Shadows problems

Started by
8 comments, last by DragonJoker 7 years, 4 months ago

Hey guys, I worked long time ago in shadows but I left it because I couldn't find the problem I still have. I've been following this tutorial and I adapted it on my project but when I render everything I'm having weird issues I don't know why. As you can see in this pic the shadow projection is well calculated from the shader etc but you also can see that the back and front faces of both cubes are getting also weird shadow and also the shadow is repeated among the "terrain". This last problem is because of GL_REPEAT but if I use others parameters to create the texture it's worst... I tried using GL_CULLFACE to remove the weird shadows in the cubes but doesn't work. I'll leave the code here:



#include "../stdafx.h"
#include "Main.h"
#include "../Renderer/Renderer.h"
#include "../GLUtils/GLUtils.h"
#include "../Camera/Camera.h"
#include "../Audio/Audio.h"
#include "../Physics/Physics.h"
#include "../Skybox/Skybox.h"
#include "../Lighting/LightManager.h"
#include "../Entity/Entity.h"
#include "../Editor/Editor.h"
#include "../NativeGeometry/NativeGeometry.h"
#include "../Textures/TextureManager.h"
#include "../RenderTargets/RenderTargets.h"
#include "../ParticleSystem/ParticleSystem.h"
#include "../CEV/CEV.h"
#include "../Timers/Timers.h"
#include "../LUA/LuaManager.h"
#include "../LUA/EventManager.h"
#include "../LUA/EventArguments.h"
#include "../ColShapes/ColShapes.h"
#include "../KeyCode/Keycode.h"


int WINAPI WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPSTR lpCmdLine, int nShowCmd)
{
	//CoInitializeEx(nullptr, COINIT_MULTITHREADED);  //wtf is this xd

	Console.reset(new CConsole());
	Console->AllocateConsole();

	Renderer = std::make_unique<CRenderer>();
	ColShapeManager = std::make_unique<CColShapeManager>();
	Physics = std::make_unique<CPhysics>();
	EventArguments = std::make_unique<CEventsArguments>();
	LuaManager = std::make_unique<CLuaManager>();
	EventManager = std::make_unique<CEventManager>();
	KeyCode = std::make_unique<CKeyCode>();
	RenderTargetManager = std::make_unique<CRenderTargetManager>();
	TextureManager = std::make_unique<CTextureManager>();
	EntityManager = std::make_unique<CEntityManager>();
	FogManager = std::make_unique<CFogManager>();
	LightManager = std::make_unique<CLightManager>();
	Skybox = std::make_unique<CSkybox>();
	Camera = std::make_unique<CCamera>(glm::vec3(10, 10, 10));
	CEAudio = std::make_unique<CAEngine>();
	GLUtils = std::make_unique<CGLUtils>();

	if (!Renderer->CreateGLWindow(1920 / 2 - (1280 / 3) + 200, 1080 / 2 - (800 / 2), 1600, 1200, 16, false)) 
		return 0;

	Renderer.release();
	ColShapeManager.release();
	Physics.release();
	EventArguments.release();
	LuaManager.release();
	EventManager.release();
	KeyCode.release();
	RenderTargetManager.release();
	TextureManager.release();
	EntityManager.release();
	FogManager.release();
	LightManager.release();
	Skybox.release();
	Camera.release();
	CEAudio.release();
	GLUtils.release();

	return 0;
}

bool first = true;

BOOL DrawGLScene()
{
	glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT | GL_STENCIL_BUFFER_BIT);

	static GLuint depthMapFBO;
	static GLuint SHADOW_WIDTH = 2048, SHADOW_HEIGHT = 2048;
	static GLuint depthMap;
	if (first)
	{
		glGenFramebuffers(1, &depthMapFBO);

		glGenTextures(1, &depthMap);
		glBindTexture(GL_TEXTURE_2D, depthMap);
		glTexImage2D(GL_TEXTURE_2D, 0, GL_DEPTH_COMPONENT, SHADOW_WIDTH, SHADOW_HEIGHT, 0, GL_DEPTH_COMPONENT, GL_FLOAT, NULL);
		glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
		glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
		glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_REPEAT);
		glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_REPEAT);

