Graphical consistency problems when resizing

Started by
9 comments, last by Supernat02 19 years, 5 months ago
I've been working on OpenGL and DirectDraw for a little while now, but I recently started doing the DirectX tutorials. I am currently working on Tut 5. Everything looks great with the preinitialized window resolution, but if you dare touch it, then it will look really ugly. I think I need to call something on the WM_SIZE msg, but then again other programs look fine even when they don't call that. What would an experienced D3D programmer tell me ? TIA
Advertisement
Well I believe that if you want your program to be resizeable then when WM_SIZE is generated you need to adjust the projection or orthographic matrix depending on the new resolution.

If you want the border to be static(non-resizeable) create the window with WS_STATICEDGE(well I don't remember exactly what it's called but it's something like that).

Also I find that when the window is at a non 4:3 ratio(400x300,640x480,800x600,etc..) resolution the FPS drops and the scene looks like crap so I just make my apps non-resizable.

Edit: OK to create a non-resizeable window use CreateWindowEx and use WS_EX_STATICEDGE as the first parameter.
------------------------------"Are you pondering what I am pondering?" - Fruny"I think so Brain, but how do we make a hockey game with DarkBASIC?" - Scint
If you're talking about your screen getting pixelly/blurry, then it's because you haven't resized the backbuffer.

To do so, you need to change the BackBufferWidth/Height in your presentParameters, the call pDevice->Reset(&presentParameters).

This can be a bit of a pain, because you need to release all non-managed resources (Like the backbuffer, any fonts and any textures created without D3DPOOL_MANAGED), then recreate them after the reset. You also need to set up all of your render-states again, and all your matrix settings.

There might be a better way of doing this, but this is the way I'm doing it at the moment.
Thanks for the responses
I see that it's not that easy :(.

I have adjusted the width and height parameters in the d3dpp.

My question now is :
Would I call reset in a WM_SIZE msg from windows? Or is there a better way of doing this within the render scene(as I've seen other programs do invisibly to my knowledge)?

Wherever the code goes, I'm assuming it would look like this :

RECT box ;
GetClientRect(hWnd, &box) ;
d3dpp.BackBufferWidth = box.right ;
d3dpp.BackBufferHeight = box.bottom ;
g_pd3dDevice->Reset(&d3dpp) ;

I can't think of a better way. I gotta read up on this stuff. The SDK doesn't cover this. =(
That's pretty much it! Just dont forget to call Release() on any fonts, additional swap chains or non-managed surfaces (If you dont have any textures, then that will probably just be your backbuffer) and then re-create them afterwards.

	// Set the screen 'present parameters'	m_presentParameters.BackBufferWidth = m_windowWidth;	m_presentParameters.BackBufferHeight = m_windowHeight;	// Release all non-managed D3D resources	m_defaultFont->Release();	m_BackSurf->Release();	// Reset the device with the new present parameters	if(FAILED(m_pDevice->Reset(&m_presentParameters)))		throw "Error resetting D3D9 device!";	// Reaquire back buffer	m_pDevice->GetBackBuffer(0, 0, D3DBACKBUFFER_TYPE_MONO, &m_BackSurf);	// Recreate default font	D3DXCreateFont(m_pDevice, (HFONT)GetStockObject(SYSTEM_FIXED_FONT), &m_defaultFont);	// Set any required renderstates below...	m_pDevice->SetRenderState( D3DRS_LIGHTING, TRUE);	m_pDevice->SetRenderState( D3DRS_ZENABLE, D3DZB_TRUE);	m_pDevice->SetRenderState(D3DRS_CULLMODE, D3DCULL_CCW);	SetPerspectiveMode(m_windowWidth, m_windowHeight, m_fov);


Note that 'SetPerspectiveMode' just recreates the projection matrix using D3DXMatrixPerspectiveFovLH(). This may seem like a slow procedure, but remember that your user wont be sizing the window every frame.

Good luck!
I understand what you're saying, but I'm not sure if I'm implementing it.

