electrosoup.com

git clone git://git.electrosoup.com/electrosoup.com
Log | Files | Refs

index.html (7345B)


      1 <!DOCTYPE html>
      2 <html lang="en-US">
      3 
      4 <head>
      5   <meta http-equiv="X-Clacks-Overhead" content="GNU Terry Pratchett" />
      6   <meta charset="utf-8">
      7   <meta name="viewport" content="width=device-width, initial-scale=1.0" />
      8 <link rel="shortcut icon" href="https://electrosoup.com/favicon.png" />
      9 <title>Deriving the View Matrix | Electrosoup</title>
     10 <meta name="title" content="Deriving the View Matrix" />
     11 <meta name="description" content="In computer graphics, the view matrix is used to transform points from
     12 world space into camera space. Below is an explanation of how to derive this
     13 matrix." />
     14 <meta name="author" content="" />
     15 <meta name="keywords" content="Graphics,Matrices,OpenGL," />
     16 
     17 
     18 
     19 
     20 
     21 
     22   
     23   <meta property="og:url" content="https://electrosoup.com/posts/view-matrix/">
     24   <meta property="og:site_name" content="Electrosoup">
     25   <meta property="og:title" content="Deriving the View Matrix">
     26   <meta property="og:description" content="In computer graphics, the view matrix is used to transform points from world space into camera space. Below is an explanation of how to derive this matrix.">
     27   <meta property="og:locale" content="en_US">
     28   <meta property="og:type" content="article">
     29     <meta property="article:section" content="posts">
     30     <meta property="article:published_time" content="2021-11-21T13:34:08-08:00">
     31     <meta property="article:modified_time" content="2021-11-21T13:34:08-08:00">
     32     <meta property="article:tag" content="Graphics">
     33     <meta property="article:tag" content="Matrices">
     34     <meta property="article:tag" content="OpenGL">
     35 
     36 
     37   
     38   
     39   <meta name="twitter:card" content="summary">
     40   <meta name="twitter:title" content="Deriving the View Matrix">
     41   <meta name="twitter:description" content="In computer graphics, the view matrix is used to transform points from world space into camera space. Below is an explanation of how to derive this matrix.">
     42 
     43 
     44   
     45   
     46   <meta itemprop="name" content="Deriving the View Matrix">
     47   <meta itemprop="description" content="In computer graphics, the view matrix is used to transform points from world space into camera space. Below is an explanation of how to derive this matrix.">
     48   <meta itemprop="datePublished" content="2021-11-21T13:34:08-08:00">
     49   <meta itemprop="dateModified" content="2021-11-21T13:34:08-08:00">
     50   <meta itemprop="wordCount" content="521">
     51   <meta itemprop="keywords" content="Graphics,Matrices,OpenGL">
     52 
     53 
     54   <link href="/herman.min.css" rel="stylesheet">
     55 
     56   
     57 
     58   
     59 
     60   
     61 <link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/katex@0.16.11/dist/katex.min.css" integrity="sha384-nB0miv6/jRmo5UMMR1wu3Gz6NLsoTkbqJghGIsx//Rlm+ZU03BU6SQNC66uf4l5+" crossorigin="anonymous">
     62 <script defer src="https://cdn.jsdelivr.net/npm/katex@0.16.11/dist/katex.min.js" integrity="sha384-7zkQWkzuo3B5mTepMUcHkMB5jZaolc2xDwL6VFqjFALcbeS9Ggm/Yr2r3Dy4lfFg" crossorigin="anonymous"></script>
     63 <script defer src="https://cdn.jsdelivr.net/npm/katex@0.16.11/dist/contrib/auto-render.min.js" integrity="sha384-43gviWU0YVjaDtb/GhzOouOXtZMP/7XUzwPTstBeZFe/+rCMvRwr4yROQP43s0Xk" crossorigin="anonymous"
     64     onload="renderMathInElement(document.body);"></script>
     65 
     66 
     67 </head>
     68 
     69 <body>
     70   <header><a class="skip-link" href="#main-content">Skip to main content</a>
     71 
     72 <nav>
     73   <a href="/">Main</a>
     74 
     75   <a href="/posts/">Posts</a>
     76 
     77   <a href="https://git.electrosoup.com">Code</a>
     78 
     79   <a href="https://christianermann.dev">Resume</a>
     80 
     81 
     82 
     83 
     84 
     85 
     86 
     87 
     88 
     89 </nav>
     90 </header>
     91   <main id="main-content">
     92 
     93 <h1>Deriving the View Matrix</h1>
     94 <p class="byline">
     95   
     96   <time datetime='2021-11-21' pubdate>
     97     2021-11-21
     98   </time>
     99   
    100 </p>
    101 
    102 <content>
    103   <p>In computer graphics, the view matrix is used to transform points from
    104 world space into camera space. Below is an explanation of how to derive this
    105 matrix.</p>
    106 <p>Let $V$ be the view matrix.
