• Global community
• Language:
• Deutsch
• English
• EspaÃ±ol
• FranÃ§ais
• PortuguÃªs
• æ—¥æœ¬èªžã‚³ãƒŸãƒ¥ãƒ‹ãƒ†ã‚£
Dedicated community for Japanese speakers
• í•œêµ­ ì»¤ë®¤ë‹ˆí‹°
Dedicated community for Korean speakers
Exit
• PBR validate node's albedo to grayscale conversion...

# PBR validate node's albedo to grayscale conversion weights?

New Here ,
Feb 03, 2022 Feb 03, 2022

Copy link to clipboard

Copied

Hey Substance3D team and community!

I was pinging Substance3D on twitter about this a while ago, but didn't get a response (https://twitter.com/olafhaag3D/status/1484321640515129352).

I was looking into the PBR BaseColor / Metallic validate node to better understand how safe values are determined. As a first step, it converts the albedo to a grayscale image, presumably to get luminance values.

I was a bit surprised about the channel weights .299R+.587G+.114B, which are typically used to calculate rec601 luma, not used with sRGB/Rec. 709 values.

Why is that? Aren't the colors linearized internally? I try to understand why the colorimetric weights (.2126R+.7152G+.0722B) weren't used, assuming nodes use a linear colorspace.

More on grayscale/luminance/luma conversion and the difference: https://en.wikipedia.org/wiki/Grayscale

The PBR Albedo safe color node, on the other hand, treats all channels equally.

I assume the 0.19 and 0.12 values in the function are just the range between 50 and 30 sRGB, which kind of hints at non-linear sRGB colorspace is used in nodes. But since sRGB are non-linear values, how useful is a linear interpolation?

I appreciate any information to clear up my confusion. Thanks!

TOPICS
Substance Graph

Views

66

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more

Have something to add?

Join the conversation
Resources
Guides
Assets
Common issues