Octane 4.01 rendering brighter than 3.08

[Mayahi]

We're in the process of switching 16 workstations from Octane 3.08 (offline mode) to V4.01.1-R2 (dongle). The plan was to keep Cinema R19 with Octane 3.08 and setup a clean Cinema R20 with Octane 4.01. Unfortunately, the licensing requires us to retire offline mode for V3 and switch our licenses to V4 with dongles. Not and ideal solution in a production environment, where best practice is to keep legacy software for old projects in tact.
We're noticing roughly a 10% brightness increase when migrating a project built in V3.08 to V4.01. I've confirmed that Camera Imager and Render Settings are identical between versions. It's a problem because we have about a dozen projects in the middle of production. There seems to be no way for us to upgrade our workstations to V4 while keeping V3 active.
Anyone else come across this issue? Thanks!

[aoktar]

Do you have octane sky object or directlight in this scenes?

[Mayahi]

Yes. OctaneSky with an HDRI for Primary Environment and RgbSpectrum (Black) for Visible Environment. Two spheres with Blackbody Emission and 5 Octane Arealights.

[Mayahi]

We have isolated the issue to Image Textures within materials having a very slight luminance shift in V4.01. Although it is a very slightly brighter, this ends up having major repercussions on the material response as the color is mapped to Gradients, Mixes etc.... in the Specular and Roughness channels.
There is a new dropdown menu in V4.01 "Texture import type" which is defaulted to Auto. Could this be interpreting the color slightly different than in V3.08?

[Mayahi]

We've solved the brightness shift after going thru the shader network node by node. The difference was the result of the Gradient assigned to "Texture2" of the Multiply node not having anything assigned to the Texture channel. By linking a FloatTexture (value 1) we were able to get the V4.01 render to match V3.08. It's strange that the unassigned Texture channel in the Gradient would produce a different result in V3.08 vs V4.01.

[Tim_Twisted]

Good to know that these differences occur. Since your Float was 1, it seems that the empty standard value was even higher which seems really weird since you expect it to be 1.

[aoktar]

I'll take a look for this in free time, thanks!