Difference between revisions of "3D Fundamentals Tutorial 2"
From Chilipedia
(→Video) |
(→Video) |
||
| (3 intermediate revisions by the same user not shown) | |||
| Line 6: | Line 6: | ||
<div class="mw-collapsible mw-collapsed"><br /> | <div class="mw-collapsible mw-collapsed"><br /> | ||
| − | * Fix an error in <code>Mat2.h</code> ( | + | * Fix an error in <code>Mat2.h</code> (multiplication of vectors with matrices) [https://youtu.be/HV1ZGyiY5C8?t=0m16s 0:16] |
* Introducing 3D space [https://youtu.be/HV1ZGyiY5C8?t=1m57s 1:57] | * Introducing 3D space [https://youtu.be/HV1ZGyiY5C8?t=1m57s 1:57] | ||
* Implementing the <code>Vec3</code> class [https://youtu.be/HV1ZGyiY5C8?t=3m19s 3:19] | * Implementing the <code>Vec3</code> class [https://youtu.be/HV1ZGyiY5C8?t=3m19s 3:19] | ||
| Line 21: | Line 21: | ||
:* And start with defining World Space (=object space for now) and Screen Space | :* And start with defining World Space (=object space for now) and Screen Space | ||
</div> | </div> | ||
| − | * Setting up a World Space ( | + | * Setting up a World Space (Chili likes to call it Pube Space) [https://youtu.be/HV1ZGyiY5C8?t=10m08s 10:08] |
<div class="mw-collapsible-content"> | <div class="mw-collapsible-content"> | ||
:* We set the Origin (0,0,0) of world space in the middle of the screen | :* We set the Origin (0,0,0) of world space in the middle of the screen | ||
| Line 40: | Line 40: | ||
:* <code>Cube</code> class created as a data structure to hold the vertices | :* <code>Cube</code> class created as a data structure to hold the vertices | ||
</div> | </div> | ||
| − | * Representing lines: vertex container + | + | * Representing lines: vertex container + vector of indices [https://youtu.be/HV1ZGyiY5C8?t=18m34s 18:34] |
<div class="mw-collapsible-content"> | <div class="mw-collapsible-content"> | ||
:* This strategy avoids having to transform vertices multiple times | :* This strategy avoids having to transform vertices multiple times | ||
Latest revision as of 06:12, 24 April 2020
In this video we lay the groundwork needed to represent and transform positions and 3D space, we define the relationship between our 3D space and the screen coordinates, and we setup a data structure to hold our test cube geometry.
Video
The tutorial video is on YouTube here.
- Fix an error in
Mat2.h(multiplication of vectors with matrices) 0:16 - Introducing 3D space 1:57
- Implementing the
Vec3class 3:19
- Inherits from
Vec2(z-component added) - Overload all operators for 3D
- Advantage of inheritence here: enables passing
Vec3s into functions that acceptVecc2&references
- Inherits from
- Calculating the length of a 3D vector (i.e., in R3) 4:57
- The
Mat33D Matrix class 5:57 - Choosing a coordinate system 7:37
- We choose a Left Handed coordinate system
- And start with defining World Space (=object space for now) and Screen Space
- Setting up a World Space (Chili likes to call it Pube Space) 10:08
- We set the Origin (0,0,0) of world space in the middle of the screen
- - This makes for better (easier) scaling
- We will use NDC (Normalized Device Coordinates)
- - Meaning: we don't use pixel values, but normalized values relative to screen size
- - (e.g. x=1.0 means far right end of screen, x=-1.0 for far left end of screen
- - Main advantage: become independent from screen size
- - Note: Direct 3D and OpenGL use NDC
- We will use the Pre-Clip 3D space definition: z=1.0 indicates half cube depth
- Transforming vertices from Pube Space to Screen Space 14:01
- Introducing the
PubeScreenTransformerclass
- Introducing the
- Testing: drawing a cube 17:07
-
Cubeclass created as a data structure to hold the vertices
-
- Representing lines: vertex container + vector of indices 18:34
- This strategy avoids having to transform vertices multiple times
- Same concept as IndexBuffer used in Direct3D to draw primitives 20:40
- Make the
Cubeclass return aIndexedLineListstruct
- Code for drawing: transform the vertices and iterate through the index buffer 21:50
- Limitations observed: not square and not 3D. Why? 23:22
- Screen size is not square, we'll use a quick fix for now
- We still lack 3D visual cues because we don't account for the z-component yet (we're still in Orthographic Projection)
Downloads
The GitHub repository for the tutorial code is here.
