Implementation: First Person and Orbit Cameras

Overview 

An OpenGL application uses specialized cameras navigate a terrain in 3D scenes or wade thru sections of a complex structure. Some of such implementations are First Person Shooter (FPS) Camera or Orbit Camera or ARC ball Camera.

Details

System  class diagram

New Camera classes derived from DualProjectionCamera are defined. They are FPSCamera and OrbitCamera,

A new mesh class  QuadMesh is defined for loading vertexdata of 3D objects such as walls, ceilings etc in a scene. It's used by another  Quad object for loading vertex data to vertex shader and texture data to the vertex and fragment shaders. 

FPSCamera

FPSCamera class is derived from DualProjectionCamera class. This camera captures objects from the viewpoint of a viewer. Some aspects have to be considered, like the characteristics of the camera (orbiting with the mouse and translation with keyboard keys).

The camera has the following characteristics:

Orbit: The character can look left, right, up, and down; however, if we imagine the character’s head, it can’t be tilted.

Translation: The character can move in four directions: forward, backward, left, and right, up and down. 

It process mouse and keyboard inputs and calculates forward, backward, left, and right, up and down movements. They are updated into MatrixData  objects.

Constructor

Initializes hwnd with the Window handle of the hosting window. Internally calls DualProjectionCamera constructor.

Members

Name	Description
camera_front	contains front of the camera which gets updated when the viewer moves forward or backward or sideways.
camera_pos	contains current position based on the movements of the viewer.
camera_up	contains front of the camera which gets updated when the viewer moves up or down.
mouse_sensitivity	Sensitivity of the mouse to calculate finer movements of the viewer.
velocity	Speed to calculate finer movements of the viewer.

Methods

Name	Description
OnKey	This method handles keyboard input events and updates camera_pos as below: W,S : Viewer Move forward or backward A,D:Viewer Move Left  or right R,F:Viewer Move Up or Down
OnMouseMove	Updates camera_pos based on the mouse movements.
setSenseivity	Sets mouse_sensitivity and velocity values
updateViewMatrix	Sets starting values of camera_pos, camera_front and camera_up values.

OrbitCamera

OrbitCamera class is derived from DualProjectionCamera class. 

An orbit camera moves in a circle around a central object. The camera always looks at the center of the object. The viewer moves around the object in a circular path while keeping the focus fixed on that center point.  To calculate polar coordinates and spherical coordinates are used.

A polar coordinate represents a point in a 2D space as radius and angle. The point lies on the circumference where the radius is computed from the center point and the angle between the point and  the polar axis.

As shown below, the blue line represents polar axis. r represents the radius.

A spherical coordinate system specifies a given point in three-dimensional space by using a distance and two angles as its three coordinates. 

These are radius r,  the polar angle, θ, and the azimuthal angle, φ, which is the angle of rotation of the radial line around the polar axis.

The three coordinates (r, θ, φ), provide a coordinate system on a sphere, typically called the spherical polar coordinates. 

Given sphere with center point c and [r, θ, φ], the corresponding [x, y, z] coordinates are:

x = c.x + r * cos(θ) * cos(φ)

y = c.y + r * sin(θ)

z = c.z + r * cos(θ) * sin(φ)

They are updated into MatrixData  objects.

Constructor

Initializes hwnd with the Window handle of the hosting window. Internally calls DualProjectionCamera constructor.

Members

Name	Description
_minRadius	The minimum value of radius set during zoom operation.
azimuthAngle_	contains the value of current azimuth angle.
center_	The center defaulted to 0,0,0.
polarAngle_	contains the value of current polar angle.
radius	The current value of the radius.
upVector_	The up vector defaulted to 0,1,0.
zoomfactor	The increments of the zoom both ways defaulted to 0.5
mouse_sensitivity	Sensitivity of the mouse to calculate finer movements of the viewer.

