Visualizing Lighting Values — Handmade Hero — Episode Guide

0:01Note Twitch's latest twitchiness¹

🗩

0:01Note Twitch's latest twitchiness¹

🗩

0:01Note Twitch's latest twitchiness¹

🗩

1:26Determine to do a lighting quality pass

🗩

1:26Determine to do a lighting quality pass

🗩

1:26Determine to do a lighting quality pass

🗩

4:50Debugging the lighting: 1) Trial-and-error eyeballing

🗩

4:50Debugging the lighting: 1) Trial-and-error eyeballing

🗩

4:50Debugging the lighting: 1) Trial-and-error eyeballing

🗩

5:28Debugging the lighting: 2) Create and validate against our own definition of correct lighting behaviour

🗩

5:28Debugging the lighting: 2) Create and validate against our own definition of correct lighting behaviour

🗩

5:28Debugging the lighting: 2) Create and validate against our own definition of correct lighting behaviour

🗩

7:48Regret not having gained much experience with lighting

🗩

7:48Regret not having gained much experience with lighting

🗩

7:48Regret not having gained much experience with lighting

🗩

8:27Demo our ability to debug the raycaster

🏃

8:27Demo our ability to debug the raycaster

🏃

8:27Demo our ability to debug the raycaster

🏃

10:19ComputeLightPropagationWork cycle use: 82%

🏃

10:19ComputeLightPropagationWork cycle use: 82%

🏃

10:19ComputeLightPropagationWork cycle use: 82%

🏃

11:23Approaching lighting transfer

🏃

11:23Approaching lighting transfer

🏃

11:23Approaching lighting transfer

🏃

13:52Approaching lighting transfer: Draw lines through the centre of each grid cube, derived from the octahedral map, with brighter directions drawn longer

🏃

13:52Approaching lighting transfer: Draw lines through the centre of each grid cube, derived from the octahedral map, with brighter directions drawn longer

🏃

13:52Approaching lighting transfer: Draw lines through the centre of each grid cube, derived from the octahedral map, with brighter directions drawn longer

🏃

15:59Try increasing the W modification from ×0.75 to ×0.1 in BuildDiffuseLightMaps()

16:23Demo the slow, but feeding back, lighting transfer, with the determination to step the lighting frame-by-frame

🏃

16:23Demo the slow, but feeding back, lighting transfer, with the determination to step the lighting frame-by-frame

🏃

16:23Demo the slow, but feeding back, lighting transfer, with the determination to step the lighting frame-by-frame

🏃

17:49Determine to draw those octahedral lines in a smaller secondary debug bounds

🏃

17:49Determine to draw those octahedral lines in a smaller secondary debug bounds

🏃

17:49Determine to draw those octahedral lines in a smaller secondary debug bounds

🏃

19:34Introduce DebugDrawOctahedralValues() for UpdateLighting() to call

21:05Note that the diffuse blur is probably providing too much power to the system

🗩

21:05Note that the diffuse blur is probably providing too much power to the system

🗩

21:05Note that the diffuse blur is probably providing too much power to the system

🗩

22:23Make DebugDrawOctahedralValues() draw out our desired debug bounds

29:44Introduce a version of * that takes a v3s and s32, and fix compile errors

31:00Check out our debug boxes

🏃

31:00Check out our debug boxes

🏃

31:00Check out our debug boxes

🏃

31:37Make DebugDrawOctahedralValues() draw our octahedral lines

39:13Check out our octahedral lines

🏃

39:13Check out our octahedral lines

🏃

39:13Check out our octahedral lines

🏃

40:00Make PushLightingRenderValues() Clamp01() the Color

42:46The octahedral lines do not remain white

🏃

42:46The octahedral lines do not remain white

🏃

42:46The octahedral lines do not remain white

🏃

43:10Fix PushLightingRenderValues() to Clamp01() the Line->Color and pass the resulting Color to PushLineSegment()

43:24Our octahedral lines still do not remain white

🏃

43:24Our octahedral lines still do not remain white

🏃

43:24Our octahedral lines still do not remain white

🏃

43:49Fix PushLightingRenderValues() to pass the clamped Color to PushLineSegment() for both ends

