8-bit Raycasting Quake Skies and Animated Textures

Monday, 20th August 2007

All of this Quake and XNA 3D stuff has given me a few ideas for calculator (TI-83) 3D.

One of my problems with calculator 3D apps is that I have never managed to even get a raycaster working. Raycasters aren't exactly very tricky things to write.

So, to help me, I wrote a raycaster in C#, limiting myself to the constraints of the calculator engine - 96×64 display, 256 whole angles in a full revolution, 16×16 map, that sort of thing. This was easy as I had floating-point maths to fall back on.


With that done, I went and ripped out all of the floating-point code and replaced it with fixed-point integer arithmetic; I'm using 16-bit values, 8 bits for the whole part and 8 bits for the fractional part.

From here, I just rewrote all of my C# code in Z80 assembly, chucking in debugging code all the way through so that I could watch the state of values and compare them with the results from my C# code.


The result is rather slow, but on the plus side the code is clean and simple. smile.gif The screen is cropped for three reasons: it's faster to only render 64 columns (naturally), you get some space to put a HUD and - most importantly - it limits the FOV to 90°, as the classic fisheye distortion becomes a more obvious problem above this.


I sneaked a look at the source code of Gemini, an advanced raycaster featuring textured walls, objects and doors. It is much, much faster than my engine, even though it does a lot more!

It appears that the basic raycasting algorithm is pretty much identical to the one I use, but gets away with 8-bit fixed point values. 8-bit operations can be done significantly faster than 16-bit ones on the Z80, especially multiplications and divisions (which need to be implemented in software). You can also keep track of more variables in registers, and restricting the number of memory reads and writes can shave off some precious cycles.

Some ideas that I've had for the raycaster, that I'd like to try and implement:

  • Variable height floors and ceilings. Each block in the world is given a floor and ceiling height. When the ray intersects the boundary, the camera height is subtracted from these values, they are divided by the length of the ray (for projection) and the visible section of the wall is drawn. Two counters would keep track of the upper and lower values currently drawn to to keep track of the last block's extent (for occlusion) and floor/ceiling colours could be filled between blocks.
  • No texturing: wall faces and floors/ceilings would be assigned dithered shades of grey. I think this, combined with lighting effects (flickering, shading), would look better than monochrome texture mapping - and would be faster!
  • Ray-transforming blocks. For example, you could have two 16×16 maps with a tunnel: the tunnel would contain a special block that would, when hit, tell the raycaster to start scanning through a different level. This could be used to stitch together large worlds from small maps (16×16 is a good value as it lets you reduce level pointers to 8-bit values).
  • Adjusting floors and ceilings for lifts or crushing ceilings.

As far as the Quake project, I've made a little progress. I've added skybox support for Quake 2:


Quake 2's skyboxes are simply made up of six textures (top, bottom, front, back, left, right). Quake doesn't use a skybox. Firstly, you have two parts of the texture - one half is the sky background, and the other half is a cloud overlay (both layers scroll at different speeds). Secondly, it is warped in a rather interesting fashion - rather like a squashed sphere, reflected in the horizon:


For the moment, I'm just using the Quake 2 box plus a simple pixel shader to mix the two halves of the sky texture.


I daresay something could be worked out to simulate the warping.


The above is from GLQuake, which doesn't really look very convincing at all.


I've reimplemented the texture animation system in the new BSP renderer, including support for Quake 2's animation system (which is much simpler than Quake 1's - rather than have magic texture names, all textures contain the name of the next frame in their animation cycle).

QuakeC VM

Wednesday, 15th August 2007

I've started serious work on the QuakeC virtual machine.


The bytecode is stored in a single file, progs.dat. It is made up of a number of different sections:

  • Definitions data - an unformatted block of data containing a mixture of floating point values, integers and vectors.
  • Statements - individual instructions, each made up of four short integers. Each statement has an operation code and up to three arguments. These arguments are typically pointers into the definitions data block.
  • Functions - these provide a function name, a source file name, storage requirements for local variables and the address of the first statement.

On top of that are two tables that break down the definitions table into global and field variables (as far as I'm aware this is only used to print "nice" names for variables when debugging, as it just attaches a type and name to each definition) and a string table.

The first few values in the definition data table are used for predefined values, such as function parameters and return value storage.

Now, a slight problem is how to handle these variables. My initial solution was to read and write types strictly as particular types using the definitions table, but this idea got scrapped when I realised that the QuakeC bytecode uses the vector store opcode to copy string pointers, and a vector isn't much use when you need to print a string.

I now use a special VariablePointer class that internally stores the pointer inside the definition data block, and provides properties for reading and writing using the different formats.