		glBindFramebuffer(GL_FRAMEBUFFER, depthMapFBO);
		glFramebufferTexture2D(GL_FRAMEBUFFER, GL_DEPTH_ATTACHMENT, GL_TEXTURE_2D, depthMap, 0);
		glDrawBuffer(GL_NONE);
		glReadBuffer(GL_NONE);
		glBindFramebuffer(GL_FRAMEBUFFER, 0);

		first = false;
	}

	//SKYBOX RENDERING, LIGHT, FOG, CAMERA UPDATE & VIEWMATRIX UPDATE
	Renderer->MainProgram.Start();
	Camera->Update();
	Renderer->SetViewMatrices();
	Skybox->RenderSkybox();
	LightManager->SetUpAmbientColor();
	LightManager->UpdateLightsInShader(Renderer->MainProgram);
	FogManager->RenderFog();

	for (int i = 0; i < RenderTargetManager->GetRenderTargets(); i++)
	{
		CRenderTarget* RT = RenderTargetManager->GetRenderTarget(i);
		if (RT)
		{
			RT->Bind();
			glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
			RT->UpdateTrans();
			Skybox->RenderSkybox();
			RenderTargetManager->BackToNormalRenderer();
		}
	}

	//SHADOWS
	static float xx = 0, yy = 3.f, zz = 0;
	if (GetAsyncKeyState(VK_NUMPAD8))
	{
		xx += 0.1f;
	}
	if (GetAsyncKeyState(VK_NUMPAD2))
	{
		xx -= 0.1f;
	}
	if (GetAsyncKeyState(VK_NUMPAD7))
	{
		yy += 0.1f;
	}
	if (GetAsyncKeyState(VK_NUMPAD1))
	{
		yy -= 0.1f;
	}
	if (GetAsyncKeyState(VK_NUMPAD9))
	{
		zz += 0.1f;
	}
	if (GetAsyncKeyState(VK_NUMPAD3))
	{
		zz -= 0.1f;
	}
	glm::mat4 mPROJ = glm::ortho<float>(-10.0f, 10.0f, -10.0f, 10.0f, 0.05f, 50.f);
	glm::mat4 lightView = glm::lookAt(glm::vec3(5.f, 5.f, 5.f), glm::vec3(0.0f, 0.0f, 0.0f), glm::vec3(0.0f, 1.0f, 0.0f));
	glm::mat4 biasMatrix(
		0.5, 0.0, 0.0, 0.0,
		0.0, 0.5, 0.0, 0.0,
		0.0, 0.0, 0.5, 0.0,
		0.5, 0.5, 0.5, 1.0
	);
	glm::mat4 lightSpaceMatrix = biasMatrix * mPROJ * lightView;

	Renderer->ShadowMapping.Start();
	Renderer->ShadowMapping.setUniform("depthMVP", lightSpaceMatrix);

	glm::mat4 mModelToCamera;
	glm::vec3 vPos = glm::vec3(xx, yy, zz);
	mModelToCamera = glm::translate(glm::mat4(1.0), vPos);
	mModelToCamera = glm::scale(mModelToCamera, glm::vec3(2.f));
	Renderer->ShadowMapping.setUniform("model", mModelToCamera);
	glViewport(0, 0, SHADOW_WIDTH, SHADOW_HEIGHT);
	glBindFramebuffer(GL_FRAMEBUFFER, depthMapFBO);
	/*glEnable(GL_CULL_FACE);
	glCullFace(GL_FRONT);*/
	glClear(GL_DEPTH_BUFFER_BIT);

	Environment::DrawCube(glm::vec3(xx, yy, zz), glm::vec3(2.f), glm::vec4(1.f), 7, true);

	vPos = glm::vec3(xx + 3.f, yy, zz);
	mModelToCamera = glm::translate(glm::mat4(1.0), vPos);
	mModelToCamera = glm::scale(mModelToCamera, glm::vec3(2.f));
	Renderer->ShadowMapping.setUniform("model", mModelToCamera);
	Environment::DrawCube(glm::vec3(xx + 3, yy, zz), glm::vec3(2.f), glm::vec4(1.f), 7, true);

