Trouble loading and drawing a .obj model using Assimp and OpenGL - c++

Question

After performing the basics of OpenGL (creating a window, making a 2D triangle, shaders, etc.) I decided to start trying to load simple .obj models. The most recommended library was Assimp so I followed some tutorials and modified my project to load models. But unfortunately, the models were displayed very strangely. I created the following code to show this:
#include <iostream>
#include <string>
#include <vector>
#include <memory>
#include <GL/glew.h>
#include <GLFW/glfw3.h>
#include <glm/glm.hpp>
#include <assimp/Importer.hpp>
#include <assimp/scene.h>
#include <assimp/postprocess.h>
struct Vertex
{
glm::vec3 position;
glm::vec3 normal;
};
struct Mesh
{
//The vertex array object, vertex buffer object and element buffer object
GLuint VAO;
GLuint VBO;
GLuint EBO;
//Vectors for the vertices and indices to put in the buffers
std::vector<Vertex> vertices;
std::vector<GLuint> indices;
//Constructor
Mesh(const std::vector<Vertex>& vertices, const std::vector<GLuint>& indices)
{
this->vertices = vertices;
this->indices = indices;
//Generate the VAO
glGenVertexArrays(1, &VAO);
//Generate the buffer objects
glGenBuffers(1, &VBO);
glGenBuffers(1, &EBO);
//Bind the VAO
glBindVertexArray(VAO);
//Bind the VBO and set the vertices
glBindBuffer(GL_ARRAY_BUFFER, VBO);
glBufferData(GL_ARRAY_BUFFER, sizeof(Vertex) * vertices.size(), &vertices.at(0), GL_STATIC_DRAW);
//Enable the first attribute pointer
glEnableVertexAttribArray(0);
//Set the attribute pointer The stride is meant to be 'sizeof(Vertex)', but it doesn't work at all that way
// \/
glVertexAttribPointer(0, 3, GL_FLOAT, GL_FALSE, 0, 0);
//Enable the second attribute pointer
glEnableVertexAttribArray(1);
//Set the attribute pointer ditto
// \/
glVertexAttribPointer(1, 3, GL_FLOAT, GL_FALSE, 0, (void*) offsetof(Vertex, normal));
//Bind the EBO and set the indices
glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, EBO);
glBufferData(GL_ELEMENT_ARRAY_BUFFER, sizeof(GLuint) * indices.size(), &indices.at(0), GL_STATIC_DRAW);
//Report any errors
GLenum error = glGetError();
if (error != GL_NO_ERROR)
{
std::cerr << "Error while creating mesh!" << std::endl;
}
glBindVertexArray(0);
}
void draw()
{
//Bind the VAO
glBindVertexArray(VAO);
//Bind the ELement Buffer Object
glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, EBO);
//Draw the mesh
glDrawElements(GL_TRIANGLES, indices.size(), GL_UNSIGNED_INT, 0);
//Unbind the VAO
glBindVertexArray(0);
}
};
int main()
{
//Intialize GLFW (no error checking for brevity)
glfwInit();
//Set the OpenGL version to 3.3
glfwWindowHint(GLFW_CONTEXT_VERSION_MAJOR, 3);
glfwWindowHint(GLFW_CONTEXT_VERSION_MINOR, 3);
//Create a new window
GLFWwindow* window = glfwCreateWindow(800, 600, "Model Testing", NULL, NULL);
glfwMakeContextCurrent(window);
//Initialize glew (no checking again)
glewInit();
glViewport(0, 0, 800, 600);
glClearColor(1.0f, 1.0f, 1.0f, 1.0f);
//Load the model
Assimp::Importer importer;
const aiScene* scene = importer.ReadFile("mymodel.obj", aiProcess_Triangulate | aiProcess_GenNormals);
//Check for errors
if ((!scene) || (scene->mFlags == AI_SCENE_FLAGS_INCOMPLETE) || (!scene->mRootNode))
{
std::cerr << "Error loading mymodel.obj: " << std::string(importer.GetErrorString()) << std::endl;
//Return fail
return -1;
}
//A vector to store the meshes
std::vector<std::unique_ptr<Mesh> > meshes;
//Iterate over the meshes
for (unsigned int i = 0; i < scene->mNumMeshes; ++i)
{
//Get the mesh
aiMesh* mesh = scene->mMeshes[i];
//Create vectors for the vertices and indices
std::vector<Vertex> vertices;
std::vector<GLuint> indices;
//Iterate over the vertices of the mesh
for (unsigned int j = 0; j < mesh->mNumVertices; ++j)
{
//Create a vertex to store the mesh's vertices temporarily
Vertex tempVertex;
//Set the positions
tempVertex.position.x = mesh->mVertices[j].x;
tempVertex.position.y = mesh->mVertices[j].y;
tempVertex.position.z = mesh->mVertices[j].z;
//Set the normals
tempVertex.normal.x = mesh->mNormals[j].x;
tempVertex.normal.y = mesh->mNormals[j].y;
tempVertex.normal.z = mesh->mNormals[j].z;
//Add the vertex to the vertices vector
vertices.push_back(tempVertex);
}
//Iterate over the faces of the mesh
for (unsigned int j = 0; j < mesh->mNumFaces; ++j)
{
//Get the face
aiFace face = mesh->mFaces[j];
//Add the indices of the face to the vector
for (unsigned int k = 0; k < face.mNumIndices; ++k) {indices.push_back(face.mIndices[k]);}
}
//Create a new mesh and add it to the vector
meshes.push_back(std::unique_ptr<Mesh>(new Mesh(std::move(vertices), std::move(indices))));
}
//While the window shouldn't be closed
while (!glfwWindowShouldClose(window))
{
//Clear the buffer
glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
//Draw all the meshes
for (auto& mesh : meshes) {mesh.get()->draw();}
//Swap the buffers
glfwSwapBuffers(window);
}
//Close the window now that it's not needed anymore
glfwDestroyWindow(window);
return 0;
}
When the program loads this teapot my screen looks like this:
And from further away at another angle (using a more complex program than the one above):
In case it's useful, I'm running Ubuntu 16.04 with an Nvidia GTX 750 Ti, driver version 361.45