/// <summary>Defines a variable.</summary>
public class VariablePointer {

	private readonly uint Offset;

	private readonly QuakeC Source;

	private void SetStreamPos() { this.Source.DefinitionsDataReader.BaseStream.Seek(this.Offset, SeekOrigin.Begin); }

	public VariablePointer(QuakeC source, uint offset) {
		this.Source = source;
		this.Offset = offset;

	#region Read/Write Properties

	/// <summary>Gets or sets a floating-point value.</summary>
	public float Float {
		get { this.SetStreamPos(); return this.Source.DefinitionsDataReader.ReadSingle(); }
		set { this.SetStreamPos(); this.Source.DefinitionsDataWriter.Write(value); }

	/// <summary>Gets or sets an integer value.</summary>
	public int Integer {
		get { this.SetStreamPos(); return this.Source.DefinitionsDataReader.ReadInt32(); }
		set { this.SetStreamPos(); this.Source.DefinitionsDataWriter.Write(value); }

	/// <summary>Gets or sets a vector value.</summary>
	public Vector3 Vector {
		get { this.SetStreamPos(); return new Vector3(this.Source.DefinitionsDataReader.BaseStream); }
		set {


	#region Extended Properties

	public bool Boolean {
		get { return this.Float != 0f; }
		set { this.Float = value ? 1f : 0f; }


	#region Read-Only Properties

	/// <summary>Gets a string value.</summary>
	public string String {
		get { return this.Source.GetString((uint)this.Integer);  }

	public Function Function {
		get { return this.Source.Functions[this.Integer]; }


Not too elegant, but it works!

If the offset for a statement is negative in a function, that means that the function being called is an internally-implemented one. The source code for the test application in the screenshot at the top of this entry is as follows:

float testVal;

void() test = {
	dprint("This is a QuakeC VM test...\n");
	testVal = 100;
	dprint(ftos(testVal * 10));
	while (testVal > 0) {
		testVal = testVal - 1;
	dprint("Lift off!");

Both dprint and ftos are internal functions; I use a simple array of delegates to reference them.

There's a huge amount of work to be done here, especially when it comes to entities (not something I've looked at at all). All I can say is that I'm very thankful that the .qc source code is available and the DOS compiler runs happily under Windows - they're going to be handy for testing.

Vista and MIDI

Tuesday, 14th August 2007


I have a Creative Audigy SE sound card, which provides hardware MIDI synthesis. However, under Vista, there was no way (that I could see) to change the default MIDI output device to this card, meaning that all apps were using the software synthesiser instead.

Vista MIDI Fix is a 10-minute application I wrote to let me easily change the default MIDI output device. Applications which use MIDI device 0 still end up with the software synthesiser, unfortunately.

To get the hardware MIDI output device available I needed to install Creative's old XP drivers, and not the new Vista ones from their site. This results in missing CMSS, but other features - such as bass redirection, bass boost, 24-bit/96kHz output and the graphical equaliser - now work.

The Creative mixer either crashes or only displays two volume sliders (master and CD audio), which means that (as far as I can tell) there's no easy way to enable MIDI Reverb and MIDI Chorus.

Quake 2 PVS, Realigned Lightmaps and Colour Lightmaps

Friday, 10th August 2007

Quake 2 stores its visibility lists differently to Quake 1 - as close leaves on the BSP tree will usually share the same visibility information, the lists are grouped into clusters (Quake 1 stored a visibility list for every leaf). Rather than go from the camera's leaf to find all of the other visible leaves directly, you need to use the leaf's cluster index to look up which other clusters are visible, then search through the other leaves to find out which reference that cluster too.

In a nutshell, I now use the visibility cluster information in the BSP to cull large quantities of hidden geometry, which has raised the framerate from 18FPS (base1.bsp) to about 90FPS.


I had a look at the lightmap code again. Some of the lightmaps appeared to be off-centre (most clearly visible when there's a small light bracket on a wall casting a sharp inverted V shadow on the wall underneath it, as the tip of the V drifted to one side). On a whim, I decided that if the size of the lightmap was rounded to the nearest 16 diffuse texture pixels, one could assume that the top-left corner was not at (0,0) but offset by 8 pixels to centre the texture. This is probably utter nonsense, but plugging in the offset results in almost completely smooth lightmaps, like the screenshot above.

2007.08.09.02.jpg   2007.08.09.03.jpg
Before and after - coloured lightmaps.

