Debugging Voxel Interpolation
References ▼
GitHub
lucasg/​dependencies
[1]2:05:43
SadBart-Small.gif
[2]2:11:37
GitHub User Content
[3]2:17:48
Filter mode:
Topics
Media
debug visualisation
hardware
lighting
rendering
(null topic)
🗹Administrivia
🗪Chat Comment
🏃In-Game
🖮Programming🟉
📖Research
🗩Speech
🎭
🏟
🔗
Link to: current timestamp
Credits
Host
Casey Muratori
Indexer
Matt Mascarenhas
?
?

Keyboard Navigation

Global Keys

[, < / ], > Jump to previous / next episode
W, K, P / S, J, N Jump to previous / next timestamp
t / T Toggle theatre / SUPERtheatre mode
V Revert filter to original state Y Select link (requires manual Ctrl-c)

Menu toggling

q Quotes r References f Filter y Link c Credits

In-Menu and Index Controls

a
w
s
d
h j k l
←
↑
↓
→

Esc Close menu / unfocus timestamp

Quotes and References Menus and Index

Enter Jump to timestamp

Quotes, References and Credits Menus

o Open URL (in new tab)

Filter Menu

x, Space Toggle category and focus next
X, ShiftSpace Toggle category and focus previous
v Invert topics / media as per focus