//------------------------------------------------------------------// Textures: renders a cylinder and wraps a texture around it//------------------------------------------------------------------#include <windows.h>#include <mmsystem.h>#include <d3dx9.h>#include "resource.h"//#define SHOW_HOW_TO_USE_TCI// Global variablesLPDIRECT3D9				g_pD3D			= NULL ; // used create the D3DDeviceLPDIRECT3DDEVICE9		g_pd3dDevice	= NULL ; // our rendering deviceLPDIRECT3DVERTEXBUFFER9	g_pVB			= NULL ; // buffer to hold verticesLPDIRECT3DTEXTURE9		g_pTexture		= NULL ; // Our texture// structure for custom vertex type.  Normal added(color is provided by the material)struct CUSTOMVERTEX{	D3DXVECTOR3 position ;	// the 3D position for the vertex	D3DCOLOR	color ;		// the surface normal for the vertex#ifndef SHOW_HOW_TO_USE_TCI	FLOAT		tu, tv ;	// The texture coordinates#endif} ;// custom FVF, describes custom vertex structure#ifdef SHOW_HOW_TO_USE_TCI#define D3DFVF_CUSTOMVERTEX (D3DFVF_XYZ | D3DFVF_DIFFUSE)#else#define D3DFVF_CUSTOMVERTEX (D3DFVF_XYZ | D3DFVF_DIFFUSE | D3DFVF_TEX1)#endifHRESULT InitD3D(HWND hWnd);HRESULT InitGeometry();VOID SetupMatrices();VOID Render();VOID Cleanup();LRESULT WINAPI MsgProc(HWND hWnd,UINT msg,WPARAM wParam,LPARAM lParam);HWND g_hWnd ;D3DPRESENT_PARAMETERS d3dpp ;int g_x = 640 ;int g_y = 480 ;INT WINAPI WinMain(HINSTANCE hInstance,HINSTANCE hPrevInstance,LPSTR lPCmdLine,int nCmdShow){	// register the window class	WNDCLASSEX wc = {sizeof(WNDCLASSEX), CS_CLASSDC, MsgProc, 0L, 0L, GetModuleHandle(NULL),		LoadIcon(hInstance, MAKEINTRESOURCE(IDI_ICON1)), NULL, NULL, NULL, "D3D Tutorial", NULL} ;	RegisterClassEx(&wc) ;	// create the application's window	HWND hWnd=CreateWindow("D3D Tutorial", "D3D Tutorial 05 : Textures", WS_OVERLAPPEDWINDOW,		100, 100, g_x, g_y, GetDesktopWindow(), NULL, wc.hInstance, NULL) ;	g_hWnd = hWnd ;	// initialize Direct3D	if(SUCCEEDED(InitD3D(hWnd)))	{		// create the geometry		if(SUCCEEDED(InitGeometry()))		{			// show the window			ShowWindow(hWnd, SW_SHOWDEFAULT) ;			UpdateWindow(hWnd) ;			// enter the message loop			MSG msg ;			ZeroMemory(&msg, sizeof(msg)) ;			while(msg.message != WM_QUIT)			{				if(PeekMessage(&msg, NULL, 0U, 0U, PM_REMOVE))				{					TranslateMessage(&msg) ;					DispatchMessage(&msg) ;				}				else					Render() ;			}		}	}	UnregisterClass("D3D Tutorial", wc.hInstance) ;	return 0 ;}HRESULT InitD3D(HWND hWnd){	// create the D3D object	if((g_pD3D = Direct3DCreate9(D3D_SDK_VERSION)) == NULL)		return E_FAIL ;	// set up the structure used to create the D3DDevice.  Since we are now	// using more complex geometry, we will create a device with a zbuffer.	ZeroMemory(&d3dpp, sizeof(d3dpp)) ;	d3dpp.Windowed					= TRUE ;	d3dpp.SwapEffect				= D3DSWAPEFFECT_DISCARD ;	d3dpp.BackBufferWidth			= g_x ;	d3dpp.BackBufferHeight			= g_y ;	d3dpp.BackBufferFormat			= D3DFMT_UNKNOWN ;	d3dpp.EnableAutoDepthStencil	= TRUE ;	d3dpp.AutoDepthStencilFormat	= D3DFMT_D16 ;	// create the D3DDevice	if(FAILED(g_pD3D->CreateDevice(D3DADAPTER_DEFAULT, D3DDEVTYPE_HAL, hWnd,		D3DCREATE_HARDWARE_VERTEXPROCESSING, &d3dpp, &g_pd3dDevice)))		return E_FAIL ;	// turn