Skip to content
This repository has been archived by the owner on Feb 28, 2024. It is now read-only.

[BUG-228781] [EEP] Specular color interpreted wrong (too bright) from point lights. #6775

Closed
6 tasks
sl-service-account opened this issue May 19, 2020 · 4 comments

Comments

@sl-service-account
Copy link

sl-service-account commented May 19, 2020

What just happened?

Surfaces with a non-white specular texture and/or with a non-white specular color shine a lot brighter than pre-EEP. The darker the specular color, the more apparent the difference. This breaks a LOT of content including mesh body skin shine, making mesh avatars look like barbie dolls. This bug is a major one because the longer it stays, the more people will modify the shine on their furniture and bodies, thinking they have to adjust. Problem is, a lot of products include shine presets set via scripts, and those presets are completely wrong in EEP. For example, some of my products have a "leather" preset which means 75 glossiness, no environment and 25% grey for the shine color. It looked fine pre-EEP, now it looks like dull plastic.

To repro :

  • Set the environment to midnight.

  • Create a sphere, tint it black.

  • Add a white specular texture to it, set its glossiness to 100, no environment, and most important : set its shine color to 13/13/13 or something equally dark.

  • Rez a point light near it, keep the default values, no need to make it a projector, and look at the surface of the black sphere.

  • Observe that there is a noticeable specular reflection in EEP, while pre-EEP the reflection was barely visible.

    What were you doing when it happened?

    Trying to figure out why there is a difference in rendering between pre-EEP and EEP on shiny surfaces.

    What were you expecting to happen instead?

    The specular reflection should be exactly the same as pre-EEP, in shine intensity, diameter and blurriness.

    Other information

    Please note that although I'm using RLV, it is based on the SL viewer and this bug occurs with the EEP SL viewer as well.

Attachments

Links

Related

Duplicates

Original Jira Fields
Field Value
Issue BUG-228781
Summary [EEP] Specular color interpreted wrong (too bright) from point lights.
Type Bug
Priority Unset
Status Closed
Resolution Accepted
Labels whirly-eep
Reporter Marine Kelley (marine.kelley)
Created at 2020-05-19T15:29:38Z
Updated at 2021-01-30T19:09:36Z
{
  'Build Id': 'unset',
  'Business Unit': ['Platform'],
  'Date of First Response': '2020-05-20T11:37:59.591-0500',
  "Is there anything you'd like to add?": "Please note that although I'm using RLV, it is based on the SL viewer and this bug occurs with the EEP SL viewer as well.",
  'ReOpened Count': 0.0,
  'Severity': 'Unset',
  'System': 'SL Viewer',
  'Target Viewer Version': 'viewer-development',
  'What just happened?': '- Set the environment to midnight.\r\nSurfaces with a non-white specular texture and/or with a non-white specular color shine a lot brighter than pre-EEP. The darker the specular color, the more apparent the difference. This breaks a LOT of content including mesh body skin shine, making mesh avatars look like barbie dolls. This bug is a major one because the longer it stays, the more people will modify the shine on their furniture and bodies, thinking they have to adjust. Problem is, a lot of products include shine presets set via scripts, and those presets are completely wrong in EEP. For example, some of my products have a "leather" preset which means 75 glossiness, no environment and 25% grey for the shine color. It looked fine pre-EEP, now it looks like dull plastic.\r\n\r\nTo repro :\r\n- Create a sphere, tint it black.\r\n- Add a white specular texture to it, set its glossiness to 100, no environment, and most important : set its shine color to 13/13/13 or something equally dark.\r\n- Rez a point light near it, keep the default values, no need to make it a projector, and look at the surface of the black sphere.',
  'What were you doing when it happened?': 'Trying to figure out why there is a difference in rendering between pre-EEP and EEP on shiny surfaces.',
  'What were you expecting to happen instead?': '- Observe that there is a noticeable specular reflection in EEP, while pre-EEP the reflection was barely visible.',
}
@sl-service-account
Copy link
Author

Whirly Fizzle commented at 2020-05-20T16:38:00Z

There is some discussion about this bug in the comments on BUG-228581 too.

@sl-service-account
Copy link
Author

Marine Kelley commented at 2020-05-20T16:59:05Z

I know, I commented there too, but that bug is about specular reflections being too weak, while this bug here is about them being too strong. And I preferred creating a bug with a clear and simple repro to make it easier to spot and fix the issue because "too strong" and "too weak" are a bit subjective a measure.

And I really wanted to insist on the fact that this bug breaks existing content and feared that a single comment wouldn't be sufficient.

@sl-service-account
Copy link
Author

Chorazin Allen commented at 2020-05-22T12:46:25Z

This looks to be due to the addition of diff.rgb = srgb_to_linear(diff.rgb); in shaders\class1\deferred\multiPointLightF.glsl as part of commit 653133b9c035e7c34563b199e991189ca3046092 however this was a big commit so a look at all the possible ramifications of reverting it is needed

@sl-service-account
Copy link
Author

Whirly Fizzle commented at 2020-07-03T18:28:12Z

Note comments from Geenz on the commit that fixes this bug.
https://bitbucket.org/lindenlab/viewer/commits/aa5bd5d6f530f576556c9256714f01f2a35854bf#comment-9035516

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

No branches or pull requests

1 participant