Filter and Link Menus

z Toggle filter / linking mode

Credits Menu

Enter Open URL (in new tab)
0:07Recap and set the stage for the day
🗩
0:07Recap and set the stage for the day
🗩
0:07Recap and set the stage for the day
🗩
2:01Run the game to see our current lighting situation
🏃
2:01Run the game to see our current lighting situation
🏃
2:01Run the game to see our current lighting situation
🏃
5:42Consider the current problems in the lighting
🏃
5:42Consider the current problems in the lighting
🏃
5:42Consider the current problems in the lighting
🏃
9:14Determine to validate RayCast()
🗩
9:14Determine to validate RayCast()
🗩
9:14Determine to validate RayCast()
🗩
12:44Make RayCast() tint the light grey
12:44Make RayCast() tint the light grey
12:44Make RayCast() tint the light grey
14:13Run the game to see less light in the scene
🏃
14:13Run the game to see less light in the scene
🏃
14:13Run the game to see less light in the scene
🏃
14:41Investigate the possibility that we are wrapping
14:41Investigate the possibility that we are wrapping
14:41Investigate the possibility that we are wrapping
16:46Make GatherFinalLighting() set the IncidentLight to pure red
16:46Make GatherFinalLighting() set the IncidentLight to pure red
16:46Make GatherFinalLighting() set the IncidentLight to pure red
17:21Run the game and do not see a red scene
🏃
17:21Run the game and do not see a red scene
🏃
17:21Run the game and do not see a red scene
🏃
18:20Run the game and switch away from the "just emitters" debug visualisation
🏃
18:20Run the game and switch away from the "just emitters" debug visualisation
🏃
18:20Run the game and switch away from the "just emitters" debug visualisation
🏃
19:04Scrutinise the lighting code
📖
19:04Scrutinise the lighting code
📖
19:04Scrutinise the lighting code
📖
22:46Make CompileZBiasProgram() multiply the DirFalloff in to the final result
22:46Make CompileZBiasProgram() multiply the DirFalloff in to the final result
22:46Make CompileZBiasProgram() multiply the DirFalloff in to the final result
27:20Run the game to see a little difference, but still no ambient lighting
🏃
27:20Run the game to see a little difference, but still no ambient lighting
🏃
27:20Run the game to see a little difference, but still no ambient lighting
🏃
28:27Read through GatherFinalLighting() and OutputLightingTextures()
📖
28:27Read through GatherFinalLighting() and OutputLightingTextures()
📖
28:27Read through GatherFinalLighting() and OutputLightingTextures()
📖
30:17Step in to OutputLightingTextures() and inspect the IncidentLight, Colour and Directional values
30:17Step in to OutputLightingTextures() and inspect the IncidentLight, Colour and Directional values
30:17Step in to OutputLightingTextures() and inspect the IncidentLight, Colour and Directional values
32:15Note that we don't have a way to directly visualise lighting quads
🏃
32:15Note that we don't have a way to directly visualise lighting quads
🏃
32:15Note that we don't have a way to directly visualise lighting quads
🏃
35:05Make CompileZBiasProgram() renormalise the contribution colour if the alpha is greater than 0
35:05Make CompileZBiasProgram() renormalise the contribution colour if the alpha is greater than 0
35:05Make CompileZBiasProgram() renormalise the contribution colour if the alpha is greater than 0
36:04Run the game to see still nonsense results
🏃
36:04Run the game to see still nonsense results
🏃
36:04Run the game to see still nonsense results
🏃
36:40Hard set the lighting colour to green in CompileZBiasProgram()
36:40Hard set the lighting colour to green in CompileZBiasProgram()
36:40Hard set the lighting colour to green in CompileZBiasProgram()
36:59Run the game to see pure green, except on the blue surface
🏃
36:59Run the game to see pure green, except on the blue surface
🏃
36:59Run the game to see pure green, except on the blue surface
🏃
37:37Verify that all pure colours hard set by CompileZBiasProgram() produce the expected results
37:37Verify that all pure colours hard set by CompileZBiasProgram() produce the expected results
37:37Verify that all pure colours hard set by CompileZBiasProgram() produce the expected results
38:56Run to see pure white and note that the voxel interpolation isn't totally busted
🏃
38:56Run to see pure white and note that the voxel interpolation isn't totally busted
🏃
38:56Run to see pure white and note that the voxel interpolation isn't totally busted
🏃
39:27Hard set the LightC to white and the DirFalloff to 1.0f in CompileZBiasProgram()
39:27Hard set the LightC to white and the DirFalloff to 1.0f in CompileZBiasProgram()