of culling	g_pd3dDevice->SetRenderState(D3DRS_CULLMODE, D3DCULL_NONE) ;	// turn off D3D lighting	g_pd3dDevice->SetRenderState(D3DRS_LIGHTING, FALSE) ;	// turn on the zbuffer	g_pd3dDevice->SetRenderState(D3DRS_ZENABLE, TRUE) ;	return S_OK ;}HRESULT InitGeometry(){	// use D3DX to create a texture from a file based image	if(FAILED(D3DXCreateTextureFromFile(g_pd3dDevice, "hovigTex.bmp", &g_pTexture)))		return E_FAIL ;	// create the vertex buffer	if(FAILED(g_pd3dDevice->CreateVertexBuffer(100 * sizeof(CUSTOMVERTEX), 0,		D3DFVF_CUSTOMVERTEX, D3DPOOL_DEFAULT, &g_pVB, NULL)))		return E_FAIL ;	// fill the vertex buffer.  We are algorithmically generating a cylinder	// here, including the normals, which are used for lighting	CUSTOMVERTEX* pVertices ;	if(FAILED(g_pVB->Lock(0, 0, (void**)&pVertices, 0)))		return E_FAIL ;	for(DWORD i = 0; i < 50; i++)	{		FLOAT theta = (2 * D3DX_PI * i) / 49 ;		pVertices[2 * i + 0].position	= D3DXVECTOR3(sinf(theta), -1.0f, cosf(theta)) ;		pVertices[2 * i + 0].color		= 0xffffffff ;#ifndef SHOW_HOW_TO_USE_TCI		pVertices[2 * i + 0].tu			= FLOAT(i) / (49) ;		pVertices[2 * i + 0].tv			= 1.0f ;#endif		pVertices[2 * i + 1].position	= D3DXVECTOR3(sinf(theta), 1.0f, cosf(theta)) ;		pVertices[2 * i + 1].color		= 0xff808080 ;#ifndef SHOW_HOW_TO_USE_TCI		pVertices[2 * i + 1].tu			= FLOAT(i) / (49) ;		pVertices[2 * i + 1].tv			= 0.0f ;#endif	}	g_pVB->Unlock() ;	return S_OK ;}VOID SetupMatrices(){	// for our world matrix, we will just leave it as the identity	D3DXMATRIXA16 matWorld ;	D3DXMatrixIdentity(&matWorld) ;	D3DXMatrixRotationX(&matWorld, timeGetTime() / 1000.0f) ;	g_pd3dDevice->SetTransform(D3DTS_WORLD, &matWorld) ;	// set up our view matrix.  A view matrix can be defined given an eye point,	// a point to lookat, and a direction for which way is up.  Here, we set the	// eye five units back along the z-axis and up three units, look at the	// origin, and define "up" to be in the y-direction.	D3DXVECTOR3 vEyePt(0.0f, 3.0f, -5.0f) ;	D3DXVECTOR3 vLookatPt(0.0f, 0.0f, 0.0f) ;	D3DXVECTOR3 vUpVec(0.0f, 1.0f, 0.0f) ;	D3DXMATRIXA16 matView ;	D3DXMatrixLookAtLH(&matView, &vEyePt, &vLookatPt, &vUpVec) ;	g_pd3dDevice->SetTransform(D3DTS_VIEW, &matView) ;		// for the projection matrix, we set up a perspective transform (which	// transforms geometry from 3D view space to 2D viewport space, with a	// perspective device making objects smaller in the distance).  To build	// a perspective transform, we need the field of view (1/4 pi is common),	// the aspect ratio, and the near and far clipping planes (which define at	// what distances geometry should no longer be rendered).	D3DXMATRIXA16 matProj ;	D3DXMatrixPerspectiveFovLH(&matProj, D3DX_PI / 4, 1.0f, 1.0f, 100.0f) ;	g_pd3dDevice->SetTransform(D3DTS_PROJECTION, &matProj) ;}VOID Render(){	// clear the backbuffer and the zbuffer to a black color	g_pd3dDevice->Clear(0, NULL, D3DCLEAR_TARGET | D3DCLEAR_ZBUFFER,						D3DCOLOR_XRGB(0, 0, 0), 1.0f, 0) ;	// begin the scene	if(SUCCEEDED(g_pd3dDevice->BeginScene()))	{		// setup the world, view, and projection matrices		SetupMatrices() ;		// Setup our texture.  