    107 If $\vec{r}$, $\vec{u}$, and $\vec{f}$ denote the right, up, and forward
    108 vectors of the camera and $\vec{p}$ is the camera&rsquo;s position, $V$ should
    109 satisfy the following properties:
    110 $$
    111 \begin{matrix}
    112 V_{1*} \times
    113 \begin{bmatrix}
    114 r_{x} \\
    115 r_{y} \\
    116 r_{z} \\
    117 0
    118 \end{bmatrix}
    119 =
    120 \begin{bmatrix}
    121 1 \\
    122 0 \\
    123 0 \\
    124 0
    125 \end{bmatrix},&amp;
    126 V_{2*} \times
    127 \begin{bmatrix}
    128 u_{x} \\
    129 u_{y} \\
    130 u_{z} \\
    131 0
    132 \end{bmatrix}
    133 =
    134 \begin{bmatrix}
    135 0 \\
    136 1 \\
    137 0 \\
    138 0
    139 \end{bmatrix},&amp;
    140 V_{3*} \times
    141 \begin{bmatrix}
    142 f_{x} \\
    143 f_{y} \\
    144 f_{z} \\
    145 0
    146 \end{bmatrix}
    147 =
    148 \begin{bmatrix}
    149 0 \\
    150 0 \\
    151 1 \\
    152 0
    153 \end{bmatrix},&amp;
    154 V_{4*} \times
    155 \begin{bmatrix}
    156 p_{x} \\
    157 p_{y} \\
    158 p_{z} \\
    159 1
    160 \end{bmatrix}
    161 =
    162 \begin{bmatrix}
    163 0 \\
    164 0 \\
    165 0 \\
    166 1
    167 \end{bmatrix}
    168 \end{matrix}
    169 $$
    170 In words, the view matrix should translate the camera to the origin and rotate
    171 it to align with the $\vec{x}$, $\vec{y}$, and $\vec{z}$ axes.</p>
    172 <p>This can be expressed more concisely as the product of two matrices:
    173 $$
    174 V \times
    175 \begin{bmatrix}
    176 r_{x} &amp; u_{x} &amp; f_{x} &amp; p_{x} \\
    177 r_{y} &amp; u_{y} &amp; f_{y} &amp; p_{y} \\
    178 r_{z} &amp; u_{z} &amp; f_{z} &amp; p_{z} \\
    179 0     &amp; 0     &amp; 0     &amp; 1
    180 \end{bmatrix}
    181 =
    182 \begin{bmatrix}
    183 1 &amp; 0 &amp; 0 &amp; 0 \\
    184 0 &amp; 1 &amp; 0 &amp; 0 \\
    185 0 &amp; 0 &amp; 1 &amp; 0 \\
    186 0 &amp; 0 &amp; 0 &amp; 1
    187 \end{bmatrix}
    188 $$</p>
    189 <p>To make this equation easier to solve, we can split the matrix containing our
    190 camera parameters into the product of a translation matrix $T$ and a rotation
    191 matrix $R$:
    192 $$
    193 V \times (T \times R) = I
    194 $$
    195 where
    196 $$
    197 \begin{matrix}
    198 T =
    199 \begin{bmatrix}
    200 1 &amp; 0 &amp; 0 &amp; p_{x} \\
    201 0 &amp; 1 &amp; 0 &amp; p_{y} \\
    202 0 &amp; 0 &amp; 1 &amp; p_{z} \\
    203 0 &amp; 0 &amp; 0 &amp; 1
    204 \end{bmatrix},&amp;
    205 R =