39:27Hard set the LightC to white and the DirFalloff to 1.0f in CompileZBiasProgram()
41:06Run the game to see white light, but some incorrectly dark spots
🏃
41:06Run the game to see white light, but some incorrectly dark spots
🏃
41:06Run the game to see white light, but some incorrectly dark spots
🏃
43:14Try to make CompileZBiasProgram() set the colour to pure yellow where there is no light sample
43:14Try to make CompileZBiasProgram() set the colour to pure yellow where there is no light sample
43:14Try to make CompileZBiasProgram() set the colour to pure yellow where there is no light sample
44:25Run the game to see spots of yellow in unexpected places
🏃
44:25Run the game to see spots of yellow in unexpected places
🏃
44:25Run the game to see spots of yellow in unexpected places
🏃
44:51Temporarily prevent CompileZBiasProgram() from setting the colour to yellow, and instead set the SurfaceReflect directly to the light colour, rather than multiplying it in
44:51Temporarily prevent CompileZBiasProgram() from setting the colour to yellow, and instead set the SurfaceReflect directly to the light colour, rather than multiplying it in
44:51Temporarily prevent CompileZBiasProgram() from setting the colour to yellow, and instead set the SurfaceReflect directly to the light colour, rather than multiplying it in
45:45Run the game to see the incoming light colour
🏃
45:45Run the game to see the incoming light colour
🏃
45:45Run the game to see the incoming light colour
🏃
46:04Make CompileZBiasProgram() set the colour to pure yellow where there is no light sample
46:04Make CompileZBiasProgram() set the colour to pure yellow where there is no light sample
46:04Make CompileZBiasProgram() set the colour to pure yellow where there is no light sample
46:20Run the game to see yellow where we cannot compute lighting results
🏃
46:20Run the game to see yellow where we cannot compute lighting results
🏃
46:20Run the game to see yellow where we cannot compute lighting results
🏃
47:09Try to make CompileZBiasProgram() set the colour to red
47:09Try to make CompileZBiasProgram() set the colour to red
47:09Try to make CompileZBiasProgram() set the colour to red
47:18Run the game to see some red places
🏃
47:18Run the game to see some red places
🏃
47:18Run the game to see some red places
🏃
47:39Make CompileZBiasProgram() track the Contrib value all the way down the pipeline and introduce VoxLerp() to use this value
47:39Make CompileZBiasProgram() track the Contrib value all the way down the pipeline and introduce VoxLerp() to use this value
47:39Make CompileZBiasProgram() track the Contrib value all the way down the pipeline and introduce VoxLerp() to use this value
50:56Run the game to see that it already looks better
🏃
50:56Run the game to see that it already looks better
🏃
50:56Run the game to see that it already looks better
🏃
51:34Begin to enable VoxLerp() to handle the case when the lighting contribution is 0
51:34Begin to enable VoxLerp() to handle the case when the lighting contribution is 0
51:34Begin to enable VoxLerp() to handle the case when the lighting contribution is 0
53:01Consider how better to handle voxels that have no lighting contribution
🗩
53:01Consider how better to handle voxels that have no lighting contribution
🗩
53:01Consider how better to handle voxels that have no lighting contribution
🗩
53:38Revert the passing of Contrib() all down the pipe in CompileZBiasProgram()
53:38Revert the passing of Contrib() all down the pipe in CompileZBiasProgram()
53:38Revert the passing of Contrib() all down the pipe in CompileZBiasProgram()
54:17Take a quick break
🗹
54:17Take a quick break
🗹
54:17Take a quick break
🗹
55:03Run the game to see our current red-tinted lighting situation
🏃
55:03Run the game to see our current red-tinted lighting situation
🏃
55:03Run the game to see our current red-tinted lighting situation
🏃
55:44Make OutputLightingTextures() gather the colour of voxels whose lighting contribution is 0 from their neighbour
55:44Make OutputLightingTextures() gather the colour of voxels whose lighting contribution is 0 from their neighbour
55:44Make OutputLightingTextures() gather the colour of voxels whose lighting contribution is 0 from their neighbour
1:00:48Run the game to see no difference
🏃
1:00:48Run the game to see no difference
🏃
1:00:48Run the game to see no difference
🏃
1:01:27Make OutputLightingTextures() zero-initialise the Neighbour
1:01:27Make OutputLightingTextures() zero-initialise the Neighbour
1:01:27Make OutputLightingTextures() zero-initialise the Neighbour
