656 reputation
417
bio website
location France
age 25
visits member for 2 years, 6 months
seen 10 hours ago

Apr
8
comment Optimize state changes within a render queue
I think that my problem will only occurr when filling the color buffer. Because it's in this buffer that I will use the material of the object and so the technique associated to it. I don't have lot of knowledge about deferred rendering, only read some article without going in depth. So I am probably wrong on a lot of things. Since I don't have a complex rendering pipeline, I should make things easy and simple and add more functionalities and adapt it when I will implement deferred rendering. Without deferred rendering implemented in the engine I can't predict which problems I will encounter.
Apr
8
comment Optimize state changes within a render queue
I probably misuse the term "pass". In my question I refer to pass for a pass in a technique in a directx effect file (.fx). That is why in my question I am saying that a material in my engine can have multiple pass. A material is bind to one technique and therefore to technique's passes. Until now, the engine doesn't have complex rendering pipeline, I only draw object "as is" with their own materials. But if an object use a material with a technique containing two passes, I need to do one draw call for each effect pass, so in my render queue I will have two times the object for each pass.
Apr
7
comment Optimize state changes within a render queue
Thanks for your response, it helped me think about my problem. Meanwhile I have found an interesting article about a method to sort a render queue. I think about implementing it but I have a lot more questions, so I edited my question with more details and the link to the article.
Apr
7
revised Optimize state changes within a render queue
Add details
Apr
1
asked Optimize state changes within a render queue
Nov
21
comment Incorrect frustum culling with center/half-size AABB
I'm sorry to haven't provided enough code. The code I added in my question is all what I wrote. Anything else is provided by XNA (Vector3, ...). To get the frustum plane I use a class called BoundingFrustum. Each frame I give the viewProj matrix to the BoundingFrustum and it computes the plane. I don't know how it computes plane and D.
Nov
21
asked Incorrect frustum culling with center/half-size AABB
Oct
7
awarded  Yearling
Aug
7
awarded  Popular Question
Jun
18
awarded  Constituent
Jun
11
awarded  Informed
Jun
10
awarded  Caucus
May
27
answered Sharing the effect
Feb
15
awarded  Popular Question
Oct
7
awarded  Yearling
Aug
9
answered Smallest increment for a SpriteBatch depth between 0.0 and 1.0?
Aug
8
revised unity3d camera movment problem
Change explanation
Aug
8
revised unity3d camera movment problem
deleted 3 characters in body
Aug
8
answered unity3d camera movment problem
Aug
5
comment Draw a 3D object at specific coordinates on the screen
Thanks for your response. But I don't understand your explanation about ray and position on the ray. I understand that a 3D point is converted in 2D screen point by doing World * View * Projection and doing the opposite will convert from 2D to 3D. I've tried to invert the WVP matrix and transform my 2D point, and set the 3D position to the matrix of my model. But it doesn't work. When I move my camera, my model moves but doesn't stay at the same position on the screen. I think i've missed something, probably what you try to explain about ray and "depth of monitor".