44:01Our octahedral lines now remain white

🏃

44:01Our octahedral lines now remain white

🏃

44:01Our octahedral lines now remain white

🏃

44:20Determine to support stepping the lighting frame-by-frame

🏃

44:20Determine to support stepping the lighting frame-by-frame

🏃

44:20Determine to support stepping the lighting frame-by-frame

🏃

44:58Prevent UpdateAndRenderWorld() from calling UpdateLighting()

46:08We get no lighting

🏃

46:08We get no lighting

🏃

46:08We get no lighting

🏃

46:19Make UpdateAndRenderWorld() call UpdateLighting() when we press F9

47:58Step the lighting frame-by-frame with F9

🏃

47:58Step the lighting frame-by-frame with F9

🏃

47:58Step the lighting frame-by-frame with F9

🏃

49:25Delete moonlight variables from GridRayCast()

50:41Our system contains no moonlight, yet our octahedral maps are seeing more light from above

🏃

50:41Our system contains no moonlight, yet our octahedral maps are seeing more light from above

🏃

50:41Our system contains no moonlight, yet our octahedral maps are seeing more light from above

🏃

52:11Make PushLight() draw the light

53:55Check out our lights, and see that our octahedral maps are not really responding to them

🏃

53:55Check out our lights, and see that our octahedral maps are not really responding to them

🏃

53:55Check out our lights, and see that our octahedral maps are not really responding to them

🏃

55:38Move the DebugGridIndex and DebugRayIndex DEBUG_VALUE() initialisers from UpdateLighting() to PushLightingRenderValues()

56:43Edit our DebugGridIndex and DebugRayIndex to 17912 and 100

🏃

56:43Edit our DebugGridIndex and DebugRayIndex to 17912 and 100

🏃

56:43Edit our DebugGridIndex and DebugRayIndex to 17912 and 100

🏃

58:30Make UpdateLighting() set DebugGridIndex and DebugRayIndex to 17912 + (SpatialGrid.CellCount.x × SpatialGrid.CellCount.y) and 100

58:55Edit our DebugGridIndex and DebugRayIndex to 17912 and 99

🏃

58:55Edit our DebugGridIndex and DebugRayIndex to 17912 and 99

🏃

58:55Edit our DebugGridIndex and DebugRayIndex to 17912 and 99

🏃

1:02:10Make UpdateLighting() set DebugGridIndex to 17912

1:02:31Edit our DebugRayIndex to 156, and wonder if we are writing back into the wrong parts of the octahedral map

🏃

1:02:31Edit our DebugRayIndex to 156, and wonder if we are writing back into the wrong parts of the octahedral map

🏃

1:02:31Edit our DebugRayIndex to 156, and wonder if we are writing back into the wrong parts of the octahedral map

🏃

1:04:21Make UpdateLighting() set DebugRayIndex to 156

1:04:30Edit our DebugRayIndex to 155

🏃

1:04:30Edit our DebugRayIndex to 155

🏃

1:04:30Edit our DebugRayIndex to 155

🏃

1:05:35Make UpdateLighting() set DebugRayIndex to 155

1:05:42Try making FullCast() pass an Ignored buffer to GridRayCast(), to rule out the debug system interfering with the SpecAtlas

1:08:57Our problem remains

🏃

1:08:57Our problem remains

🏃

1:08:57Our problem remains

🏃

1:09:38Make FullCast() draw the ExpectedDirection that our rays point into their octahedral map

1:15:38Our rays and octahedral directions are closely aligned

🏃

1:15:38Our rays and octahedral directions are closely aligned

🏃

1:15:38Our rays and octahedral directions are closely aligned

🏃

1:16:52Try increasing the W modification from ×0.1 to ×0.75 in BuildDiffuseLightMaps()

1:17:16Step through the lighting

🏃

1:17:16Step through the lighting

🏃

1:17:16Step through the lighting

🏃

1:17:43Make DebugDrawOctahedralValues() increase the length of our octahedral lines

1:18:34Step through the lighting

🏃

1:18:34Step through the lighting

🏃

1:18:34Step through the lighting

🏃

1:19:07Make DebugDrawOctahedralValues() draw the diffuse atlas lines, introducing DebugDrawColorDir()

1:22:03Step through the lighting, to see that the diffuse atlas grows quickly

🏃

1:22:03Step through the lighting, to see that the diffuse atlas grows quickly

🏃

1:22:03Step through the lighting, to see that the diffuse atlas grows quickly

🏃

1:23:06Temporarily prevent DebugDrawOctahedralValues() from drawing the diffuse atlas lines

