Page 1 of 1
Light Falloff Mystery
Posted: August 27th, 2009, 6:14 pm
by randys
I've tried building light falloff curves for my lenses by taking pictures in sun of a white foamboard, converting the raws in pw and then pushing the compute button in the LIght Falloff transform. Each time I have gotten a 'Failed to Find a good Fit' message. Went thru this about a dozen times. Any hints for to take the images?
Thanks
Randy
Re: Light Falloff Mystery
Posted: August 27th, 2009, 8:01 pm
by Staylor2
Are any of your channels saturating? You could try changing the exposure. I had a problem like this and I converted the image to grayscale and the computation went a lot better. There was something wrong in the way the program handled RGB. That was during the Beta test period however, and I think this problem was fixed before the final release. It still may be worth a try.
Re: Light Falloff Mystery
Posted: August 28th, 2009, 6:58 am
by jsachs
This error message indicates that the light falloff curve is very uneven and cannot be accurately modeled. Make sure the white card is evenly lit and that you expose it so none of the channels is clipped and nothing is obscuring part of the field of view (e.g. a filter cutting off the corners of the frame).
Re: Light Falloff Mystery
Posted: August 29th, 2009, 1:29 pm
by randys
Success! I switched to 1ev underexposed jpgs and was able to compute curves with no problem. Could not get images with even enough lighting using PW's raw converter. Perhaps my settings influenced the result. Getting an evenly lit image from a full frame fisheye was a little tough even on my 40d. The presence of the camera a few inches from the white board influenced lighting in ways you couldn't see.
Thanks for your help
Randy