1:01:36Run the game and consider that the artifacts may only be occurring on voxels at the edge of the scene
🏃
1:01:36Run the game and consider that the artifacts may only be occurring on voxels at the edge of the scene
🏃
1:01:36Run the game and consider that the artifacts may only be occurring on voxels at the edge of the scene
🏃
1:04:47Temporarily assert in OutputLightingTextures() that Neighbor != 0
1:04:47Temporarily assert in OutputLightingTextures() that Neighbor != 0
1:04:47Temporarily assert in OutputLightingTextures() that Neighbor != 0
1:05:10Run the game to determine that we are doing some filling
🏃
1:05:10Run the game to determine that we are doing some filling
🏃
1:05:10Run the game to determine that we are doing some filling
🏃
1:07:44Make CompileZBiasProgram() colour the voxels by their location
1:07:44Make CompileZBiasProgram() colour the voxels by their location
1:07:44Make CompileZBiasProgram() colour the voxels by their location
1:09:10Run the game to see this debug visualisation confirming that the voxel interpolation is fine
🏃
1:09:10Run the game to see this debug visualisation confirming that the voxel interpolation is fine
🏃
1:09:10Run the game to see this debug visualisation confirming that the voxel interpolation is fine
🏃
1:10:02Make CompileZBiasProgram() colour the voxels green / red to indicate the presence / absence of lighting samples
1:10:02Make CompileZBiasProgram() colour the voxels green / red to indicate the presence / absence of lighting samples
1:10:02Make CompileZBiasProgram() colour the voxels green / red to indicate the presence / absence of lighting samples
1:10:49Run the game to see this debug visualisation
🏃
1:10:49Run the game to see this debug visualisation
🏃
1:10:49Run the game to see this debug visualisation
🏃
1:12:14Toggle OutputLighting() on
1:12:14Toggle OutputLighting() on
1:12:14Toggle OutputLighting() on
1:12:27Run the game to see the lighting solution visualised in voxel form
🏃
1:12:27Run the game to see the lighting solution visualised in voxel form
🏃
1:12:27Run the game to see the lighting solution visualised in voxel form
🏃
1:14:06Change OutputLightingTextures() to perform the Neighbor filling in a for() loop expanded to the whole voxel
1:14:06Change OutputLightingTextures() to perform the Neighbor filling in a for() loop expanded to the whole voxel
1:14:06Change OutputLightingTextures() to perform the Neighbor filling in a for() loop expanded to the whole voxel
1:17:36Run the game to see that we are filling nearly everything
🏃
1:17:36Run the game to see that we are filling nearly everything
🏃
1:17:36Run the game to see that we are filling nearly everything
🏃
1:17:55Introduce ClampVox() to clamp the voxel lookups only to the neighbouring voxels that are in the scene
1:17:55Introduce ClampVox() to clamp the voxel lookups only to the neighbouring voxels that are in the scene
1:17:55Introduce ClampVox() to clamp the voxel lookups only to the neighbouring voxels that are in the scene
1:20:17Run the game to see that the red patches are gone
🏃
1:20:17Run the game to see that the red patches are gone
🏃
1:20:17Run the game to see that the red patches are gone
🏃
1:20:55Make CompileZBiasProgram() take the LIGHT_LOOKUP positions to enable easy tweaking of the clamp
1:20:55Make CompileZBiasProgram() take the LIGHT_LOOKUP positions to enable easy tweaking of the clamp
1:20:55Make CompileZBiasProgram() take the LIGHT_LOOKUP positions to enable easy tweaking of the clamp
1:23:37Run the game to see the same debug visualisation
🏃
1:23:37Run the game to see the same debug visualisation
🏃
1:23:37Run the game to see the same debug visualisation
🏃
1:23:50Remove the red / green debug visualisation from CompileZBiasProgram()
1:23:50Remove the red / green debug visualisation from CompileZBiasProgram()
1:23:50Remove the red / green debug visualisation from CompileZBiasProgram()
1:24:17Run the game to see spherical weirdness
🏃
1:24:17Run the game to see spherical weirdness
🏃
1:24:17Run the game to see spherical weirdness
🏃
1:24:50Consider how CompileZBiasProgram() should compute the Contrib value
🗩
1:24:50Consider how CompileZBiasProgram() should compute the Contrib value
🗩
1:24:50Consider how CompileZBiasProgram() should compute the Contrib value
🗩
1:29:36Make CompileZBiasProgram() multiply in the VoxInvCelDim to the Contrib value
1:29:36Make CompileZBiasProgram() multiply in the VoxInvCelDim to the Contrib value
1:29:36Make CompileZBiasProgram() multiply in the VoxInvCelDim to the Contrib value
1:32:30Run the game and still see the weird gradient falloff patterns
🏃
1:32:30Run the game and still see the weird gradient falloff patterns