1:23:19Step through the lighting, and note the upwards bias

🏃

1:23:19Step through the lighting, and note the upwards bias

🏃

1:23:19Step through the lighting, and note the upwards bias

🏃

1:23:54Prevent GridRayCast() from transferring the light for non-hits

1:27:16Our SpecAtlas is getting light from the wrong directions

🏃

1:27:16Our SpecAtlas is getting light from the wrong directions

🏃

1:27:16Our SpecAtlas is getting light from the wrong directions

🏃

1:28:14Prevent GridRayCast() from transferring the light for hits

1:28:32We correctly see no light

🏃

1:28:32We correctly see no light

🏃

1:28:32We correctly see no light

🏃

1:29:00Our bug possibilities: 1) False hit reports; 2) Incorrectly writing the direction; 3) Incorrect debug drawing

🗩

1:29:00Our bug possibilities: 1) False hit reports; 2) Incorrectly writing the direction; 3) Incorrect debug drawing

🗩

1:29:00Our bug possibilities: 1) False hit reports; 2) Incorrectly writing the direction; 3) Incorrect debug drawing

🗩

1:30:41Scrutinise FullCast() for write-back errors

📖

1:30:41Scrutinise FullCast() for write-back errors

📖

1:30:41Scrutinise FullCast() for write-back errors

📖

1:33:38Check out our apparent ray hit

🏃

1:33:38Check out our apparent ray hit

🏃

1:33:38Check out our apparent ray hit

🏃

1:33:56Prevent GridRayCast() from transferring the light for non-hits

1:34:21Our target ray is correctly not drawn

🏃

1:34:21Our target ray is correctly not drawn

🏃

1:34:21Our target ray is correctly not drawn

🏃

1:34:55Move the debug line drawing branch in GridRayCast() outside of the SomethingHit condition

1:35:25Check out our misinformed octahedral lines

🏃

1:35:25Check out our misinformed octahedral lines

🏃

1:35:25Check out our misinformed octahedral lines

🏃

1:37:00Edit our DebugGridIndex to 17788

🏃

1:37:00Edit our DebugGridIndex to 17788

🏃

1:37:00Edit our DebugGridIndex to 17788

🏃

1:38:08Make UpdateLighting() set DebugGridIndex to 17788

1:38:29Edit our DebugRayIndex, hopefully towards a light

🏃

1:38:29Edit our DebugRayIndex, hopefully towards a light

🏃

1:38:29Edit our DebugRayIndex, hopefully towards a light

🏃

1:42:01Try decreasing the W modification from ×0.75 to ×0.1 in BuildDiffuseLightMaps()

1:42:51Edit our DebugRayIndex to 325, directly towards a light

🏃

1:42:51Edit our DebugRayIndex to 325, directly towards a light

🏃

1:42:51Edit our DebugRayIndex to 325, directly towards a light

🏃

1:43:31Make UpdateLighting() set DebugRayIndex to 325

1:44:16Check out our direct hit

🏃

1:44:16Check out our direct hit

🏃

1:44:16Check out our direct hit

🏃

1:45:16Rename ControlMask to FPControlMask in SetDefaultFPBehavior()²

1:46:40uplinkcoder Q: Maybe just have one light source?

🗪

1:46:40uplinkcoder Q: Maybe just have one light source?

🗪

1:46:40uplinkcoder Q: Maybe just have one light source?

🗪

1:47:07Temporarily prevent GenerateRoom() from adding lamps

1:47:26Check out our one light

🏃

1:47:26Check out our one light

🏃

1:47:26Check out our one light

🏃

1:47:53Try increasing the W modification from ×0.1 to ×0.75 in BuildDiffuseLightMaps()