Methods

Name	Description
OnKey	This method handles keyboard input events and updates radius as below: Left Arrow : Makes radius bigger Right Arrow : Makes radius smaller
OnMouseMove	Left Button: Rotates all the objects Right Button: Changes Azimuth and Polar angles resulting in circular motion of the objects Middle Button: Changes center point resulting in panning motion of the objects horizontally and vertically.
geteye	Returns current viewer position
getNormalizedViewVector	Returns normalized value of  (center_ - getEye())
getUpVector	Returns upVector_
getViewPoint	Returns center_
moveHorizontal	Pans entire scene horizontally as a result of mouse movement
moveVertical	Pans entire scene vertically as a result of mouse movement
rotateAzimuth	Rotates entire scene horizontally as a result of mouse movement
rotatePolar	Rotates entire scene vertically as a result of mouse movement
updateViewMatrix	Sets view matrix data based on current positions of the eye, center_ and upVector_.
zoom	Sets fov based on keyboard input resuting in zoom out and zoom in actions.

Implementation: Importing WaveFront OBJ Models

Overview 

As discussed before, WaveFront OBJ and MTL files needs to be parsed to get geometry, Material information. A new mesh class and new geometry object are required for this.

Details

A new utility class  GenericObjParser is defined for loading OBJ and MTL files and generating VertexData, Material and texture information. A model can consists of 1 or more Mesh. Each mesh is associated with 1 material which in turn is associated with 0 or 1 texture. Each MeshInfo object represents a mesh which contains Vertex Data consisting of vertices.

System  class diagram

The GenericParser is an abstract class. There are derived two derived classes WFObjParser and AssetImpParser to implement parsing functionality.

A new mesh class  GenericObjMesh is defined for loading vertexdata, Material and texture data. It's used by another new class  GenericObj object for loading vertex data to vertex shader and texture data to the vertex and fragment shaders. Note that  a texture is not mandatory. A GenericObj  is created for each MeshInfo object generated by the GenericParser derivative.

MeshInfo

MeshInfo is an utility class that holds vertex data of the various meshes defined by the OBJ file.

GenericParser

GenericParser the base class for parser classes that parses OBJ file and MTL files to provide geometry  and material color information for rendering. It supplies vertices, texture coordinates, normals, Ambient, Diffuse, Specular Colors and Texture information.

WFObjParser 

WFObjParser derives from GenericParser class. It parses OBJ file and MTL files to provide geometry  and material color information for rendering. It supplies vertices, texture coordinates, normals, Ambient, Diffuse, Specular Colors and Texture information.

AssetImpParser

AssetImpParser derives from GenericParser class. It parses OBJ file and MTL files to provide geometry  and material color information for rendering. It supplies vertices, texture coordinates, normals, Ambient, Diffuse, Specular Colors and Texture information. Internally it uses AssetImport library.

GenericObjMesh

This class derives from IGeometryMesh class. It contains information about Geometry information and Material color information obtained after parsing .obj and .mtl file.

GenericObj

GenericObj derives from BaseGeometry class. It renders objects after parsing obj and mtl files.

Output

Implementation: DualProjectionCamera Camera

Overview

In the previous post, we understood how view matrix, perspective matrix and orthographic matrix. Two new cameras are used to provide perspective and orthographic views.

Details

This post discusses camera  and camera data that are used in projections and animations.

The Camera data holds transformation information such as pitch, yaw and roll angles, translation, scaleby as well as Model, View and Projection matrix information of the 3D object. 

Camera and CameraData

There are many specialized camera classes that are associated with different types of camera data.

The class  diagram below shows their association, The light blue classes are camera data and the others are camera classes. 

The details will be provided in the next posts.

ViewMatrixData

ViewMatrixData class contains Model to World transformation information such as camera position, object position in the world, and the World up vector. They are used by DualProjectionCamera class for calling lookAt function to generate view matrix.

Members

Name	Description
Position	Camera position in world Coordinates.
Target	Position of the object in world Coordinates.
Up	Worlds Up vector.
V	Contains view Matrix

Methods

Name	Description
getViewMatrix	Returns latest View matrix.
setViewMatrix	Stores inputs such as positions, Target and up. Computes view matrix by internally calling lookAt function. Stores the result in V.

PerspectiveProjectionMatrixData

PerspectiveProjectionMatrixData class contains perspective view frustum information such as near plane, far plane and fov angle. Also the aspect ratio. They are used by by DualProjectionCamera class to redraw.