	RenderTargetManager->BackToNormalRenderer();
	//glDisable(GL_CULL_FACE);
	Renderer->ShadowMapping.Stop();
	glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);

	Renderer->MainProgram.Start();
	Environment::DrawPlane(glm::vec3(0, 20.f, -20.f), glm::vec3(0.f, 90.f, 90.f), glm::vec3(10.f, 1.f, 10.f), depthMap);
	Renderer->MainProgram.setUniform("EngineMatrices.DepthBiasMVP", lightSpaceMatrix);
	Renderer->MainProgram.setUniform("shadowMap", 1);
	Renderer->MainProgram.setUniform("bShadowsOn", 1);
	glActiveTexture(GL_TEXTURE1);
	glBindTexture(GL_TEXTURE_2D, depthMap);
	glActiveTexture(GL_TEXTURE0);
	Environment::DrawCube(glm::vec3(xx, yy, zz), glm::vec3(2.f), glm::vec4(1.f), 7, false);
	Environment::DrawCube(glm::vec3(xx + 3.f, yy, zz), glm::vec3(2.f), glm::vec4(1.f), 7, false);

	//SHADOWS
	//SHADOWS


#ifdef ENGINE_EDITOR
	//CEditor::UpdateSelectedEntity();
#endif

	//MODEL TYPE RENDERING
	if (Renderer->IsCullFaceEnabled())
	{
		/*glEnable(GL_CULL_FACE);
		glCullFace(GL_FRONT);
		glFrontFace(GL_CW);*/
	}

	auto RenderEntity = [](Entity* _Entity)
	{
		if (_Entity->UpdateTransform())
		{
			_Entity->StartShader();
			_Entity->Mesh->Render();
			_Entity->StopShader();
		}
		if (_Entity->ShowCollider)
		{
			btVector3 one, two;
			_Entity->Phys->RigidBody->getAabb(one, two);
			Environment::DrawPoint(glm::vec3(one.x(), one.y(), one.z()), glm::vec3(2.f), glm::vec3(0.f, 1.f, 0.f));
			Environment::DrawPoint(glm::vec3(one.x() + (two.x() - one.x()), one.y(), one.z()), glm::vec3(2.f), glm::vec3(0.f, 1.f, 0.f));
			Environment::DrawPoint(glm::vec3(one.x(), one.y(), one.z() + (two.z() - one.z())), glm::vec3(2.f), glm::vec3(0.f, 1.f, 0.f));
			Environment::DrawPoint(glm::vec3(one.x(), one.y() + (two.y() - one.y()), one.z()), glm::vec3(2.f), glm::vec3(0.f, 1.f, 0.f));
			Environment::DrawPoint(glm::vec3(two.x() - (two.x() - one.x()), two.y(), two.z()), glm::vec3(2.f), glm::vec3(0.f, 1.f, 0.f));
			Environment::DrawPoint(glm::vec3(two.x(), two.y() - (two.y() - one.y()), two.z()), glm::vec3(2.f), glm::vec3(0.f, 1.f, 0.f));
			Environment::DrawPoint(glm::vec3(two.x(), two.y(), two.z() - (two.z() - one.z())), glm::vec3(2.f), glm::vec3(0.f, 1.f, 0.f));
			Environment::DrawPoint(glm::vec3(two.x(), two.y(), two.z()), glm::vec3(2.f), glm::vec3(0.f, 1.f, 0.f));
		}
	};

	//COL SHAPES
#ifdef ENGINE_EDITOR
	for (uint i = 0; i < ColShapeManager->GetShapes(SHAPE_RECTANGLE); i++)
	{
		RectangleCol* currentShape = ColShapeManager->GetShape<RectangleCol>(SHAPE_RECTANGLE, i);
		if (currentShape && currentShape->IsDebugEnabled())
			currentShape->DebugShowCol();
	}

	for (uint i = 0; i < ColShapeManager->GetShapes(SHAPE_SPHERE); i++)
	{
		SphereCol* currentShape = ColShapeManager->GetShape<SphereCol>(SHAPE_SPHERE, i);
		if (currentShape && currentShape->IsDebugEnabled())
			currentShape->DebugShowCol();
	}
#endif