Answer 1

The stride should be sizeof(Vertex). If it doesn't work with that stride then something else is wrong!

Answer 2

Try moving the binding of the ebo to right after the vbo in the mesh constructor like this:
//Bind the VBO and set the vertices
glBindBuffer(GL_ARRAY_BUFFER, VBO);
glBufferData(GL_ARRAY_BUFFER, sizeof(Vertex) * vertices.size(), &vertices.at(0), GL_STATIC_DRAW);
//Bind the EBO and set the indices
glBindBuffer(GL_ELEMENT_ARRAY_BUFFER, EBO);
glBufferData(GL_ELEMENT_ARRAY_BUFFER, sizeof(GLuint) * indices.size(), &indices.at(0), GL_STATIC_DRAW);
Thats the way they have it on the mesh page you linked. I have had a similar problem with a different loader. Your indices are are not being loaded correctly so some verticies are positioned correctly while others are not.

Answer 3

My experience with OpenGL is pathetic so I can be mistaken. I see that your vertices are made as: x,y,z,nx,ny,nz where xyz is vertex coords and nxnynz is normal coords. Therefore stride is 6*sizeof(float).
glVertexAttribPointer(0, 3, GL_FLOAT, GL_FALSE, 6*sizeof(float), 0);
glVertexAttribPointer(1, 3, GL_FLOAT, GL_FALSE, 6*sizeof(float), (void*) 3*sizeof(float);
just look at second answer here: Opengl Vertex attribute stride to learn more about stride calculation
if this doesn't help check if indices where formed correctly
and small advice: work with cubes not with teapots (just made cube in blender or write it yourself in notepad)