Members

Name	Description
AspectRatio	An aspect ratio is the proportional relationship between the width and height of a shape, image, or screen, expressed as width:height (e.g., 16:9). It defines an object's proportions rather than its physical size, ensuring visuals scale cleanly without distortion.
FOV	Field of view (FOV) is the angular extent of the observable world that can be seen or captured at any given moment. It dictates how wide or narrow your view is.
NearPlane	A near plane (often called the clipping or projection plane) is the imaginary flat boundary closest to the camera in a 3D perspective view. It acts as a "window" that captures the scene. Everything closer to the camera than this plane is invisible to prevent visual clipping and rendering errors.
FarPlane	A far plane (or far clipping plane) is a boundary defining the maximum distance a virtual camera can "see". Any object or geometry beyond this plane is clipped, meaning it is excluded from the rendered image
P	Contains latest Perspective Projection Matrix.

Methods

Name	Description
getProjectionMatrix	returns latest Perspective Projection Matrix stored in P
setProjectionMatrix	Stores inputs such as near plane and far plane and computes Perspective Projection Matrix. Stores the result in P
setFOV	Stores FOV angle after clipping the input is lower or higher than tolerance.  Later computes Perspective Projection Matrix. Stores the result in P
setAspectRatio	If the input is a float, aspect ratio is stored from the input parameter. If the input is an integer, the current FOV is doubled and sign is applied. typically used for zooming in or out. Later computes Perspective Projection Matrix. Stores the result in P

OrthographicProjectionMatrixData

OrthographicProjectionMatrixData class contains orthographic view frustum information such as xmin, xmax, ymin, ymax, zmin, zmax. They are used by by DualProjectionCamera class to redraw.

Members

Name	Description
XMinMax	xmin and xmax define the left and right clipping planes (boundaries) of the 3D viewing volume in an orthographic projection.
YMinMax	ymin and ymax define the left and right clipping planes (boundaries) of the 3D viewing volume in an orthographic projection.
ZMinMax	zmin and zmax define the nearest and farthest clipping planes (boundaries) of the 3D viewing volume in an orthographic projection.
P	Contains latest projection matrix.

Methods

Name	Description
getProjectionMatrix	returns latest Orthographic Projection Matrix stored in P
setProjectionMatrix	Stores inputs such as X,Y, Z min and max distance and computes Orthographic Projection Matrix. Stores the result in P

DualProjectionCamera

DualProjectionCamera class is derived from ThreeDCamera class. It process mouse and keyboard inputs and calculates zoom factor, view matrix, perspective and orthographic projection matrices. They are updated into ViewMatrixData, PerspectiveProjectionMatrixData, OrthographicProjectionMatrixData object.

Constructor

Initializes hwnd with the Window handle of the hosting window. Internally calls ThreeDCamera constructor.

Members

Name	Description
VM	contains 3D transformation information such as camera position, object position in the world, and the World up vector. . They are updated by based on mouse and keyboard inputs and are shared with 3D objects based on BaseGeometry class  to redraw.
PPM	contains perspective view frustum information such as near plane, far plane and fov angle. Also the aspect ratio. They are updated by based on mouse and keyboard inputs and are shared with 3D objects based on BaseGeometry class  to redraw Perspective projection.
OPM	contains orthographic view frustum information such as xmin, xmax, ymin, ymax, zmin, zmax.  They are updated by based on mouse and keyboard inputs and are shared with 3D objects based on BaseGeometry class  to redraw Orthographic projection.

Methods

Name	Description
CopyToClipboard	Copies Unproject() information to clipboard.
OnKey	This method handles keyboard input events as below: P  => Captures x,y,z coordinates at current mouse position  in the world space  and copies to clipboard. PageUp  => doubles current FOV resulting in Zooms Out effect or items on the screen becomes smaller.  This affects only Perspective projection. PageUp  => current FOV is halved resulting in Zooms in effect or items on the screen becomes bigger.  This affects only Perspective projection.
OnMouseWheel	This simulates ZoomOut when scroll wheel is moved up and ZoomIn when scroll wheel is moved down.
setOrthographicProjectionMatrix	Applies latest Orthographic Projection Matrix stored in OPM to the geometry object.
setPerspectiveProjectionMatrix	Applies latest Perspective Projection Matrix stored in PPM to the geometry object.
setViewMatrix	Applies latest View Matrix stored in VM to the geometry object.
Unproject	Internally called by keyboard handler to capture  x,y,z coordinates at the mouse cursor
UpdateWH	Calls BaseCamera's method to reset window sizes and sets aspect ratio.