1:48:00Our octahedral maps are still seeing light from the wrong directions

🏃

1:48:00Our octahedral maps are still seeing light from the wrong directions

🏃

1:48:00Our octahedral maps are still seeing light from the wrong directions

🏃

1:48:38A few words on languages with non-helpful "error-catching" features

🏃

1:48:38A few words on languages with non-helpful "error-catching" features

🏃

1:48:38A few words on languages with non-helpful "error-catching" features

🏃

1:50:56Consider our problem to be an indexing one

🏃

1:50:56Consider our problem to be an indexing one

🏃

1:50:56Consider our problem to be an indexing one

🏃

1:53:12Q&A

🗩

1:53:12Q&A

🗩

1:53:12Q&A

🗩

1:53:29uplinkcoder Q: Try moving the light

🗪

1:53:29uplinkcoder Q: Try moving the light

🗪

1:53:29uplinkcoder Q: Try moving the light

🗪

1:53:42Step through the lighting

🏃

1:53:42Step through the lighting

🏃

1:53:42Step through the lighting

🏃

1:53:57Make UpdateAndRenderWorld() offset the debug light in X

1:54:23See how that affects our octahedral lines

🏃

1:54:23See how that affects our octahedral lines

🏃

1:54:23See how that affects our octahedral lines

🏃

1:55:01Make UpdateAndRenderWorld() offset the debug light the opposite direction in X

1:55:17See how that affects our octahedral lines

🏃

1:55:17See how that affects our octahedral lines

🏃

1:55:17See how that affects our octahedral lines

🏃

1:56:32Prevent UpdateAndRenderWorld() from offsetting the debug light

1:57:02miltondts Q: Why are there two lit areas in the map with only one light source?

🗪

1:57:02miltondts Q: Why are there two lit areas in the map with only one light source?

🗪

1:57:02miltondts Q: Why are there two lit areas in the map with only one light source?

🗪

1:57:14The lighting gets tiled around the world

🏃

1:57:14The lighting gets tiled around the world

🏃

1:57:14The lighting gets tiled around the world

🏃

1:58:37Disable LIGHTING_USE_GRID

1:58:53Our debug visualisation is wrong

🏃

1:58:53Our debug visualisation is wrong

🏃

1:58:53Our debug visualisation is wrong

🏃

1:59:27Let UpdateAndRenderWorld() call UpdateLighting() every frame

1:59:40Our debug visualisation is wrong

🏃

1:59:40Our debug visualisation is wrong

🏃

1:59:40Our debug visualisation is wrong

🏃

2:00:01Scrutinise DebugDrawOctahedralValues() and related functions for bugs

📖

2:00:01Scrutinise DebugDrawOctahedralValues() and related functions for bugs

📖

2:00:01Scrutinise DebugDrawOctahedralValues() and related functions for bugs

📖

2:06:11Check out our octahedral lines while moving the light

🏃

2:06:11Check out our octahedral lines while moving the light

🏃

2:06:11Check out our octahedral lines while moving the light

🏃

2:07:41Let LIGHT_ATLAS_ASSERT() actually Assert()

2:09:51Happily fail to hit any LIGHT_ATLAS_ASSERT()

🏃

2:09:51Happily fail to hit any LIGHT_ATLAS_ASSERT()

🏃

2:09:51Happily fail to hit any LIGHT_ATLAS_ASSERT()

🏃

2:11:03uplinkcoder Q: PushDebugLine() parameter order maybe?

🗪

2:11:03uplinkcoder Q: PushDebugLine() parameter order maybe?

🗪

2:11:03uplinkcoder Q: PushDebugLine() parameter order maybe?

🗪

2:12:05uplinkcoder Q: Direction inverted?

🗪

2:12:05uplinkcoder Q: Direction inverted?

🗪

2:12:05uplinkcoder Q: Direction inverted?

🗪

2:12:11Eyeball the octahedral lines, not seeing inversion

🏃

2:12:11Eyeball the octahedral lines, not seeing inversion

🏃

2:12:11Eyeball the octahedral lines, not seeing inversion

🏃

2:13:35Briefly scrutinise DirectionFromTxTy() for bugs

📖

2:13:35Briefly scrutinise DirectionFromTxTy() for bugs

📖

2:13:35Briefly scrutinise DirectionFromTxTy() for bugs

📖

2:14:29vaualbus Q: Does this not pass through the thing we did for having the 3D texture? Maybe we are tilting the line in the shader

🗪

2:14:29vaualbus Q: Does this not pass through the thing we did for having the 3D texture? Maybe we are tilting the line in the shader

🗪

2:14:29vaualbus Q: Does this not pass through the thing we did for having the 3D texture? Maybe we are tilting the line in the shader

🗪

2:15:17Thank you, everyone

🗩

2:15:17Thank you, everyone

🗩

2:15:17Thank you, everyone

🗩

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