I quite like Quake 2's colour lightmaps, and I also quite like the chunky look of the software renderer. I've modified the pixel shader for the best of both worlds. I calculate the three components of the final colour individually, taking the brightness value for the colourmap from one of the three channels in the lightmap.

float4 Result = 1;

ColourMapIndex.y = 1 - tex2D(LightMapTextureSampler, vsout.LightMapTextureCoordinate).r;
Result.r = tex2D(ColourMapSampler, ColourMapIndex).r;

ColourMapIndex.y = 1 - tex2D(LightMapTextureSampler, vsout.LightMapTextureCoordinate).g;
Result.g = tex2D(ColourMapSampler, ColourMapIndex).g;

ColourMapIndex.y = 1 - tex2D(LightMapTextureSampler, vsout.LightMapTextureCoordinate).b;
Result.b = tex2D(ColourMapSampler, ColourMapIndex).b;

return Result;
There is no impact on framerate at this stage (the rest of the code is the problem - I'm not even batching by texture at the moment).

Journals need more animated GIFs

Thursday, 9th August 2007

Pixel shaders are fun.

I've implemented support for decoding mip-maps from mip textures (embedded in the BSP) and from WAL files (external).

Now, I know that non-power-of-two textures are naughty. Quake uses a number of them, and when loading textures previously I've just let Direct3D do its thing which has appeared to work well.


However, now that I'm directly populating the entire texture, mip-maps and all, I found that Texture2D.SetData was throwing exceptions when I was attempting to shoe-horn in a non-power-of-two texture. Strange. I hacked together a pair of extensions to the Picture class - GetResized(width, height) which returns a resized picture (nearest-neighbour, naturally) - and GetPowerOfTwo(), which returns a picture scaled up to the next power-of-two size if required.


All textures now load correctly, and I can't help but notice that the strangely distorted textures - which I'd put down to crazy texture coordinates - now render correctly! It turns out that all of the distorted textures were non-power-of-two.

The screenshots above demonstrate that Quake 2 is also handled by the software-rendering simulation. The current effect file for the world is as follows:

uniform extern float4x4 WorldViewProj : WORLDVIEWPROJECTION;

uniform extern float Time;
uniform extern bool Rippling;

uniform extern texture DiffuseTexture;
uniform extern texture LightMapTexture;

uniform extern texture ColourMap;

struct VS_OUTPUT {
	float4 Position : POSITION;
	float2 DiffuseTextureCoordinate : TEXCOORD0;
	float2 LightMapTextureCoordinate : TEXCOORD1;
	float3 SourcePosition: TEXCOORD2;

sampler DiffuseTextureSampler = sampler_state {
	texture = <DiffuseTexture>;
	mipfilter = POINT;

sampler LightMapTextureSampler = sampler_state {
	texture = <LightMapTexture>;
	mipfilter = LINEAR;
	minfilter = LINEAR;
	magfilter = LINEAR;

sampler ColourMapSampler = sampler_state {
	texture = <ColourMap>;
	addressu = CLAMP;
	addressv = CLAMP;

VS_OUTPUT Transform(float4 Position : POSITION0, float2 DiffuseTextureCoordinate : TEXCOORD0, float2 LightMapTextureCoordinate : TEXCOORD1) {


	// Transform the input vertex position:
	Out.Position = mul(Position, WorldViewProj);
	// Copy the other values straight into the output for use in the pixel shader.
	Out.DiffuseTextureCoordinate = DiffuseTextureCoordinate;
	Out.LightMapTextureCoordinate = LightMapTextureCoordinate;
	Out.SourcePosition = Position;

	return Out;

float4 ApplyTexture(VS_OUTPUT vsout) : COLOR {

	// Start with the original diffuse texture coordinate:
	float2 DiffuseCoord = vsout.DiffuseTextureCoordinate;

	// If the surface is "rippling", wobble the texture coordinate.
	if (Rippling) {
		float2 RippleOffset = { sin(Time + vsout.SourcePosition.x / 32) / 8, cos(Time + vsout.SourcePosition.z / 32) / 8 };
		DiffuseCoord += RippleOffset;

	// Calculate the colour map look-up coordinate from the diffuse and lightmap textures:
	float2 ColourMapIndex = {
		tex2D(DiffuseTextureSampler, DiffuseCoord).a,
		1 - (float)tex2D(LightMapTextureSampler, vsout.LightMapTextureCoordinate).rgba
	// Look up and return the value from the colour map.
	return tex2D(ColourMapSampler, ColourMapIndex).rgba;

technique TransformAndTexture {
	pass P0 {
		vertexShader = compile vs_2_0 Transform();
		pixelShader  = compile ps_2_0 ApplyTexture();

It would no doubt be faster to have two techniques; one for rippling surfaces and one for still surfaces. It is, however, easier to use the above and switch the rippling on and off when required (rather than group surfaces and switch techniques). Given that the framerate rises from ~135FPS to ~137FPS on my video card if I remove the ripple effect altogether, it doesn't seem worth it.