🏃
1:32:30Run the game and still see the weird gradient falloff patterns
🏃
1:33:12Temporarily make CompileZBiasProgram() hard code the Contrib to 1
1:33:12Temporarily make CompileZBiasProgram() hard code the Contrib to 1
1:33:12Temporarily make CompileZBiasProgram() hard code the Contrib to 1
1:34:27Run the game to see no gradient falloff pattern
🏃
1:34:27Run the game to see no gradient falloff pattern
🏃
1:34:27Run the game to see no gradient falloff pattern
🏃
1:34:35Try to make CompileZBiasProgram() factor in the dimensions of all neighbouring voxels to the Contrib value
1:34:35Try to make CompileZBiasProgram() factor in the dimensions of all neighbouring voxels to the Contrib value
1:34:35Try to make CompileZBiasProgram() factor in the dimensions of all neighbouring voxels to the Contrib value
1:36:09Run the game to see that it's a little better
🏃
1:36:09Run the game to see that it's a little better
🏃
1:36:09Run the game to see that it's a little better
🏃
1:36:29Make CompileZBiasProgram() factor in to the Contrib value merely the lighting sample distance
1:36:29Make CompileZBiasProgram() factor in to the Contrib value merely the lighting sample distance
1:36:29Make CompileZBiasProgram() factor in to the Contrib value merely the lighting sample distance
1:37:12Run the game to see no gradient falloff
🏃
1:37:12Run the game to see no gradient falloff
🏃
1:37:12Run the game to see no gradient falloff
🏃
1:37:36Prevent CompileZBiasProgram() from stomping on the LightC
1:37:36Prevent CompileZBiasProgram() from stomping on the LightC
1:37:36Prevent CompileZBiasProgram() from stomping on the LightC
1:38:02Run the game to see fewer artifacts
🏃
1:38:02Run the game to see fewer artifacts
🏃
1:38:02Run the game to see fewer artifacts
🏃
1:38:24Make CompileZBiasProgram() hard set the DirFalloff to 1.0f
1:38:24Make CompileZBiasProgram() hard set the DirFalloff to 1.0f
1:38:24Make CompileZBiasProgram() hard set the DirFalloff to 1.0f
1:39:18Run the game to see pure red everywhere
🏃
1:39:18Run the game to see pure red everywhere
🏃
1:39:18Run the game to see pure red everywhere
🏃
1:39:58Make CompileZBiasProgram() set the LightC to the LightD
1:39:58Make CompileZBiasProgram() set the LightC to the LightD
1:39:58Make CompileZBiasProgram() set the LightC to the LightD
1:41:00Run the game to see this debug visualisation
🏃
1:41:00Run the game to see this debug visualisation
🏃
1:41:00Run the game to see this debug visualisation
🏃
1:45:00Read through the lighting code to consider which routines we could verify
📖
1:45:00Read through the lighting code to consider which routines we could verify
📖
1:45:00Read through the lighting code to consider which routines we could verify
📖
1:48:44Make CompileZBiasProgram() set the LightC to WorldN
1:48:44Make CompileZBiasProgram() set the LightC to WorldN
1:48:44Make CompileZBiasProgram() set the LightC to WorldN
1:48:56Run the game to see that the world normals look reasonable
🏃
1:48:56Run the game to see that the world normals look reasonable
🏃
1:48:56Run the game to see that the world normals look reasonable
🏃
1:49:56Prevent CompileZBiasProgram() from stomping on the LightC
1:49:56Prevent CompileZBiasProgram() from stomping on the LightC
1:49:56Prevent CompileZBiasProgram() from stomping on the LightC
1:50:02Run the game and consider what may be causing the artifacts
🏃
1:50:02Run the game and consider what may be causing the artifacts
🏃
1:50:02Run the game and consider what may be causing the artifacts
🏃
1:53:00Prevent GatherFinalLighting() from stomping on the IncidentLight and AverageDirectionToLight
1:53:00Prevent GatherFinalLighting() from stomping on the IncidentLight and AverageDirectionToLight
1:53:00Prevent GatherFinalLighting() from stomping on the IncidentLight and AverageDirectionToLight
1:53:11Run the game to see the lighting solution
🏃
1:53:11Run the game to see the lighting solution
🏃
1:53:11Run the game to see the lighting solution
🏃
1:53:53Prevent RayCast() from tinting the light grey
1:53:53Prevent RayCast() from tinting the light grey
1:53:53Prevent RayCast() from tinting the light grey
1:54:17Run the game and note that we're getting closer to sane
🏃
1:54:17Run the game and note that we're getting closer to sane
🏃
1:54:17Run the game and note that we're getting closer to sane
🏃
1:54:48Make CompileZBiasProgram() attenuate the SurfaceReflect by the Light
1:54:48Make CompileZBiasProgram() attenuate the SurfaceReflect by the Light
1:54:48Make CompileZBiasProgram() attenuate the SurfaceReflect by the Light
1:55:15Run the game to see that it's getting a lot better