Implementation: Drawing Text and Images

Overview 

Modern OpenGL does not support drawing text or Images so it needs to be handled independently. There many popular libraries available such as freetype to render text in a scene.

Details

In this post we shall discuss drawing text using GDI+ APIs. The plan is to  draw text and Images onto a memory based bitmap which is later loaded as a texture.

A new mesh class  TextMesh is defined for loading bitmap as texture. It's used by TextImageSketcher object for loading vertex data and texture data to the vertex and fragment shaders.

TextImageSketcher uses GDIPlus to create ARGB in memory bitmap containing text to be drawn. The size of the bitmap should be multiples of 128. i.e.,  128, 256, 512 etc.

First bitmap is filled with black color and then Text is drawn using the font and color. Note the text color needs to be non black. Font can be changed by passing LOGFONT structure, Also text color can be changed by passing COLLOREF of the color.

Fragment shader is modified for blending fragments with the background.

The client needs to first Startup() GDIPlus in the beginning of the application and Shutdown()  at the end.

Instantiate TextImageSketcher object by Init() method supply unique texture id and size.

DrawText() can be called to draw text and supply font details and text color. 

DrawImage() can be called to render images.

The TextImageSketcher object can be translated, rotated etc.  just like any 3D object.

System  class diagram

TextImageSketcher derives from BaseGeometry class. It overrides mesh with an instance of TextMesh to generate 2D geometry. It also uses TextureUtil to render text and images into the texture.

TextMesh

TextMesh derives from IGeometryMesh class. It provides 2D geometry for drawing text and images. It supplies vertices and texture coordinates.

Members

Name	Description
texturemap	Provides texture map.
vertices	Provides 2D vertices.

Methods

Name	Description
Init	This method generates vertex data for the GPU to render the geometrical shape. This method is called if the data is sent in  non indexed mode. It generates position and vertex data.

TextureUtil

TextureUtil implements handling textures from image files and bitmaps. It's used by multiple purposes such as rendering 3D objects or text.

Members

Name	Description
textureID	Unique texture handle returned after creating the texture.
texunit	Unique texture unit. It's value should be one of the 80 values supported by the system.

Methods

Name	Description
Cleanup	Releases resources.
Init	Assigns unique texture unit to texunit.
LoadTextTexture	Used to render text. It generates texture from GDI bitmap.
LoadTextTextureImage	Used to render text. It wraps 2D image from GDI bitmap.
LoadTexture	Loads texture from an image file.
MakeActive	Makes the texture active.

TextImageSketcher

TextImageSketcher derives from BaseGeometry class. It renders text and images. The utility class TextureUtil is used for loading text and images onto the texture.

Members

Name	Description
wd ht	Specifies the dimensions of the bitmap.
texutl	An instance of TextureUtil. It helps with loading text and images onto the texture.

Methods

Name	Description
Init	This method overrides the base class Init method to load texture.  Generates the vertices data consisting of surface normals in non indexed mode and later sets them up in VBO buffer.  It also prepares the bitmap for rendering text and images on it. It also initializes the texutl object with texture id for loading textures.
UpdateUniforms	This method overrides the base class UpdateUniforms method. First it updates  "cameraposition" uniform. It also updates "tex" uniform to pass texture object.
vertexShaderSource	This method overrides the base class vertexShaderSource method. It returns the vertex shader code to render the texture object.
fragmentShaderSource	This method overrides the base class fragmentShaderSource method. It returns the fragment shader code to render the  texture object.
Cleanup	This method overrides the base class Cleanup method. It releases the resources used by the host object.
Init() Shutdown()	These methods are called by the client during start up and shutdown to initialize GDI Plus environment.
ClearCanvas	This method clears the bitmap for fresh renderings of text and images.
Drawtext	This method renders input text string based on the input such as font, color and string formats.
Drawimage	This method reenders the image in the file name and resizesto fir based on the clipping width and height inputs
DrawCanvas	This method loads the bitmap onto the texture.