Sorting out the order in which polygons are drawn looks like it's going to get important, as I need to support alpha-blended surfaces for Quake 2, and there are some nasty areas of Z-fighting cropping up.

Alpha-blending in 8-bit? Software Quake didn't support any sort of alpha blending (hence the need to re-vis levels for use with Quake GL as underneath the opaque waters were marked as invisible), and Quake 2 has a data file that maps 16-bit colour values to 8-bit palette indices. Quake 2 also had a "stipple alpha" mode used a dither pattern to handle the two translucent surface opacities (1/3 and 2/3 ratios).


Tuesday, 7th August 2007

Following sirob's prompting, I dropped the BasicEffect for rendering and rolled my own effect. After seeing the things that could be done with them (pixel and vertex shaders) I'd assumed they'd be hard to put together, and that I'd need to change my code significantly.

In reality all I've had to do is copy and paste the sample from the SDK documentation, load it into the engine (via the content pipeline), create a custom vertex declaration to handle two sets of texture coordinates (diffuse and lightmap) and strip out all of the duplicate code I had for creating and rendering from two vertex arrays.

2007.08.06.01.jpg   2007.08.06.02.jpg

public struct VertexPositionTextureDiffuseLightMap {

	public Xna.Vector3 Position;
	public Xna.Vector2 DiffuseTextureCoordinate;
	public Xna.Vector2 LightMapTextureCoordinate;

	public VertexPositionTextureDiffuseLightMap(Xna.Vector3 position, Xna.Vector2 diffuse, Xna.Vector2 lightMap) {
		this.Position = position;
		this.DiffuseTextureCoordinate = diffuse;
		this.LightMapTextureCoordinate = lightMap;

	public readonly static VertexElement[] VertexElements = new VertexElement[]{
		new VertexElement(0, 0, VertexElementFormat.Vector3, VertexElementMethod.Default, VertexElementUsage.Position, 0),
		new VertexElement(0, 12, VertexElementFormat.Vector2, VertexElementMethod.Default, VertexElementUsage.TextureCoordinate, 0),
		new VertexElement(0, 20, VertexElementFormat.Vector2, VertexElementMethod.Default, VertexElementUsage.TextureCoordinate, 1)

uniform extern float4x4 WorldViewProj : WORLDVIEWPROJECTION;

uniform extern texture DiffuseTexture;
uniform extern texture LightMapTexture;

uniform extern float Time;

struct VS_OUTPUT {
    float4 Position : POSITION;
    float2 DiffuseTextureCoordinate : TEXCOORD0;
    float2 LightMapTextureCoordinate : TEXCOORD1;

sampler DiffuseTextureSampler = sampler_state {
    Texture = <DiffuseTexture>;
    mipfilter = LINEAR;

sampler LightMapTextureSampler = sampler_state {
	Texture = <LightMapTexture>;
	mipfilter = LINEAR;

VS_OUTPUT Transform(float4 Position : POSITION, float2 DiffuseTextureCoordinate : TEXCOORD0, float2 LightMapTextureCoordinate : TEXCOORD1) {

    Out.Position = mul(Position, WorldViewProj);
    Out.DiffuseTextureCoordinate = DiffuseTextureCoordinate;
    Out.LightMapTextureCoordinate = LightMapTextureCoordinate;

    return Out;

float4 ApplyTexture(VS_OUTPUT vsout) : COLOR {
	float4 DiffuseColour = tex2D(DiffuseTextureSampler, vsout.DiffuseTextureCoordinate).rgba;
	float4 LightMapColour = tex2D(LightMapTextureSampler, vsout.LightMapTextureCoordinate).rgba;
    return DiffuseColour * LightMapColour;

technique TransformAndTexture {
    pass P0 {
        vertexShader = compile vs_2_0 Transform();
        pixelShader  = compile ps_2_0 ApplyTexture();

Of course, now I have that up and running I might as well have a play with it...


By adding up and dividing the individual RGB components of the lightmap texture by three you can simulate the monochromatic lightmaps used by Quake 2's software renderer. Sadly I know not of a technique to go the other way and provide colourful lightmaps for Quake 1. smile.gif Not very interesting, though.


I've always wanted to do something with pixel shaders as you get to play with tricks that are a given in software rendering with the speed of dedicated hardware acceleration. I get the feeling that the effect (or a variation of it, at least) will be handy for watery textures.

float4 ApplyTexture(VS_OUTPUT vsout) : COLOR {
	float2 RippledTexture = vsout.DiffuseTextureCoordinate;
	RippledTexture.x += sin(vsout.DiffuseTextureCoordinate.y * 16 + Time) / 16;
	RippledTexture.y += sin(vsout.DiffuseTextureCoordinate.x * 16 + Time) / 16;
	float4 DiffuseColour = tex2D(DiffuseTextureSampler, RippledTexture).rgba;
	float4 LightMapColour = tex2D(LightMapTextureSampler, vsout.LightMapTextureCoordinate).rgba;
	return DiffuseColour * LightMapColour;

My code is no doubt suboptimal (and downright stupid).