    206 \begin{bmatrix}
    207 r_{x} &amp; u_{x} &amp; f_{x} &amp; 0 \\
    208 r_{y} &amp; u_{y} &amp; f_{y} &amp; 0 \\
    209 r_{z} &amp; u_{z} &amp; f_{z} &amp; 0 \\
    210 0     &amp; 0     &amp; 0     &amp; 1
    211 \end{bmatrix}
    212 \end{matrix}
    213 $$</p>
    214 <p>Solving for $V$ reveals that $V=R^{-1} \times T^{-1}$:
    215 $$
    216 V =
    217 \begin{bmatrix}
    218 r_{x} &amp; r_{y} &amp; r_{z} &amp; 0 \\
    219 u_{x} &amp; u_{y} &amp; u_{z} &amp; 0 \\
    220 f_{x} &amp; f_{y} &amp; f_{z} &amp; 0 \\
    221 0     &amp; 0     &amp; 0     &amp; 1
    222 \end{bmatrix}
    223 \times
    224 \begin{bmatrix}
    225 1 &amp; 0 &amp; 0 &amp; -p_{x} \\
    226 0 &amp; 1 &amp; 0 &amp; -p_{y} \\
    227 0 &amp; 0 &amp; 1 &amp; -p_{z} \\
    228 0 &amp; 0 &amp; 0 &amp; 1
    229 \end{bmatrix}
    230 $$</p>
    231 <p>And with some minor simplification,
    232 $$
    233 V
    234 =
    235 \begin{bmatrix}
    236 r_{x} &amp; r_{y} &amp; r_{z} &amp; -(\vec{r} \cdot \vec{p}) \\
    237 u_{x} &amp; u_{y} &amp; u_{z} &amp; -(\vec{u} \cdot \vec{p}) \\
    238 f_{x} &amp; f_{y} &amp; f_{z} &amp; -(\vec{f} \cdot \vec{p}) \\
    239 0     &amp; 0     &amp; 0     &amp; 1
    240 \end{bmatrix}
    241 $$</p>
    242 <p>When we&rsquo;re ready to render our scene, we can apply $V$ to each vertex $\vec{q}$
    243 by calculating
    244 $$
    245 V \times \vec{q}
    246 $$
    247 in our vertex shader.</p>
    248 
    249 </content>
    250 <p>
    251   
    252     <a class="blog-tags" href="/tags/graphics/">#graphics</a>
    253   
    254     <a class="blog-tags" href="/tags/matrices/">#matrices</a>
    255   
    256     <a class="blog-tags" href="/tags/opengl/">#opengl</a>
    257   
    258 </p>
    259 
    260 
    261   <p>
    262     <a href='mailto:christianermann@gmail.com?subject=Reply%20to%20"Deriving%20the%20View%20Matrix"'>
    263       Reply to this post by email ↪
    264     </a>
    265   </p>
    266 
    267 
    268 
    269   </main>
    270   <footer><footer>
    271     
    272         <a href="https://git.electrosoup.com">Code</a>.
    273     
    274         <a href="https://christianermann.dev">Resume</a>.
    275     
    276     <a href='https://electrosoup.com/index.xml'>RSS</a>.
    277     <p>Copyright &copy; 2026 Christian Ermann. | Made with <a href="https://github.com/clente/hugo-bearcub">Bear Cub</a></p>
    278 </footer>
    279 </footer>
    280 
    281     
    282 </body>
    283 
    284 </html>