.com/wiki?title=FXY_File_Format&oldid=103398"