To fix it, I started with normal calculations. I had forgotten to include the view matrix in the normal transform. Also, the object was dissapearing on window resize, which I fixed by reinitializing the gBuffer as stated in the homework pdf (since it depends on screen dimensions).
Z-fighting was a particularly annoying bug which is caused by how I defined gPosition<\/code> and gNormal<\/code> in the fragment shader. When they were vec3<\/code>, depth values weren’t initialized properly, causing Z-fighting. I changed them to vec4<\/code> and the problem solved magically. I tried a lot of things up to this point because the problem was so unexpected for me.<\/p>\n\n\n\nAnother bug: I noticed that the gPosition<\/code> and gNormal<\/code> textures looked identical in the geometry visualization shaders. Both the gPosition<\/code> and gNormal<\/code> data were being written correctly to the G-buffer, so the issue wasn\u2019t with the rendering, just with how the textures were being attached to the shader.first<\/em> (normals or positions) would end up being used for both, which clued me in that the problem was in the fragment shader bindings.gNormal<\/code> sampler in the shader. Adding this line fixed it:glUniform1i(glGetUniformLocation(visualizeProgram, “gNormal”), 1);<\/strong>gNormal<\/code> texture was correctly bound to texture unit 1. After that, both the position and normal visualizations rendered properly.<\/p>\n\n\n\nAfter some little additions this is a showcase video and some screenshots. Sorry for the flashbang :).<\/p>\n\n\n\n<\/video><\/figure>\n\n\n\n<\/p>\n\n\n\n
Shading with G-Buffers<\/strong><\/h4>\n\n\n\nBlinn-Phong shading<\/strong> with four point lights, one near the head, one at the back, and two around the legs. The initial result was decent, but I noticed lighting didn\u2019t behave consistently with earlier stages. It turned out that I had stored vertex positions in object space instead of world space during the geometry pass. Fixing that brought lighting back to expected behavior.<\/p>\n\n\n\n\n\nAt the end of the homework I used the same lighting setup with the given shaders to make the scene consistent.<\/p>\n\n\n\n
<\/p>\n\n\n\n
Mouse Movements + HDR Image Fix<\/strong><\/h4>\n\n\n\nNext up: mouse movement. I implemented it using glm::lookAt<\/code>, and it worked nicely. But then I saw a mysterious black patch on the sun in the tonemapped image. My cubemap texture was using GL_RGB16F<\/code>, not enough precision. Switching to GL_RGB32F<\/code> fixed it. You can clearly see the black patch with a low key value tonemap.<\/p>\n\n\n\nAfter the fix:<\/p>\n\n\n\n<\/p>\n\n\n\n
Motion Blur<\/strong><\/h4>\n\n\n\nTo finish the homework, I implemented a motion blur effect<\/strong> as a post-processing step. I did this by sampling the color of several neighboring pixels in the fragment shader and blending them with different weights (simple linear motion blur). The closer a neighboring pixel was to the center (detected with a simple linear calculation), the more influence it had on the final color. This effect is particularly noticeable when the camera or object is rotating quickly. Configuring and deciding the parameters of the motion blur and movement pace detection was harder than the motion blur algorithm.<\/p>\n\n\n\nTo apply this effect to the scene I rendered the scene to a frame buffer and use that framebuffer to render the scene to the screen by motion blur shaders.<\/p>\n\n\n\n
<\/p>\n\n\n\n
Known Issues<\/strong>\nCurrent motion blur is simple. Log average luminance may be used to make a better effect as stated in the homework pdf.Final Video<\/video><\/figure>\n","protected":false},"excerpt":{"rendered":"This assignment was about putting together a basic deferred renderer with HDR cubemap support, tone mapping, motion blur and some basic camera movements. I rolled up my sleeves, dove into OpenGL, and here’s how it all unfolded Code Cleanup I started by refactoring the given code to understand what\u2019s what and build a more organized […]<\/p>\n","protected":false},"author":8372,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_links_to":"","_links_to_target":""},"categories":[1],"tags":[],"class_list":["post-28","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/posts\/28","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/users\/8372"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/comments?post=28"}],"version-history":[{"count":0,"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/posts\/28\/revisions"}],"wp:attachment":[{"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/media?parent=28"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/categories?post=28"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.metu.edu.tr\/e251884\/wp-json\/wp\/v2\/tags?post=28"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
![\"\"](\"https:\/\/blog.metu.edu.tr\/e251884\/files\/2025\/05\/first-screenshot.png\")
<\/figure>\n\n\n\n![\"\"](\"https:\/\/blog.metu.edu.tr\/e251884\/files\/2025\/05\/after-step-7.png\")
<\/figure>\n\n\n\n