Output

Implementation: LightedTexCube

Overview 

In previous examples, we saw that Position, Color, Texture data were sent to draw cubes. In this post we  will discuss how to implement  Phong lighting models and other lights such as Directional, Point and Spot Lights. For this surface normals needs to be sent as VBO data. They look as shown at the bottom. They look elongated because aspect ratio is not applied.

Details

System  class diagram

LightedTexCube derives from TexturedCube class. It overrides mesh with an instance of CubeMesh to generate geometry. 

LightedTexCube

LightedTexCube is derived from TexturedCube class. It illuminates the textured cube as per settings done in the light parameter. Utility class gLightingUtil is used for lighting related settings.

Members

Name	Description
light	An instance of LightingUtil. It contains settings for illumination.

Methods

Name	Description
Init	This method overrides the base class Init method to load texture.  Generates the vertices data consisting of surface normals in non indexed mode and later sets them up in VBO buffer.  It also passes an input to use BlinnPhong or Phong illumination and also type of light source as defined by LightSource enum.
UpdateUniforms	This method overrides the base class UpdateUniforms method. Firs t it updates  "cameraposition" uniform. Next it updates lighting related uniforms by calling Updateshader method on light. Finally it c alls base class UpdateUniforms method to send rest.
vertexShaderSource	This method overrides the base class vertexShaderSource method. It returns the vertex shader code to render the cube object.
fragmentShaderSource	This method overrides the base class fragmentShaderSource method. It returns the fragment shader code to render the cube object based on light source type.

MaterialInfo

This class contains information about a single Material obtained after parsing a mtl file. This information includes  Ambient, Diffuse, Specular Colors and Texture information etc.  Note that a mtl file can have multiple Materials applicable to a range of meshes.

These are stored in LightingUtil object discussed below.

Members

Name	Description
name	Name of the material.
ambientColor	Contains ambient color
diffuseColor	Contains Diffuse color
specularColor	Contains Specular color
Shininess	Contains shininess value

LightSrcInfo

This class contains information about the light source and viewer position for lighting calculations.

These are stored in GenericObj object discussed below. 

Members

Name	Description
src	Contains light source type information. It has to be one of LightSourceType
name	Name of the light source.
ambientCoefficient ambientColor	Contains Ambient Coefficient and color
diffuseCoefficient diffuseColor	Contains Diffuse Coefficient and color
specularColor specularCoefficient	Contains Specular Coefficient and color
blinn	Contains boolean value to indicate to render blinn-phong(true) or phong(false)  lighting.
position	Contains light position  information required for basic, point and spot lighting.
direction	Contains direction information required for directional lighting.
attconstant attlinear attquadratic	Contains light attenuation information required for point and spot lighting.
spotlightinner spotlightouter	Contains light angle information required for spot lighting.

LightingUtil

It defines illumination properties of the light source and reflection properties of the material in the light and material members.

Members

Name	Description
material	An instance of Material class. It contains settings of the material property for reflection.
light	An instance of LightSrcInfo class. It contains settings for illumination.

Methods

Name	Description
UpdateShader	Updates the uniforms that contains copies of the information stored in light and material members.

Output

The output looks as shown below.

Implementation: Textured Cube

Overview 

In the previous discussions, we covered sending vertex data - Position and color. In this post we will try to send texture vertex data. It's implemented in TexturedCube class.

It looks as shown at the bottom. It looks elongated because aspect ratio is not applied.

Details

A Texture is 2D image that can be wrapped around a 3D object like a gift wrapper. For example, resources\textures\bricks2.jpg is a 2D Texture file. 

Unlike cartesian coordinates, Texture follow UV  System as shown below.

TextureUtil Class is used for loading textures. 

Loading Texture

OpenGL supports up to 80 Textures and each with its own identifier. Loading textures is implemented in LoadTexture method in TextureUtil class. It calls the following SOIL library method. Refer to SOIL  documentation for further info.