🏃
1:55:15Run the game to see that it's getting a lot better
🏃
1:55:15Run the game to see that it's getting a lot better
🏃
1:58:33Consider the next steps
🏃
1:58:33Consider the next steps
🏃
1:58:33Consider the next steps
🏃
1:59:48Re-enable RayCast() to tint the light grey
1:59:48Re-enable RayCast() to tint the light grey
1:59:48Re-enable RayCast() to tint the light grey
2:00:08Run the game to see the brighter scene
🏃
2:00:08Run the game to see the brighter scene
🏃
2:00:08Run the game to see the brighter scene
🏃
2:00:44Make GatherFinalLighting() divide the RayCount in to the IncidentLight
2:00:44Make GatherFinalLighting() divide the RayCount in to the IncidentLight
2:00:44Make GatherFinalLighting() divide the RayCount in to the IncidentLight
2:01:57Run the game to see the ambient grey light
🏃
2:01:57Run the game to see the ambient grey light
🏃
2:01:57Run the game to see the ambient grey light
🏃
2:03:15Q&A
🗩
2:03:15Q&A
🗩
2:03:15Q&A
🗩
2:03:47sadisttrombone Q: Regarding win calc, the size of the window determines how many buttons are shown. You can either stretch the window vertically or you can press the up arrow button (in the UI) to make the shown buttons shift into the other set like 1/x and e^x
🗪
2:03:47sadisttrombone Q: Regarding win calc, the size of the window determines how many buttons are shown. You can either stretch the window vertically or you can press the up arrow button (in the UI) to make the shown buttons shift into the other set like 1/x and e^x
🗪
2:03:47sadisttrombone Q: Regarding win calc, the size of the window determines how many buttons are shown. You can either stretch the window vertically or you can press the up arrow button (in the UI) to make the shown buttons shift into the other set like 1/x and e^x
🗪
2:04:45vkar2 Q: I think the error in the color are in the same position where the normals calculation looks funny
🗪
2:04:45vkar2 Q: I think the error in the color are in the same position where the normals calculation looks funny
🗪
2:04:45vkar2 Q: I think the error in the color are in the same position where the normals calculation looks funny
🗪
2:05:43radovik Q: Not sure if you heard, but someone rewrote depends.exe for win 101
🗪
2:05:43radovik Q: Not sure if you heard, but someone rewrote depends.exe for win 101
🗪
2:05:43radovik Q: Not sure if you heard, but someone rewrote depends.exe for win 101
🗪
2:08:22dragoonx6 Q: Seeing you used emacs without code completion, and now 4coder, how do you work without code completion? Don't you think it could help you with both productivity and discoverability? I'm a fan of YouCompleteMe for vim
🗪
2:08:22dragoonx6 Q: Seeing you used emacs without code completion, and now 4coder, how do you work without code completion? Don't you think it could help you with both productivity and discoverability? I'm a fan of YouCompleteMe for vim
🗪
2:08:22dragoonx6 Q: Seeing you used emacs without code completion, and now 4coder, how do you work without code completion? Don't you think it could help you with both productivity and discoverability? I'm a fan of YouCompleteMe for vim
🗪
2:10:26longboolean Q: When the lighting is in realtime will that tend to hide some of the more minor artifacts?
🗪
2:10:26longboolean Q: When the lighting is in realtime will that tend to hide some of the more minor artifacts?
🗪
2:10:26longboolean Q: When the lighting is in realtime will that tend to hide some of the more minor artifacts?
🗪
2:11:37dragoonx6 Q: Like in this gif2
🗪
2:11:37dragoonx6 Q: Like in this gif2
🗪
2:11:37dragoonx6 Q: Like in this gif2
🗪
2:13:05dragoonx6 Q: It's showing a heads up display of the possible arguments
🗪
2:13:05dragoonx6 Q: It's showing a heads up display of the possible arguments
🗪
2:13:05dragoonx6 Q: It's showing a heads up display of the possible arguments
🗪
2:13:56sired22 Q: I know finding a bad API is not hard, but out of curiosity did you ever run into an API that was great to work with? One that you seemed to magically just click with?
🗪
2:13:56sired22 Q: I know finding a bad API is not hard, but out of curiosity did you ever run into an API that was great to work with? One that you seemed to magically just click with?
🗪
2:13:56sired22 Q: I know finding a bad API is not hard, but out of curiosity did you ever run into an API that was great to work with? One that you seemed to magically just click with?
🗪
2:17:48dragoonx6 What about this gif?3
🗪
2:17:48dragoonx6 What about this gif?3
🗪
2:17:48dragoonx6 What about this gif?3
🗪
2:20:40Close everything down
🗩
2:20:40Close everything down
🗩
2:20:40Close everything down
🗩