Menu
▾
▴
#39 Cannot find lens via lf_db_find_lenses_hd with exact name match
The lens to find is 'Nikon AF-S Zoom-Nikkor 17-35mm f/2.8D IF-ED'. I use the 'Nikon D3S' as a camera for searching. The search function I use is lf_db_find_lenses_hd. When I don't use fuzzy search, it doesn't return any matches. If I use fuzzy search I get the 'Tamron SP AF 17-35mm f/2.8-4 Di LD Aspherical (IF)' lens with score 65.
When I manually iterate over all lenses via lf_db_get_lenses() and check for a perfect match of the lens name, I can find the lens, meaning it is really there. I guess there's a bug with the search logic somewhere.
Note that this problem is not new. I discovered that last year but never got around to file a bug report... well, better late than never!
The problem is the crop factor. The lens has a too small image circle and cannot be used with a FF camera.
Well, the actual lens is suitable for FF cameras, right? It's just that the calibration data was not created with a FF camera. In that case, yes, it's not actually a bug.
By the way I find it really confusing that <cropfactor> is a direct child of the <lens> element, instead of the <calibration> element. After all, there are two crop factors, the physical minimum for that lens, and the one used for calibration. Or am I misunderstanding something?
You are right, this is all not optimal. It's on the todo list, but I don't think it will be implemented soon. For example, http://permalink.gmane.org/gmane.comp.graphics.lensfun/300 discusses the possibility to detect that a Nikon FX camera works in DX mode. Another related issue is https://sourceforge.net/p/lensfun/features/19/. See also https://sourceforge.net/p/lensfun/wiki/Lenses%20vs.%20calibrations/. But this means a lot of internal change, and doing this without breaking anything will be hard.
Can I close this issue for now?
I see, ok. Good to know that it's already tracked elsewhere. Please close it.