Noise profile from assembled selections?

Leo · Post by **Leo** » Tue Jan 18, 2005 6:10 am

The Problem: Many scattered areas of featureless sky, none large enough to allow a rough profile, but several ok to use for manual fine-tuning.

The Madcap Idea: Copy and paste together several such areas to assemble a 100x100 sample image (with an absolute bare-minimum of cloning), then load it into Neat Image and obtain a rough profile, and save that. Then open the source photo, load the saved rough profile, and manual fine-tune using the remaining featureless areas.

Would this be a useful approach to getting an accurate noise profile?

Post by **NITeam** » Tue Jan 18, 2005 12:58 pm

That is quite similar to option 6 in this page. It is difficult to make an accurate assemble so the accuracy of noise profile will suffer too.

Vlad

Leo · Post by **Leo** » Wed Jan 19, 2005 6:20 am

Thanks Vlad. Well it's worth a try. If I get surprising results I'll post them.

taob · Post by **taob** » Fri Jan 21, 2005 4:45 am

Leo,

Why not shoot a calibration target (following the directions provided in the documentation) at the same settings as your sky shot, and use that to develop a profile?

Leo · Post by **Leo** » Fri Jan 21, 2005 7:27 am

taob wrote:Why not shoot a calibration target (following the directions provided in the documentation) at the same settings as your sky shot, and use that to develop a profile?

Taob, for that shot I used my old Kodak DC240 which captures amazing color, but the ISO is automatic and can't be pre-set. I did find another shot in the series though, which provided a useable profile & good results. I am still going to try the patchwork method however. I'm currently experimenting with ways to piece together selections without using the clone brush or changing any pixel values. It is truly maddening.