3.6.1 cycles light tree will cause normal map looks not right #110674

New Issue

MOTO · 2023-08-01T10:46:59+02:00

MOTO commented

2023-08-01 10:46:59 +02:00

System Information
Operating system: windows 11
Graphics card: RTX4090

Blender Version
Broken: 3.6.1
Worked: 3.5.1
Short description of error

When the normal strength value is over 1.0 ( even just a little bit higher like 1.05), and the light tree function is enabled, the normal map will look wrong ( or you can say too strong) and when the light tree is off, the material will look quite different.

But all these differences won't happen if the normal strength is 1.0 or below. The normal map result won't change this much whether the light tree is on or off.

And 3.5.1 doesn't have this problem.

Exact steps for others to reproduce the error
download the blend file and try the combination I've shown in the snapshot
https://www.udrop.com/Ks3C/normal_light_tree_glitch.blend

Based on the default startup or an attached .blend file (as simple as possible).

**System Information** Operating system: windows 11 Graphics card: RTX4090 **Blender Version** Broken: 3.6.1 Worked: 3.5.1 **Short description of error** When the normal strength value is over 1.0 ( even just a little bit higher like 1.05), and the light tree function is enabled, the normal map will look wrong ( or you can say too strong) and when the light tree is off, the material will look quite different. But all these differences won't happen if the normal strength is 1.0 or below. The normal map result won't change this much whether the light tree is on or off. And 3.5.1 doesn't have this problem. **Exact steps for others to reproduce the error** download the blend file and try the combination I've shown in the snapshot https://www.udrop.com/Ks3C/normal_light_tree_glitch.blend Based on the default startup or an attached .blend file (as simple as possible).

2023-08-01 161400.jpg

917 KiB

2023-08-01 161416.jpg

879 KiB

2023-08-01 161407.jpg

718 KiB

2023-08-01 161422.jpg

931 KiB

MOTO added the

labels 2023-08-01 10:47:00 +02:00

Pratik Borhade referenced this issue

2023-08-01 10:56:26 +02:00

3.6.1 cycles light tree makes normal map looks not right #110673

Pratik Borhade commented

2023-08-01 12:03:42 +02:00

Hi, thanks for the report. Can confirm. Displacement is also comparatively high in 3.6
Multiple factors are involved here, not sure if light tree is culprit here.
I suspect Normal map and displacement node are affecting the result

Hi, thanks for the report. Can confirm. Displacement is also comparatively high in 3.6 Multiple factors are involved here, not sure if light tree is culprit here. I suspect Normal map and displacement node are affecting the result

Pratik Borhade commented

2023-08-01 12:10:45 +02:00

Results are improved though in 4.0
Could you check again in newer 3.6.1 / 4.0 build: https://builder.blender.org/download/daily/

Results are improved though in 4.0 Could you check again in newer 3.6.1 / 4.0 build: https://builder.blender.org/download/daily/

Pratik Borhade added

Status

Needs Information from User

and removed

Status

Needs Triage

labels 2023-08-01 12:10:53 +02:00

MOTO commented

2023-08-01 12:57:20 +02:00

In 4.0 ALPHA, it still has noticeable differences but not that "glitch-ish"

2023-08-01 185530.png

1.6 MiB

2023-08-01 185540.png

1.6 MiB

👀 1

Pratik Borhade added

Status

Needs Triage

and removed

Status

Needs Information from User

labels 2023-08-01 13:16:33 +02:00

Philipp Oeser commented

2023-08-07 11:14:00 +02:00

Might be related to #110271 / dfe7b839bc , will check...

Might be related to #110271 / dfe7b839bc , will check...

Pratik Borhade commented

2023-08-22 08:33:17 +02:00

@lichtwerk , any updates? :)
Render results initially changed between 3eff28a158 - 2de2db0f79
Later some commit between 8191b152ec - 3e88a2f44c improved the result.
Confirming for now.

@lichtwerk , any updates? :) Render results initially changed between 3eff28a158ca - 2de2db0f79d1 Later some commit between 8191b152ecf1 - 3e88a2f44c41 improved the result. Confirming for now.

Pratik Borhade added

and removed

labels 2023-08-22 08:33:38 +02:00

Philipp Oeser commented

2023-08-22 09:07:38 +02:00

@lichtwerk , any updates? :)
Render results initially changed between 3eff28a158 - 2de2db0f79
Later some commit between 8191b152ec - 3e88a2f44c improved the result.
Confirming for now.

Ah, somehow this slipped through, will try to get down to specific commits...

> @lichtwerk , any updates? :) > Render results initially changed between 3eff28a158ca - 2de2db0f79d1 > Later some commit between 8191b152ecf1 - 3e88a2f44c41 improved the result. > Confirming for now. Ah, somehow this slipped through, will try to get down to specific commits...

Philipp Oeser commented

2023-08-22 17:41:07 +02:00

So yeah, this first changed with dfe7b839bc, so it is basically the same as #110271 as mentioned above (but this also means this is actually intended behavior).

NOTE: mentioned in the release notes:
https://wiki.blender.org/wiki/Reference/Release_Notes/3.6/Cycles
#105776

The second change (which "improved the result") came with b767a62f32

So I guess this leaves us with two "fix" commits, both intended.
And while the result differs, there does not seem to be a bug here.
Of course feel free to comment again if there is some misunderstanding, but will close for now.

So yeah, this first changed with dfe7b839bca6a66c3eefa2b5968c061b425e2df1, so it is basically the same as #110271 as mentioned above (but this also means this is actually intended behavior). NOTE: mentioned in the release notes: https://wiki.blender.org/wiki/Reference/Release_Notes/3.6/Cycles #105776 The second change (which "improved the result") came with b767a62f32ad5d6342ea57363776d7224db63043 So I guess this leaves us with two "fix" commits, both intended. And while the result differs, there does not seem to be a bug here. Of course feel free to comment again if there is some misunderstanding, but will close for now.

Philipp Oeser closed this issue

2023-08-22 17:41:09 +02:00

Blender Bot added

and removed

labels 2023-08-22 17:41:16 +02:00

Sign in to join this conversation.