Am 20.02.2015 um 16:32 schrieb scott:

> You might get better results by rendering with a white light - once
> without attentuation and another render with attenuation. The reason
> being that HDR formats (I assume you are outputing to an HDR image
> format for better precision) IIRC only store a "magnitude" for each
> pixel, so if red and green are vastly different trying to calculate the
> ratio will have a large error.

You're right as far as the Radiance HDR file format is concerned; not so 
OpenEXR though: It uses three separate 16-bit ("half precision") 
floating point values per pixel.

Post a reply to this message