Off-line Lighting and Per-vertex Reflectors — Handmade Hero — Episode Guide

0:07Introduce the new location, recap and set the stage for the day

1:58Run the game and consider the current state of the block lighting

5:59Blackboard: Multiresolution and Quality

16:45Blackboard: Aggregating reflectors

18:52Run the game and consider switching to rectangular reflectors

22:27Blackboard: Figuring out a multiresolution raycasting scheme

29:25Run the game and consider switching to an offline lighting solution, for experimentation purposes

30:50handmade_render_group.cpp: Make LightingTest() cast out 64 and then 256 rays and run the game to demonstrate how unusable this is

32:32handmade_render_group.cpp: Set up LightingTest() to compute the lighting offline

34:55handmade_render_group.h: Introduce lighting_solution for LightingTest() to use

36:24handmade_world_mode.cpp: Set up UpdateAndRenderWorld() to output the lighting on a key press

36:54handmade_render_group.cpp: Split out some of LightingTest() into a new OutputLighting() function

49:37handmade_world_mode.cpp: Make LightingTest() store the bitmaps when copying the lighting in

52:21win32_handmade.cpp: Enable Win32ProcessPendingMessages() to handle the F-keys

56:28handmade_world_mode.cpp: Enable UpdateAndRenderWorld() to toggle ShowLighting mode on a key press

59:07Run the game, try toggling to ShowLighting mode and hit an assert in PushQuad()

1:00:51handmade_render_group.cpp: Make OutputLighting() call GetCurrentQuads() up front

1:01:46Run the game and toggle to ShowLighting mode

1:02:25Infinite waters^α

1:03:07handmade_render_group.cpp: Make OutputLighting() loop through all the Elements, and LightingTest() write out the correct Bitmaps

1:04:21Run the game and toggle to ShowLighting mode

1:04:41win32_handmade.cpp: Make Win32ProcessPendingMessages() correctly clear the pressed key

1:05:56Run the game, toggle to ShowLighting mode and consider the solution

1:08:50handmade_render_group.cpp: Temporarily make ComputeLighting() set the lighting to full white, and run the game to confirm that we are correctly writing out the lighting

1:09:56Get some water

1:10:38Return with water

1:11:08handmade_render_group.cpp: Make LightingTest() cast 8 rays, run the game and consider if we are accumulating and not averaging

1:12:42handmade_render_group.cpp: Make LightingTest() correctly accumulate the rays

1:14:04Run the game to see that our lighting solution is a lot better

1:15:25handmade_render_group.cpp: Make OutputLighting() wind the vectors in the correct order

1:16:10Run the game and consider using the quad structures themselves as reflectors

1:21:06handmade_render_group.cpp: Make LightingTest() cast 256 rays, run the game and consider using the ray count as a divisor

1:23:06handmade_render_group.cpp: Make LightingTest() factor in the ray count

1:23:57Run the game to see a more stable lighting solution, in terms of its brightness

1:26:03Run the game and consider smoothing out the lighting solution across reflectors

1:35:22handmade_render_group.cpp: Split out LightingTest() into a new ExtractReflectorsFromQuads() function

1:38:46handmade_render_group.cpp: Introduce ExtractReflectorsFromVerts()

1:43:44Run the game and view just the reflectors

1:46:02handmade_render_group.cpp: Make ExtractReflectorsFromVerts() correctly create the four vertices

1:46:32Run the game to see our reflectors' vertices

1:48:12Run the game and temporarily believe that it is correct

1:49:19handmade_render_group.cpp: Consider making ExtractReflectorsFromVerts() position the vertices correctly

1:50:27Blackboard: Evenly distributing reflectors across a face

1:51:24handmade_render_group.cpp: Make ExtractReflectorsFromVerts() distribute the reflectors

1:53:46Run the game and take a look at the fairly regularly tiled reflectors

1:54:19handmade_render_group.cpp: Introduce OutputLightingQuads() and OutputLightingVerts()

1:59:20Run the game and check out the lighting solution, now with vertex-based banding

2:00:52Q&A

🗩

2:00:52Q&A

🗩

2:00:52Q&A

🗩

2:02:21Till_us You can triangle order the verts for the banding issue?