	//ENTITIES
	std::map<float, Entity*> TransparentEntities;
	std::map<float, Entity*> FarestEntities;

	EntityManager->LoadQueueEntities();
#ifdef ENGINE_EDITOR
	EntityManager->ResetEntitiesRendered();
#endif

	for (uint i = 0; i < EntityManager->GetEntities(); i++)
	{
		Entity* currentEntity = EntityManager->GetEntity(i);
		float EntityToCamera = glm::distance(Camera->Position, currentEntity->Position);
		if (currentEntity && currentEntity->CanRender() && EntityToCamera <= Renderer->GetRenderMaxDistance() && currentEntity->IsVisible())
		{
			if (currentEntity->IsDestroying())
			{
				float _alpha = currentEntity->GetAlpha();
				if (_alpha <= 0.f)
				{
					EntityManager->DestroyEntity(currentEntity);
					continue;
				}
				currentEntity->SetAlpha(_alpha - 0.01f);
			}
#ifdef ENGINE_EDITOR
			EntityManager->IncreaseEntitiesRendered();
#endif
			if (currentEntity->GetAlpha() < 1.f)
				TransparentEntities[EntityToCamera] = currentEntity;
			else
				FarestEntities[EntityToCamera] = currentEntity;
		}
	}

	//RENDER TRANSPARENT ENTITIES & FAREST ENTITIES FIRST (lol?)

	for (auto it = FarestEntities.begin(); it != FarestEntities.end(); ++it)
		RenderEntity(it->second);

	for (auto it = TransparentEntities.rbegin(); it != TransparentEntities.rend(); ++it)
		RenderEntity(it->second);

	Renderer->MainProgram.setUniform("bShadowsOn", 0);

	TransparentEntities.empty();
	TransparentEntities.clear();

	//PARTICLE SYSTEM
	for (uint i = 0; i < ParticleSystemManager::GetParticleSystems(); i++)
	{
		if (ParticleSystems[i] && ParticleSystems[i]->ID != INT_MAX)
		{
			ParticleSystems[i]->SetMatrices(&Renderer->GetProjectionMatrix(), Camera->Position, Camera->vView, Camera->vUp);
			ParticleSystems[i]->UpdateParticles(0.025f);
			ParticleSystems[i]->RenderParticles();
		} 
	}

	Renderer->ResetColor();
	Renderer->MainProgram.Stop();

	if (Renderer->IsCullFaceEnabled())
		glDisable(GL_CULL_FACE);

	OGLImGui::DrawGUI();

	EventManager->CallEvent(ON_RENDER, EventArguments->GetEventArguments(ON_RENDER));

	return true;
}

Sorry for the messed code but I use to do that when I'm testing something instead of organizing it first. Maybe I'm missing something stupid but I can't figure out lol. I appreciate any help. Thanks in advance.

Advertisement

show us the code of all shadowmapping shaders

i believe you are not clamping the projection. from light into the direction of light but its just a thought i don't even know if you are using directionla 'light' for that. everything that falls outside of light projection stays "not shadowed"

Thanks for the reply WiredCat! It's directional light as you can see in the tutorial and also in my code where I transform the matrices using the projection matrix and the light direction matrix too. I've fixed the clamping and I know why it's repeated along the terrain because the terrain it's supposed to be a cube originally modelled with (1, 1, 1) scale and I scale it manually runtime to (100, 1, 100), I've tested with others big models and it doesn't repeat. The last problem is that I can't achieve properly the shadow for the entities that are casting the shadows. Take a look at this and this. You can see that it's not properly projected into the entities and I don't know what's happening. There is the shaders I to project:

MainFrag.frag:


#version 430
#extension GL_ARB_shader_storage_buffer_object : enable

smooth in vec2 texCoord;
smooth in vec3 vNormal;
smooth in vec3 vWorldPos;
smooth in vec4 vEyeSpacePos;
smooth in vec4 ShadowCoord;

uniform bool ison = false;
uniform int bShadowsOn;

uniform vec4 vColor;
uniform vec4 vAmbientColor = vec4(1.0);
uniform vec3 vEyePosition;
uniform sampler2D gSampler;
uniform sampler2D gNormalMap;
uniform sampler2D shadowMap;
uniform bool bUseOnlyColor = false;
uniform bool bSkybox = false;
uniform sampler2DShadow testt;

uniform int bEnableBumpMap; 
in mat3 vLightTangent;

#include "Directional_light.frag"
#include "Point_light.frag"
#include "Spot_light.frag"
#include "Fog.frag"
#include "Plasma.frag"
#include "Text_1.frag"
#include "Water.frag"
#include "RoomEditor.frag"


/*vec2 poissonDisk[16] = vec2[]( 
   vec2( -0.94201624, -0.39906216 ), 
   vec2( 0.94558609, -0.76890725 ), 
   vec2( -0.094184101, -0.92938870 ), 
   vec2( 0.34495938, 0.29387760 ), 
   vec2( -0.91588581, 0.45771432 ), 
   vec2( -0.81544232, -0.87912464 ), 
   vec2( -0.38277543, 0.27676845 ), 
   vec2( 0.97484398, 0.75648379 ), 
   vec2( 0.44323325, -0.97511554 ), 
   vec2( 0.53742981, -0.47373420 ), 
   vec2( -0.26496911, -0.41893023 ), 
   vec2( 0.79197514, 0.19090188 ), 
   vec2( -0.24188840, 0.99706507 ), 
   vec2( -0.81409955, 0.91437590 ), 
   vec2( 0.19984126, 0.78641367 ), 
   vec2( 0.14383161, -0.14100790 ) 
);

float GetVisibility()
{
    if (bShadowsOn == 0)
		return 1.0;

    float visibility = 1.0;
    float bias = 0.000;

    for (int i = 0; i < 4; i++)
    {
      int index = i;
		  vec4 vShadowSmooth = vec4(ShadowCoord.x + poissonDisk[index].x/700.0, ShadowCoord.y + poissonDisk[index].y/700.0, (ShadowCoord.z-bias)/ShadowCoord.w, 1.0);
		vShadowSmooth = vShadowSmooth * 0.5 + 0.5;
		vShadowSmooth.w = 1.0;
		  float fSub = texture(shadowMap, vShadowSmooth.xy).r; 
		  visibility -= 0.25 * (1.0 - fSub);
    }
    return visibility;
}*/

float ShadowCalculation(vec4 fragPosLightSpace)
{
    vec3 projCoords = fragPosLightSpace.xyz / fragPosLightSpace.w;
    projCoords = projCoords * 0.5 + 0.5;
    float closestDepth = texture(shadowMap, projCoords.xy).r; 
    float currentDepth = projCoords.z;
    float shadow = currentDepth > closestDepth ? 1.0 : 0.0;

    if (projCoords.z > 1.0)
        return 0.0;

    return shadow * 0.6;
}


void main()
{
	vec4 vTexColor = texture2D(gSampler, texCoord);
	vec3 vFragColor;
	vec3 vNormalized = normalize(vNormal);
	if (!bUseOnlyColor)
		gl_FragColor = vTexColor * vColor * vAmbientColor;				  //texture without any kind of effect applied (lights etc.)
	else
		gl_FragColor = vColor * vAmbientColor;							  //texture without any kind of effect applied (lights etc.)

	vFragColor = gl_FragColor.xyz;

	//ENGINE SHADERS, will change gl_FragColor if the entity has the shader enabled

	PlasmaShader(texCoord);                    //Plasma.frag
	Text_1Shader(texCoord);                    //Text_1.frag
	Water_Shader(texCoord, vColor.w);          //Water.frag
	Arrow_Shader(texCoord);					   //RoomEditor.frag