Naturally, I needed to try and duplicate Scet's software rendering simulation trick. smile.gif

The colour map (gfx/colormap.lmp) is a 256×64 array of bytes. Each byte is an index to a colour palette entry, on the X axis is the colour and on the Y axis is the brightness: ie, RGBColour = Palette[ColourMap[DiffuseColour, Brightness]]. I cram the original diffuse colour palette index into the (unused) alpha channel of the ARGB texture, and leave the lightmaps untouched.

float2 LookUp = 0;
LookUp.x = tex2D(DiffuseTextureSampler, vsout.DiffuseTextureCoordinate).a;
LookUp.y = (1 - tex2D(LightMapTextureSampler, vsout.LightMapTextureCoordinate).r) / 4;
return tex2D(ColourMapTextureSampler, LookUp);
2007.08.06.06.jpg   2007.08.06.07.jpg

2007.08.06.08.jpg   2007.08.06.09.jpg

As I'm not loading the mip-maps (and am letting Direct3D handle generation of mip-maps for me) I have to disable mip-mapping for the above to work, as otherwise you'd end up with non-integral palette indices. The results are therefore a bit noisier in the distance than in vanilla Quake, but I like the 8-bit palette look. At least the fullbright colours work.

Less Colourful Quake 2

Monday, 6th August 2007


I've transferred the BSP rendering code to use the new level loading code, so I can now display correctly-coloured Quake 2 levels. smile.gif The Quake stuff is in its own assembly, and is shared by the WinForms resource browser project and the XNA renderer.


I'm also now applying lightmaps via multiplication rather than addition, so they look significantly better.


A shader solution would be optimal. I'm currently just drawing the geometry twice, the second time with some alpha blending enabled.


Keyboard Handler Fix

Friday, 3rd August 2007

ArchG indicated a bug in the TextInputHandler class I posted a while back - no reference to the delegate instance used for the unmanaged callback is held, so as soon as the garbage collector kicks in things go rather horribly wrong.

/* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
 * XnaTextInput.TextInputHandler - benryves@benryves.com                                     *
 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
 * This is quick and very, VERY dirty.                                                       *
 * It uses Win32 message hooks to grab messages (as we don't get a nicely wrapped WndProc).  *
 * I couldn't get WH_KEYBOARD to work (accessing the data via its pointer resulted in access *
 * violation exceptions), nor could I get WH_CALLWNDPROC to work.                            *
 * Maybe someone who actually knows what they're  doing can work something out that's not so *
 * kludgy.                                                                                   *
 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
 * This quite obviously relies on a Win32 nastiness, so this is for Windows XNA games only!  *
 * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */

#region Using Statements
using System;
using System.Runtime.InteropServices;
using System.Windows.Forms; // This class exposes WinForms-style key events.

namespace XnaTextInput {

	/// <summary>
	/// A class to provide text input capabilities to an XNA application via Win32 hooks.
	/// </summary>
	class TextInputHandler : IDisposable {

		#region Win32

		/// <summary>
		/// Types of hook that can be installed using the SetWindwsHookEx function.
		/// </summary>
		public enum HookId {
			WH_CBT = 5,
			WH_DEBUG = 9,
			WH_MAX = 11,
			WH_MIN = -1,
			WH_MOUSE_LL = 14,
			WH_SHELL = 10,

		/// <summary>
		/// Window message types.
		/// </summary>
		/// <remarks>Heavily abridged, naturally.</remarks>
		public enum WindowMessage {
			WM_KEYDOWN = 0x100,
			WM_KEYUP = 0x101,
			WM_CHAR = 0x102,

		/// <summary>
		/// A delegate used to create a hook callback.
		/// </summary>
		public delegate int GetMsgProc(int nCode, int wParam, ref Message msg);

		/// <summary>
		/// Install an application-defined hook procedure into a hook chain.
		/// </summary>
		/// <param name="idHook">Specifies the type of hook procedure to be installed.</param>
		/// <param name="lpfn">Pointer to the hook procedure.</param>
		/// <param name="hmod">Handle to the DLL containing the hook procedure pointed to by the lpfn parameter.</param>
		/// <param name="dwThreadId">Specifies the identifier of the thread with which the hook procedure is to be associated.</param>
		/// <returns>If the function succeeds, the return value is the handle to the hook procedure. Otherwise returns 0.</returns>
		[DllImport("user32.dll", EntryPoint = "SetWindowsHookExA")]
		public static extern IntPtr SetWindowsHookEx(HookId idHook, GetMsgProc lpfn, IntPtr hmod, int dwThreadId);