Using textures introduces the texture state states		// which govern how textures get blended together (in the case of multiple		// textures) and lighting information.  In this case, we are modulating		// (blending) our texture with the diffuse color of the vertices.		g_pd3dDevice->SetTexture(0, g_pTexture) ;		g_pd3dDevice->SetTextureStageState(0, D3DTSS_COLOROP,	D3DTOP_MODULATE) ;		g_pd3dDevice->SetTextureStageState(0, D3DTSS_COLORARG1,	D3DTA_TEXTURE) ;		g_pd3dDevice->SetTextureStageState(0, D3DTSS_COLORARG2,	D3DTA_DIFFUSE) ;		g_pd3dDevice->SetTextureStageState(0, D3DTSS_ALPHAOP,	D3DTOP_DISABLE) ;#ifdef SHOW_HOW_TO_USE_TCI		// Note : to use D3D texture coordinate generation, use teh stage state		// D3DTSS_TEXCOORDINDEX, as shown below.  In this example, we are using		// the position of the vertex in camera space to generate texture		// coordinates.  The tex coord index (TCI) parameters are passed into a		// texture transform, which is a 4x4 matrix which trasnforms the x, y, z		// TCI coordinates into tu, tv texture coordinates.		// In this example, the texture matrix is setup to		// transform the texture from (-1, +1) position coordinates to (0, 1)		// texture coordinate space:		//		tu =  0.5 * x + 0.5 		//		tv = -0.5 * y + 0.5		D3DXMATRIXA16 mat ;				mat._11 = 0.25f ; mat._12 =  0.00f ; mat._13 = 0.00f ; mat._14 = 0.00f ;		mat._21 = 0.00f ; mat._22 = -0.25f ; mat._23 = 0.00f ; mat._24 = 0.00f ;		mat._31 = 0.00f ; mat._32 =  0.00f ; mat._33 = 1.00f ; mat._34 = 0.00f ;		mat._41 = 0.50f ; mat._42 =  0.50f ; mat._43 = 0.00f ; mat._44 = 1.00f ;		g_pd3dDevice->SetTransform(D3DTS_TEXTURE0, &mat) ;		g_pd3dDevice->SetTextureStageState(0, D3DTSS_TEXTURETRANSFORMFLAGS, D3DTTFF_COUNT2) ;		g_pd3dDevice->SetTextureStageState(0, D3DTSS_TEXCOORDINDEX, D3DTSS_TCI_CAMERASPACEPOSITION) ;#endif		// render the vertex buffer contents		g_pd3dDevice->SetStreamSource(0, g_pVB, 0, sizeof(CUSTOMVERTEX)) ;		g_pd3dDevice->SetFVF(D3DFVF_CUSTOMVERTEX) ;		g_pd3dDevice->DrawPrimitive(D3DPT_TRIANGLESTRIP, 0, 98) ;				// end the scene		g_pd3dDevice->EndScene() ;	}	// present the backbuffer contents to the display	g_pd3dDevice->Present(NULL, NULL, NULL, NULL) ;}VOID Cleanup(){	if(g_pTexture != NULL)		g_pTexture->Release() ;	if(g_pVB != NULL)		g_pVB->Release() ;	if(g_pd3dDevice != NULL)		g_pd3dDevice->Release() ;	if(g_pD3D != NULL)		g_pD3D->Release() ;}LRESULT WINAPI MsgProc(HWND hWnd, UINT msg, WPARAM wParam, LPARAM lParam){	D3DXMATRIXA16 matProj ;	RECT rect ;	switch(msg)	{	case WM_SIZE :		GetClientRect(hWnd, &rect) ;		g_x = rect.right ;		g_y = rect.bottom ;		d3dpp.BackBufferWidth			= g_x ;		d3dpp.BackBufferHeight			= g_y ;		g_pd3dDevice->Reset(&d3dpp) ;		g_pd3dDevice->SetRenderState(D3DRS_CULLMODE, D3DCULL_NONE) ;		// turn off D3D lighting		g_pd3dDevice->SetRenderState(D3DRS_LIGHTING, FALSE) ;		// turn on the zbuffer		g_pd3dDevice->SetRenderState(D3DRS_ZENABLE, TRUE) ;			D3DXMatrixPerspectiveFovLH(&matProj, D3DX_PI / 4, 1.0f, 1.0f, 100.0f) ;		g_pd3dDevice->SetTransform(D3DTS_PROJECTION, &matProj) ;		return 0 ;	case WM_DESTROY :		Cleanup() ;		PostQuitMessage(0) ;		return 0 ;	}	return DefWindowProc(hWnd, msg, wParam, lParam) ;}