🗪

2:02:21Till_us You can triangle order the verts for the banding issue?

🗪

2:02:21Till_us You can triangle order the verts for the banding issue?

🗪

2:02:43Blackboard: Triangle vertex order picking

2:04:29Brian Forgive me, but is the game currently running in the hardware renderer, or software?

🗪

2:04:29Brian Forgive me, but is the game currently running in the hardware renderer, or software?

🗪

2:04:29Brian Forgive me, but is the game currently running in the hardware renderer, or software?

🗪

2:04:44garryjohanson So I had the idea of maintaining all geometry in a geometrically determined hash table, computing the trajectory of the reflections and looking up where the resulting geometry is. Is that something that would be workable, or am I missing something?

🗪

2:04:44garryjohanson So I had the idea of maintaining all geometry in a geometrically determined hash table, computing the trajectory of the reflections and looking up where the resulting geometry is. Is that something that would be workable, or am I missing something?

🗪

2:04:44garryjohanson So I had the idea of maintaining all geometry in a geometrically determined hash table, computing the trajectory of the reflections and looking up where the resulting geometry is. Is that something that would be workable, or am I missing something?

🗪

2:05:20Blackboard: Using a geometrically determined hash table to compute the trajectory of reflections

2:13:36Brian WeeChat works pretty well, if this is on Linux

🗪

2:13:36Brian WeeChat works pretty well, if this is on Linux

🗪

2:13:36Brian WeeChat works pretty well, if this is on Linux

🗪

2:14:26kriminator117 If you figure it out, tell Autodesk how to fix how MAYA's weight painting looks

🗪

2:14:26kriminator117 If you figure it out, tell Autodesk how to fix how MAYA's weight painting looks

🗪

2:14:26kriminator117 If you figure it out, tell Autodesk how to fix how MAYA's weight painting looks

🗪

2:15:21derpzerker Would it be possible to write a transpiler so that we could take this C code and run it on the GPU trivially after debugging on CPU?

🗪

2:15:21derpzerker Would it be possible to write a transpiler so that we could take this C code and run it on the GPU trivially after debugging on CPU?

🗪

2:15:21derpzerker Would it be possible to write a transpiler so that we could take this C code and run it on the GPU trivially after debugging on CPU?

🗪

2:17:23chaoscom In your lighting formulae, you use negative spans like "-0.25f*Span10". What if everything is supposed to be a displacement off Vertex0? Shouldn't you have "+0.75f" at some point instead of something negative?

🗪

2:17:23chaoscom In your lighting formulae, you use negative spans like "-0.25f*Span10". What if everything is supposed to be a displacement off Vertex0? Shouldn't you have "+0.75f" at some point instead of something negative?

🗪

2:17:23chaoscom In your lighting formulae, you use negative spans like "-0.25f*Span10". What if everything is supposed to be a displacement off Vertex0? Shouldn't you have "+0.75f" at some point instead of something negative?

🗪

2:18:15username_password_admin You're using square, so why not filter bilinearly along the square axes rather than interpolating over the triangles? That should avoid the diagonal seams

🗪

2:18:15username_password_admin You're using square, so why not filter bilinearly along the square axes rather than interpolating over the triangles? That should avoid the diagonal seams

🗪

2:18:15username_password_admin You're using square, so why not filter bilinearly along the square axes rather than interpolating over the triangles? That should avoid the diagonal seams

🗪

2:19:01Blackboard: Interpolating perpendicularly between the squares' edges

2:21:07chaoscom Light doesn't behave linearily. That's why the banding occurs in the first place

🗪

2:21:07chaoscom Light doesn't behave linearily. That's why the banding occurs in the first place

🗪

2:21:07chaoscom Light doesn't behave linearily. That's why the banding occurs in the first place

🗪

2:22:12Wrap it up

🗩

2:22:12Wrap it up

🗩

2:22:12Wrap it up

🗩

2:22:53"It's not live now"^β

🗩

2:22:53"It's not live now"^β

🗩

2:22:53"It's not live now"^β

🗩

Keyboard Navigation

Global Keys

Menu toggling

In-Menu and Index Controls

Quotes and References Menus and Index

Quotes, References and Credits Menus

Filter Menu

Filter and Link Menus

Credits Menu