		/// <summary>
		/// Removes a hook procedure installed in a hook chain by the SetWindowsHookEx function. 
		/// </summary>
		/// <param name="hHook">Handle to the hook to be removed. This parameter is a hook handle obtained by a previous call to SetWindowsHookEx.</param>
		/// <returns>If the function fails, the return value is zero. To get extended error information, call GetLastError.</returns>
		public static extern int UnhookWindowsHookEx(IntPtr hHook);

		/// <summary>
		/// Passes the hook information to the next hook procedure in the current hook chain.
		/// </summary>
		/// <param name="hHook">Ignored.</param>
		/// <param name="ncode">Specifies the hook code passed to the current hook procedure.</param>
		/// <param name="wParam">Specifies the wParam value passed to the current hook procedure.</param>
		/// <param name="lParam">Specifies the lParam value passed to the current hook procedure.</param>
		/// <returns>This value is returned by the next hook procedure in the chain.</returns>
		public static extern int CallNextHookEx(int hHook, int ncode, int wParam, ref Message lParam);

		/// <summary>
		/// Translates virtual-key messages into character messages.
		/// </summary>
		/// <param name="lpMsg">Pointer to an Message structure that contains message information retrieved from the calling thread's message queue.</param>
		/// <returns>If the message is translated (that is, a character message is posted to the thread's message queue), the return value is true.</returns>
		public static extern bool TranslateMessage(ref Message lpMsg);

		/// <summary>
		/// Retrieves the thread identifier of the calling thread.
		/// </summary>
		/// <returns>The thread identifier of the calling thread.</returns>
		public static extern int GetCurrentThreadId();


		#region Hook management and class construction.

		/// <summary>Handle for the created hook.</summary>
		private readonly IntPtr HookHandle;

		private readonly GetMsgProc ProcessMessagesCallback;

		/// <summary>Create an instance of the TextInputHandler.</summary>
		/// <param name="whnd">Handle of the window you wish to receive messages (and thus keyboard input) from.</param>
		public TextInputHandler(IntPtr whnd) {
			// Create the delegate callback:
			this.ProcessMessagesCallback = new GetMsgProc(ProcessMessages);
			// Create the keyboard hook:
			this.HookHandle = SetWindowsHookEx(HookId.WH_GETMESSAGE, this.ProcessMessagesCallback, IntPtr.Zero, GetCurrentThreadId());

		public void Dispose() {
			// Remove the hook.
			if (this.HookHandle != IntPtr.Zero) UnhookWindowsHookEx(this.HookHandle);


		#region Message processing

		private int ProcessMessages(int nCode, int wParam, ref Message msg) {
			// Check if we must process this message (and whether it has been retrieved via GetMessage):
			if (nCode == 0 && wParam == 1) {

					// We need character input, so use TranslateMessage to generate WM_CHAR messages.
					TranslateMessage(ref msg);

					// If it's one of the keyboard-related messages, raise an event for it:
					switch ((WindowMessage)msg.Msg) {
						case WindowMessage.WM_CHAR:
							this.OnKeyPress(new KeyPressEventArgs((char)msg.WParam));
						case WindowMessage.WM_KEYDOWN:
							this.OnKeyDown(new KeyEventArgs((Keys)msg.WParam));
						case WindowMessage.WM_KEYUP:
							this.OnKeyUp(new KeyEventArgs((Keys)msg.WParam));


			// Call next hook in chain:
			return CallNextHookEx(0, nCode, wParam, ref msg);


		#region Events

		public event KeyEventHandler KeyUp;
		protected virtual void OnKeyUp(KeyEventArgs e) {
			if (this.KeyUp != null) this.KeyUp(this, e);

		public event KeyEventHandler KeyDown;
		protected virtual void OnKeyDown(KeyEventArgs e) {
			if (this.KeyDown != null) this.KeyDown(this, e);

		public event KeyPressEventHandler KeyPress;
		protected virtual void OnKeyPress(KeyPressEventArgs e) {
			if (this.KeyPress != null) this.KeyPress(this, e);


I wrote a crude ZSoft PCX loader (only handles 8-bit per plane, single-plane images, which is sufficient for Quake 2).


Using this loader I found colormap.pcx, which appears to perform the job of palette and colour map for Quake II.


.wal files now open with the correct palette. I've also copied over most of the BSP loading code, but it needs a good going-over to make it slightly more sane (especially where the hacks for Quake II support have been added).

Loader Change

Thursday, 2nd August 2007

I've started rewriting the underlying resource loading code to better handle multiple versions of the game.


