piersyf opened this issue on Sep 16, 2014 · 57 posts
piersyf posted Tue, 10 February 2015 at 12:46 AM
Thanks to a recent post by Miss Nancy I decided to investigate the Tone Mapping options in Firefly as an alternative to cranking up light values and ambient values to simulate different 'exposure settings'. Starting with the same premise of 100ASA film shooting outdoors, a base setting of ambient sky set to 1 and sun infinite light set to 100% equates to an exposure time on a standard SLR camera of 1/60th of a second at around f.4 or 5.6.
So, Tone mapping for the default light setting (100% amb 1), off.
Tone mapping for 1/30th of a second, base set to 2, exponential set to 1.Equates to light at 200%, ambient 2 on the skydome.
for 1/15th of a second, base 2, exp. 2 (400% sunlight, ambient 4 on skydome)
for 1/8th of a second, base 2, exp. 3
for 1/4 of a second, base 2, exp. 4, and so on.Base 2 meant I only needed to change the exponential value to get a progressive doubling of light values.
To replicate the light levels for the very first image in this thread, I reset ambient to 1 and light to 100, set tone mapping base to 2 and exponential to 7.
Results? Image is somewhat washed out (regardless of whether I use exponential or HSV exp, although HSV is a bit better). Tones are evened across the board.
Also, and this was a surprise, the artifacts caused by light leaking through geometry seams appear at exactly the same level they do when ramping up light settings. So regardless that the light was at 100%, Firefly ramps it up itself in the calculations.
So... who cares? Well, I though it interesting to get a similar result by two methods. It also means that exposure settings ARE in Poser (at least in Firefly), they just aren't explained well and aren't very good. If the algorithm can be fixed so that Firefly recognises intersecting geometry as a 'no go' for light rays and they can tweak the colour averaging, they're almost there...
If you're wondering why no pictures, I spent 6 hours doing many renders with this testing settings and doing comparisons, then considered it a failed experiment. I didn't realise it might be useful to someone else until 2 days later... like today...