Syntax

unsigned int
SOIL_load_OGL_texture
(
    const char *filename,
    int force_channels,
    unsigned int reuse_texture_ID,
    unsigned int flags
)

Name	Description
filename	The name of the file to upload as a texture
force_channels	0-image format, 1-luminous, 2-luminous/alpha, 3-RGB, 4-RGBA
reuse_texture_ID	0-generate a new texture ID, otherwise reuse the texture ID (overwriting the old texture)
flags	flags can be any of SOIL_FLAG_POWER_OF_TWO | SOIL_FLAG_MIPMAPS | SOIL_FLAG_TEXTURE_REPEATS | SOIL_FLAG_MULTIPLY_ALPHA | SOIL_FLAG_INVERT_Y | SOIL_FLAG_COMPRESS_TO_DXT | SOIL_FLAG_DDS_LOAD_DIRECT | SOIL_FLAG_NTSC_SAFE_RGB  |  SOIL_FLAG_CoCg_Y  | SOIL_FLAG_TEXTURE_RECTANGLE | SOIL_FLAG_PVR_LOAD_DIRECT | SOIL_FLAG_ETC1_LOAD_DIRECT | SOIL_FLAG_GL_MIPMAPS   | SOIL_FLAG_SRGB_COLOR_SPACE
return value	0-failed, otherwise returns the OpenGL texture handle

MIPMAP

Mipmapping is a technique where a high-resolution texture is downscaled and filtered so that each subsequent mip level is a quarter of the area of the previous level. This means that the texture and all of its generated mips requires no more than 1.5 times the original texture size. An example is shown below.

Inversion of Y Axis

After loading the image needs to be inverted since the V axis of the texture and Y axis of OpenGL  have opposite polarity.

Wrapping

As shown below, the wrapping mode determines how the texture is wrapped.

Texture Filtering

Texture filtering is a method that is used to improve the texture quality in a scene. Without texture filtering, artifacts like aliasing generally look worse. Texture filtering makes textures look better and less blocky.

Applying Texture

As discussed previously, Texture coordinates are passed during rendering. The following mapping in relation to the UxV axes is used as texture coordinates for the vertices consecutively.

//First Triangle
{ 0.0f, 1.0f },
{ 0.0f, 0.0f },
{ 1.0f, 0.0f },
//Second Triangle
{ 1.0f, 0.0f },
{ 1.0f, 1.0f },
{ 0.0f, 1.0f }

System  class diagram

Every 3D object such as IndexedCube  derives from BaseGeometry class. It overrides mesh with an instance of CubeMesh to generate geometry. It uses a helper class texutl to help render texture.

TextureUtil

TextureUtil implements handling textures from image files and bitmaps. It's used by multiple purposes such as rendering 3D objects or text.

Members

Name	Description
textureID	Unique texture handle returned after creating the texture.
texunit	Unique texture unit. It's value should be one of the 80 values supported by the system.

Methods

Name	Description
Cleanup	Releases resources.
Init	Assigns unique texture unit to texunit.
LoadTextTexture	Used to render text. It generates texture from GDI bitmap.
LoadTextTextureImage	Used to render text. It wraps 2D image from GDI bitmap.
LoadTexture	Loads texture from an image file.
MakeActive	Makes the texture active.

TexturedCube

TexturedCube is derived from BaseGeometry class. It assembles vertex data generated by mesh in VBO and EBO buffers. Later  sent to GPU using shader programs.

Members

Name	Description
filename	File name of the image file containing the texture.
texutl	Texture utility to aid rendering texture.

Methods

Name	Description
Init	This method overrides the base class Init method. It creates the mesh object. Generates the vertices and texture data in non indexed mode and later sets them up in VBO buffers of the GPU. It later loads the texture from the image file.
UpdateUniforms	This method overrides the base class UpdateUniforms method. The base class method updates the "transform" matrix uniform by calling GetTransformationMatrix method. It updates the uniform "tex" containing the texture coordinates from to the loaded texture.
vertexShaderSource	This method overrides the base class vertexShaderSource method. It returns the vertex shader code to render the cube object.
fragmentShaderSource	This method overrides the base class fragmentShaderSource method. It returns the fragment shader code to render the cube object.
Cleanup	Releases resources. Calls base class method and the texutl.

Output