To help with this I'm writing a WinForms-based resource browser.

(That's the only real Quake-related change visible in the above screenshot. I've written a cinematic (.cin, used in Quake 2) loader).

To aid loading resources I've added a number of new generic types. For example, the Picture class always represents a 32-bit per pixel ARGB 2D picture. The decoders for various formats will always have access to the resource manager, so they can request palette information if they need it. To further aid issues, there are some handy interfaces that a specific format class can implement - for example, a class (such as WallTexture for handling .wal files) implementing IPictureLoader will always have a GetPicture() method.

The loader classes are also given attributes specifying which file extensions are handled. (This project uses quite a bit of reflection now). The only issue I can see with this are files that use the same extension but have different types, such as the range of .lmp files.

In addition, certain single files within the packages have multiple sub-files (for example, the .wad files in Quake). I'm not sure how I'll handle this, but I'm currently thinking of having the .wad loader implement IPackage so you could access files via gfx/somewad.wad/somefileinthewad, but some files don't have names or extensions.

Quake 2 and Emulation

Wednesday, 1st August 2007

The current design of the Quake project is that there are a bunch of classes in the Data namespace that are used to decode Quake's structures in a fairly brain-dead manner. To do anything useful with it you need to build up your own structures suitable for the way you intend on rendering the level.

The problem comes in when you try to load resources from different versions of Quake. Quake 1 and Quake 2 have quite a few differences. One major one is that every BSP level in Quake contains its own mip textures. You can call a method in the BSP class which returns sane texture coordinates as it can inspect the texture dimensions inside itself. Quake 2 stores all of the textures externally in .wal resources - the BSP class can no longer calculate texture coordinates as it can't work out how large the textures are as it can't see outside itself.

I guess the only sane way to work this out is to hide the native types from the end user and wrap everything up, but I've never liked this much as you might neglect to wrap up something that someone else would find very important, or you do something that is unsuitable for the way they really wanted to work.

Anyhow. I've hacked around the BSP loader to within an inch of its life and it seems to be (sort of) loading Quake 2 levels for brute-force rendering. Quake 2 boasts truecolour lightmaps, improving the image quality quite significantly!


The truecolour lightmaps show off the Strogg disco lighting to its best effect. One of the problems with the Quake II BSP file format is that the indexing of lumps inside the file has changed. Not good.


That's a bit better. smile.gif Quake II's lightmaps tend to stick to the red/brown/yellow end of the spectrum, but that is a truecolour set of lightmaps in action!

2007.07.31.03.jpg   2007.07.31.04.jpg

The lightmaps tend to look a bit grubby where they don't line up between faces. Some trick to join all lightmaps for a plane together into a single texture should do the trick, and reduce the overhead of having to load thousands of tiny textures (which I'm guessing have to be scaled up to a power-of-two). I'll have to look into it.

On to .wal (wall texture) loading - and I can't find a palette anywhere inside the Quake II pack files. I did find a .act (Photoshop palette) that claimed to be for Quake II, but it doesn't quite seem to match. It's probably made up of the right colours, but not in the right order.


Fortunately I have some PAK files with replacement JPEG textures inside them and can load those instead for the moment.

2007.07.31.06.jpg   2007.07.31.07.jpg   2007.07.31.08.jpg

The brightness looks strange due to the bad way I apply the lightmaps - some kludgy forced two-pass affair with alpha blending modes set to something that sort of adds the two textures together in a not-very-convincing manner.

Can anyone recommend a good introduction to shaders for XNA? I'm not really trying to do anything that exciting.

This is a really bad and vague overview of the emulation technique I use in Cogwheel, so I apologise in advance. Emulation itself is very simple when done in the following manner - all you really need is a half-decent knowledge of how the computer you're emulating works at the assembly level. The following is rather Z80-specific.

At the heart of the system is its CPU. This device reads instructions from memory and depending on the value it reads it performs a variety of different actions. It has a small amount of memory inside itself which it uses to store its registers, variables used during execution. For example, the PC register is used as a pointer to the next instruction to fetch and execute from memory, and the SP register points at the top of the stack.

It can interact with the rest of the system in three main ways:

  • Read/Write Memory
  • Input/Output Hardware
  • Interrupt Request

I assume you're familiar with memory. smile.gif The hardware I refer to are peripheral devices such as video display processors, keypads, sound generators and so on. Data is written to and read from these devices on request. What the hardware device does with that data is up to it. I'll ignore interrupt requests for the moment.

The CPU at an electronic level communicates with memory and hardware using two buses and a handful of control pins. The two buses are the address bus and data bus. The address bus is read-only (when viewed from outside the CPU) and is used to specify a memory address or a hardware port number. It is 16 bits wide, meaning that 64KB memory can be addressed. Due to the design, only the lower 8-bits are normally used for hardware addressing, giving you up to 256 different hardware devices.