	//LIGHTING
	if (iDirectionalLights > 0)
	{
		vec4 Diffuse_DirectionalLightsFinalColor = vec4(1.0);
		vec4 Specular_DirectionalLightsFinalColor = vec4(1.0);
		vec4 Normal_DirectionalLightsFinalColor = vec4(1.0);
		for (int i = 0; i < iDirectionalLights; i++)
		{
			if (DirectionalLights.Lights[i].bEnabled == 1 && !bSkybox)
			{
				Diffuse_DirectionalLightsFinalColor += GetDirectionalLightColor(DirectionalLights.Lights[i], vNormalized);
				Specular_DirectionalLightsFinalColor += GetDirectionalLightSpecularColor(vWorldPos, vEyePosition, DirectionalLights.Lights[i], vNormalized);
				
				if (bEnableBumpMap == 1)
					Normal_DirectionalLightsFinalColor += GetDirectionalLightNormalColor(DirectionalLights.Lights[i], vLightTangent, texCoord, gNormalMap);
			}
		}
		gl_FragColor *= Diffuse_DirectionalLightsFinalColor * Specular_DirectionalLightsFinalColor * Normal_DirectionalLightsFinalColor;
	}
	if (iPointLights > 0)
	{
		vec4 Diffuse_PointLightsFinalColor = vec4(1.0);
		vec4 Specular_PointLightsFinalColor = vec4(1.0);
		for (int i = 0; i < iPointLights; i++)
		{
			if (PointLights.Lights[i].bEnabled == 1 && !bSkybox)
			{
				Diffuse_PointLightsFinalColor += GetPointLightColor(PointLights.Lights[i], vWorldPos, vNormalized);
				Specular_PointLightsFinalColor += GetPointLightSpecularColor(vWorldPos, vEyePosition, PointLights.Lights[i], vNormalized);
			}
		}
		gl_FragColor *= Diffuse_PointLightsFinalColor * Specular_PointLightsFinalColor;  // vec4(pow(vFragColor, vec3(1.0/2.2)), 1.0)
		
	}
	if (iSpotLights > 0)
	{
		vec4 SpotLightsFinalColor = vec4(1.0);
		for (int i = 0; i < iSpotLights; i++)
		{
			if (SpotLights.Lights[i].bEnabled == 1)
			{
				SpotLightsFinalColor += GetSpotLightColor(SpotLights.Lights[i], vWorldPos);
			}
		}
		gl_FragColor *= SpotLightsFinalColor;
	}

	if (bShadowsOn == 1)
	{
		//gl_FragColor *= GetVisibility();
		
		gl_FragColor *= 1.0 - ShadowCalculation(ShadowCoord);
		gl_FragColor.w = 1.0;
	}

	if (FogData.Enabled)
		gl_FragColor = mix(gl_FragColor, FogData.FogColor, GetFogFactor(abs(vEyeSpacePos.z / vEyeSpacePos.w))); 
}





ShadowMap.vert:


#version 330 core

layout(location = 0) in vec3 inPous;

uniform mat4 depthMVP;
uniform mat4 model;

void main()
{
	gl_Position = depthMVP * model * vec4(inPous, 1.0);
}


ShadowFrag.frag:


#version 330 core

// Ouput data
layout(location = 0) out float fragmentdepth;


void main()
{
	//fragmentdepth = gl_FragCoord.z;
}

These are all the shaders I use and everything related to shadows it's in the first code I sent in the main post so you can compare. I think I'm missing something in my C++ main code because as you can see in the picture the shadow is applied triangle by triangle and that should be something I'm missing not sure. Also in the commented code of the MainFrag you can see GetVisibility function which I tested and it works like the one I'm using but with smooth only, the shadow it's applied to the whole entity that casts the shadow and also happens the same error that there are parts in the entity that have no shadow in the borders...

mhm lets say i didint check the code.

first image shows that maybe you have depth precision problems, that means either bias it out by adding small number to the

tested fragment so actual_fragment_depth+0.0005 > shadow_depth_fragment then is in shadow might fix it.

but i define point light as something like. position, radius, color, direction, and i fit the frustum of the light as follows:

i define z far value for shadowmap, lets do not care for now for anything like torch lights etc. lets assume we have just directional light as sunlight. you define just light direction, color and z_far (the same one you define when you draw your shadowmap)

then

if fragment is in shadow frustum then


float actual_fragment_depth = distance(LightPOS, world_vertex_pos) / z_far;

after that


if (actual_fragment_depth > shadow_depth)
color = vec4(0.0, 0.0, 0.0, 1.0);
else
color = vec4(1.0, 1.0, 1.0, 0.0);

and second image

you need to add this code too you see there resolution precision problems, but theres an easy fix for that das dot product


if (dot( normalize( vectorAB(LightPOS, world_vertex_pos) ), VertexNormal) >= 0.0)
		color = vec4(0.0, 0.0, 0.0, 1.0);

Did i forget to mention that you need to define vertice normals too, so each polygon has its own individual normal facing outwards the box center.