[Edited by - Coder on October 24, 2004 12:26:49 PM]
You might want to place some source /source tags around that code to make it easier to read. Here's some ideas, maybe they'll be useful.

Instead of using the WM_SIZE message in the WndProc or WinMain functions, I'd suggest creation a function similar to ProcessMessage(MSG Message) in your game engine. You'll need it for a lot more than just WM_SIZE eventually. You can then just put a call to this function in the WndProc message handler, passing the message to your game engine first.

Now, we're in the game engine ProcessMessage function and we case out WM_SIZE. Here's a code snipet from my engine.

ProcessMessage(MSG fpMessage){...case WM_SIZE:      if (pGraphicsManager && (fpMessage.wParam == SIZE_MAXIMIZED) &&           bGoFullScreenOnMaximize)        pGraphicsManager->SetFullScreen(true);      else        if (pGraphicsManager)          pGraphicsManager->ResizeWindow();}void ResizeWindow(void){  RECT rClientArea;  // hWindowHandle is passed in to the creation of the Graphics Manager  GetClientRect(hWindowHandle, &rClientArea);  uiWidth = rClientArea.right - rClientArea.left;  uiHeight = rClientArea.bottom - rClientArea.top;  ppPresentParameters.BackBufferWidth = uiWidth;  ppPresentParameters.BackBufferHeight = uiHeight;  pInterface->GetAdapterDisplayMode(uiCurrentAdapter, &dmDisplayMode);  pViewManager->GetActivePerspective()->SetFovLH(PI/4*((float)uiHeight/dmDisplayMode.Height), ((float)uiWidth)/uiHeight,                                                  1.0f, 1000.0f);  BeginReset();}


If the max button is pressed (which sends a WM_SIZE event) and if the game engine is set up to toggle to full screen on a maximize, then we go to full screen. Otherwise, we resize the window.

When resizing, I get the new size of the window, then store the values, replace the values in the backbuffer, and change the FOV. My calculation here of PI/4*((float)...) is so that my FOV will change with the screen resolution based on the max resolution currently available. If the window gets bigger, changing the FOV to be proportional allows the view to "spread out" as the window gets wider, so you can see more around. This may not be what you want to do. Finally, I call BeginReset which sets in motion a destruction of all objects that I manage. It also sets a flag "bResetting" so that the next time I enter the Render loop, I check this flag, and if it's set I call EndReset, and otherwise I render as usual. See the code below:

bool cGraphicsManager::EndReset(void){  bool ReturnVal = true;  // Attempt to reset the Device  if (pDevice->TestCooperativeLevel() == D3D_OK)    pDevice->Reset(&ppPresentParameters);  else    return false;  // End the resets  pMeshManager->EndReset();  pTextureManager->EndReset();  pCursorManager->EndReset();  ...  if(ppPresentParameters.Windowed)  {    RECT rc;    const DWORD dwStyle = WS_CAPTION | WS_MINIMIZEBOX | WS_SYSMENU;    rc.left = rc.top = 0;    rc.right = uiWidth;    rc.bottom = uiHeight;    rc.right -= rc.left;    rc.bottom -= rc.top;    DWORD XWidth = GetSystemMetrics(SM_CXSCREEN);    DWORD YHeight = GetSystemMetrics(SM_CYSCREEN);    ShowCursor(TRUE);  }  // Change state to Normal Render  if (ReturnVal)    bResetting = false;  return ReturnVal;}


Hope this helps,
Chris
Chris ByersMicrosoft DirectX MVP - 2005
Lots of helpful info in there. I guess the code is fairly similar to mine.

I am calling the Reset function improperly. I know this because I tried a

if(FAILED(device->Reset())
MessageBox(xxxx)

which ended up showing the message box.

From my code above I use d3dpool textures and I don't have fonts or anything that need to be reseted.

I appreciate all the help, but
I'm about to give up on this and just ignore it. I know in OpenGL this is done for you, you don't have to worry about it. I think another reason for the problem is because the tutorials were done using D3DX, and the code other people have used to explain their solution doesn't seem to be in D3DX(make difference?).

I'll try a couple more things.

Edit : One thing i did try is g_pVB->Release() ;, I then called the function that recreates it, but with no luck.

The screen is black

[Edited by - Toonkides on October 24, 2004 4:47:01 PM]
Your vertex buffer was not managed, so you have to release it and recreate it. That would go as follows.

g_vb->Release();
g_pd3dDevice->Reset();
Create(g_vb);

Is this how you did it? And it made the screen black?

Chris
Chris ByersMicrosoft DirectX MVP - 2005
Aye, that's exactly what did it.

I released the VB then the g_pd3dDevice.

Now the create of the VB in the tutorial was from the InitGeometry() function which has the code :

HRESULT InitGeometry(){	// use D3DX to create a texture from a file based image	if(FAILED(D3DXCreateTextureFromFile(g_pd3dDevice, "hovigTex.bmp", &g_pTexture)))		return E_FAIL ;	// create the vertex buffer	if(FAILED(g_pd3dDevice->CreateVertexBuffer(100 * sizeof(CUSTOMVERTEX), 0,		D3DFVF_CUSTOMVERTEX, D3DPOOL_DEFAULT, &g_pVB, NULL)))		return E_FAIL ;	// fill the vertex buffer.  We are algorithmically generating a cylinder	// here, including the normals, which are used for lighting	CUSTOMVERTEX* pVertices ;	if(FAILED(g_pVB->Lock(0, 0, (void**)&pVertices, 0)))		return E_FAIL ;	for(DWORD i = 0; i < 50; i++)	{		FLOAT theta = (2 * D3DX_PI * i) / 49 ;		pVertices[2 * i + 0].position	= D3DXVECTOR3(sinf(theta), -1.0f, cosf(theta)) ;		pVertices[2 * i + 0].color		= 0xffffffff ;#ifndef SHOW_HOW_TO_USE_TCI		pVertices[2 * i + 0].tu			= FLOAT(i) / (49) ;		pVertices[2 * i + 0].tv			= 1.0f ;#endif		pVertices[2 * i + 1].position	= D3DXVECTOR3(sinf(theta), 1.0f, cosf(theta)) ;		pVertices[2 * i + 1].color		= 0xff808080 ;#ifndef SHOW_HOW_TO_USE_TCI		pVertices[2 * i + 1].tu			= FLOAT(i) / (49) ;		pVertices[2 * i + 1].tv			= 0.0f ;#endif	}	g_pVB->Unlock() ;	return S_OK ;}


Tried that didn't work. Do I need to do something with the device when I reset it?

I even tried releasing the texture and recreating it, still a black screen. I tried looking for DX programs, but they sure do a good job of making sure no one knows how they deal with the resizing of stuff=(
TIA

This topic is closed to new replies.

Advertisement