The data bus is 8-bits wide (making the Z80 an "8-bit" CPU). It can be read from or written to, depending on the current instruction.

The exact function of these buses - whether you're addressing memory or a hardware device, or whether you're reading or writing - is relayed to the external hardware via some control pins on the CPU itself. The emulator author doesn't really need to emulate these. Rather, we can do something like this:

class CpuEmulator {

	public virtual void WriteMemory(ushort address, byte value) {
		// Write to memory.
	public virtual byte ReadMemory(ushort address) {
		// Read from memory.
		return 0x00;

	public virtual void WriteHardware(ushort address, byte value) {
		// Write to hardware.
	public virtual byte ReadHardware(ushort address) {
		// Read from hardware.
		return 0x00;


A computer with a fixed 64KB RAM, keyboard on hardware port 0 and console (for text output) on port 1 might look like this:

class SomeBadComputer : CpuEmulator {

	private byte[] AllMemory = new byte[64 * 1024];

	public override void WriteMemory(ushort address, byte value) {
		AllMemory[address] = value;
	public override byte ReadMemory(ushort address) {
		return AllMemory[address];

	public override void WriteHardware(ushort address, byte value) {
		switch (address & 0xFF) {
			case 1:
	public override byte ReadHardware(ushort address) {
		switch (address & 0xFF) {
			case 0:
				return (byte)Console.ReadKey();
				return 0x00;


This is all very well, but how does the CPU actually do anything worthwhile?

It needs to read instructions from memory, decode them, and act on them. Suppose our CPU had two registers - 16-bit PC (program counter) and 8-bit A (accumulator) and this instruction set:

00nn   : Load 'nn' into accumulator.
01nn   : Output accumulator to port N.
02nn   : Input to accumulator from port N.
03nnnn : Read from memory address nnnn to accumulator.
04nnnn : Write accumulator to memory address nnnn.
05nnnn : Jump to address nnnn.

Extending the above CpuEmulator class, we could get something like this:

partial class CpuEmulator {

	public ushort RegPC = 0;
	public byte RegA = 0;
	private int CyclesPending = 0;
	public void FetchExecute() {
		switch (ReadMemory(RegPC++)) {
			case 0x00:
				RegA = ReadMemory(RegPC++);
				CyclesPending += 8;
			case 0x01:
				WriteHardware(ReadMemory(RegPC++), RegA);
				CyclesPending += 8;
			case 0x02:
				RegA = ReadHardware(ReadMemory(RegPC++));
				CyclesPending += 16;
			case 0x03:
				RegA = ReadMemory((ushort)(ReadMemory(RegPC++) + ReadMemory(RegPC++) * 256));
				CyclesPending += 16;
			case 0x04:
				WriteMemory((ushort)(ReadMemory(RegPC++) + ReadMemory(RegPC++) * 256), RegA);
				CyclesPending += 24;
			case 0x05:
				RegPC = (ushort)(ReadMemory(RegPC++) + ReadMemory(RegPC++) * 256);
				CyclesPending += 24;
				// NOP
				CyclesPending += 4;


The CyclesPending variable is used for timing. Instructions take a variable length of time to run (depending on complexity, length of opcode, whether it needs to access memory and so on). This time is typically measured in the number of clock cycles taken for the CPU to execute the instruction.

Using the above CyclesPending += x style one can write a function that will execute a particular number of cycles:

partial class CpuEmulator {

	public void Tick(int cycles) {
		CyclesPending -= cycles;
		while (CyclesPending < 0) FetchExecute();


For some truly terrifying code, an oldish version of Cogwheel's instruction decoding switch block. That code has been automatically generated from a text file, I didn't hand-type it all.

Um... that's pretty much all there is. The rest is reading datasheets! Your CPU would need to execute most (if not all) instructions correctly, updating its internal state (and registers) as the hardware would. The non-CPU hardware (video processor, sound processor, controllers and so on) would also need to conform to data reads and writes correctly.

As far as timing goes, various bits of hardware need to run at their own pace. One scanline (of the video processor) is a good value for the Master System. Cogwheel provides this method to run the emulator for a single frame:

public void RunFrame() {
	this.VDP.RunFramePending = false;
	while (!this.VDP.RunFramePending) {

In the Master System's case, one scanline is displayed every 228 clock cycles. Some programs update the VDP on every scanline (eg changing the background horizontal scroll offset to skew the image in a driving game).

The above is embarrassingly vague, so if anyone is interested enough to want clarification on anything I'd be happy to give it.

FirstJuly 2007September 2007Last RSSSearchBrowse by dateIndexTags