At the moment all darkened JPGs are darkened by the same amount.
Some HDRIs have weak or large light sources, so darkening them by the same consistent amount might put their brightest value below 1.0, meaning there'd be unnecessary banding.
For such cases, only darken them until the brightest pixel is at 1.0. A list of the amount each HDRI is darkened would have to be stored so that they can be brightened by the same amount later.
We could maybe do all HDRIs like this (even very bright ones), but then we'd have to do per-case gamma adjustments to avoid excessive banding in the shadows.
At the moment all darkened JPGs are darkened by the same amount.
Some HDRIs have weak or large light sources, so darkening them by the same consistent amount might put their brightest value below 1.0, meaning there'd be unnecessary banding.
For such cases, only darken them until the brightest pixel is at 1.0. A list of the amount each HDRI is darkened would have to be stored so that they can be brightened by the same amount later.
We could maybe do all HDRIs like this (even very bright ones), but then we'd have to do per-case gamma adjustments to avoid excessive banding in the shadows.