That should do, cheers/

Thanks WiredCat! I've been the whole day working on this and struggling with the shaders and finally got it working taking into account your notes. My fear now is that maybe in the future when I use more complex models the bias I use will fuck up everything in shadows, I set the bias as 0.00005 which is pretty low but it's the one working for all models at the moment. Here's the result:

f99e5d516f1c7e546486f9c96aa7ed65.jpg

well, i dont use any bias anyway :P butone thing i can tell you that bias is connected to z_near/z_far values i cant remember exact connection between these.

Staying in topic, after I finished directional light I started with spot light since they are easy because they don't need a cube map like point lights. I've been programming the shadows and FBO's for it but I encounter a curious problem projection the shadow, In this pic you can see that there are some black lines but the real shadow of the cube it's properly projected, I need a way to remove those lines somehow, maybe you WiredCat know about this issue.

The shadow calculation function:


float ShadowCalculation()
{
    vec3 projCoords = ShadowCoord.xyz / ShadowCoord.w;
    vec2 texelSize = 1.0 / textureSize(shadowMap, 0);
    projCoords = projCoords * 0.5 + 0.5;
    float closestDepth = texture(shadowMap, projCoords.xy).r; 
    float currentDepth = projCoords.z; 
    float shadow = 1.0;

	for (int x = -1; x <= 4; x++)
		for (int y = -1; y <= 4; y++)
			shadow += currentDepth > (texture(shadowMap, projCoords.xy + vec2(x, y) * texelSize).r) ? 1.0 : 0.0;

	shadow *= 0.01;

	shadow = currentDepth > closestDepth ? shadow : 0.0;

	if (dot(normalize(vWorldPos - LightPos), normalize(vNormal)) >= 0.0)
	{
		BackFaceShadow = true;
		return 0.0;
	}

	if (projCoords.z <= 0.0 || projCoords.z >= 1.0 || projCoords.x >= 1.0 || projCoords.y >= 1.0 || projCoords.x <= 0.0 || projCoords.y <= 0.0)
		return 0.0;
	
	BackFaceShadow = false;
    return shadow * 0.8;
}

I added the


if (projCoords.z <= 0.0 || projCoords.z >= 1.0 || projCoords.x >= 1.0 || projCoords.y >= 1.0 || projCoords.x <= 0.0 || projCoords.y <= 0.0)

to remove the shadow outside of the perpective projection by the spot light. And where I call it:


if (bShadowsOn == 1)
{
	ShadowFactor = 1.0 - ShadowCalculation();
	if (BackFaceShadow)
		FinalColor *= vec4(vec3(length(cross(normalize(vWorldPos - LightPos), vNormal))), 1.0);
	else
		FinalColor *= vec4(vec3(ShadowFactor), 1.0);
}

I also added the line when BackFaceShadow it's true to make progressive color changing to avoid faces change directly from shadow to light.

Looks like traditional shadow acne, to me?

If you can't find anything, look for something else.

You're right I added some bias to remove the acne and now it works, do you know any article or website to fix the peter-panning without having to render the front faces? I'm modelling something and it doesn't have front faces due to remove useless geometry, the model is this and from this is from other perspective. At the moment I have no idea if this is possible to achieve but many games do that, for example GTA5 has a terrain with no front faces and the shadows of the hills are perfect, maybe this is more complex but I want to at least try or read how it's achieved.

I solved the Peter Panning by virtually translating the world space position along its normal, instead of applying a depth bias

If you can't find anything, look for something else.

This topic is closed to new replies.

Advertisement