I recently solved a problem which prevented my lighting from working in an OpenGL ES iOS app:
I solved the problem by replacing
glEnable(GL_DEPTH_TEST);
glGenVertexArraysOES(1, &_vertexArray);
glBindVertexArrayOES(_vertexArray);
glGenBuffers(1, &_vertexBuffer);
glBindBuffer(GL_ARRAY_BUFFER, _vertexBuffer);
glBufferData(GL_ARRAY_BUFFER, loader.currentCountOfVerticies * sizeof(GLfloat) * 3, arrayOfVerticies, GL_STATIC_DRAW);
glEnableVertexAttribArray(GLKVertexAttribPosition);
glVertexAttribPointer(GLKVertexAttribPosition, 3, GL_FLOAT, GL_FALSE, 12, BUFFER_OFFSET(0));
glGenVertexArraysOES(1, &_normalArray);
glBindVertexArrayOES(_normalArray);
glGenBuffers(1, &_normalBuffer);
glBindBuffer(GL_ARRAY_BUFFER, _normalBuffer);
glBufferData(GLKVertexAttribNormal, loader.currentCountOfNormals * sizeof(GLfloat) * 3,loader.arrayOfNormals , GL_STATIC_DRAW);
glEnableVertexAttribArray(GLKVertexAttribNormal);
glVertexAttribPointer(GLKVertexAttribNormal, 3, GL_FLOAT, GL_FALSE, 12, BUFFER_OFFSET(0));
glBindVertexArrayOES(0);
with
glEnable(GL_DEPTH_TEST);
glGenVertexArraysOES(1, &_vertexArray);
glBindVertexArrayOES(_vertexArray);
glGenBuffers(1, &_vertexBuffer);
glBindBuffer(GL_ARRAY_BUFFER, _vertexBuffer);
glBufferData(GL_ARRAY_BUFFER, total * sizeof(GLfloat), mergedArray, GL_STATIC_DRAW);
glEnableVertexAttribArray(GLKVertexAttribPosition);
glVertexAttribPointer(GLKVertexAttribPosition, 3, GL_FLOAT, GL_FALSE, 24, BUFFER_OFFSET(0));
glEnableVertexAttribArray(GLKVertexAttribNormal);
glVertexAttribPointer(GLKVertexAttribNormal, 3, GL_FLOAT, GL_FALSE, 24, BUFFER_OFFSET(12));
glBindVertexArrayOES(0);
Combining the array of Vertices and normals to a single GLfloat array and passing that array to a single buffer. This resolved my problem, but I don’t understand why. To my knowledge I should have been able to use 2 buffers?
The problem wasn’t the usage of multiple vertex buffers, but the usage of multiple vertex array objects. A vertex array object (VAO) is a lightweight object (meaning it doesn’t contain any actual vertex attribute data) encapsulating all the state required for rendering a bunch of vertex arrays with a single draw call, in particular
glVertexAttribPointerfor each attribute indexIt is therefore one level higher than the individual vertex attribute arrays, comprising all the vertex attribute array settings of a single (conceptual) scene object, or more correctly, a single draw call.
But in your original code you create a new vertex array object for each individual attribute. When then rendering you only bind
_vertexArraywhich in turn only sets and enables theGLKVertexAttribPositionattribute, thus no normals or whatever else.So you should have rather replaced the original code with: