Le 08/06/2014 22:38, Leonard Evens a écrit : > I should have explained more carefully what I meant. I also have a > D800, an FX camera. The entry in slr-nikon.xml for all 70-300 mm lens > specifies the factor 1.5X, meaning that the calibration was done with a > DX camera. When you do the calibration, you must specify that factor. > If in fact you use the lens with an FX camera, for which the factor > would be 1.0X, you might get different results. > > I plan mainly to use the lens with the D7100, but there may be occasions > when I want to use it with my D800. In that case, to use darktable's > Lens Correction, I would have to specify that the camera was some DX > camera, in order to pull up the lens from the database, since there is > no entry for my lens with a 1.0X factor. I can, of course, do that, but > I suspect the corrections wouldn't be quite accurate. To get accurate > corrections, I might have to calibrate the lens on my D800 myself, > unless someone has already done that. > > I speak from experience since I've already been through this with my > 70-200f/4 lens. The entry in the database for that lens is for DX > format, and I had to calibrate it myself. > > I suspect these lens all have very little distortion, so the differences > are in fact very small.
Thank you for explaining. You're right. Rigorously, An FX lens should be calibrated using an FX camera. And it would be recognizable in the database from the 1.0 crop factor. If using a DX camera to calibrate the lens, obviously, all the area outside DX will not be available for calibration. The numbers wouldn't be exactly the same if you calibrated on an FX body because a real lens doesn't exactly match the model. Applying correction from that DX calibration to a FX photo would at best (and I think it would, because lensfun doesn't get confused and does its job) get a nominal correction inside the DX area and an extrapolated correction around it. So one might fear good correction inside the DX area and wrong (or even horribly wrong) outside (for FX photos). Actually, I would not be surprised if the whole thing would come out similar to what you would have with a FX calibration. The reason is the model as you can see in the lensfun database is simple : three real numbers (for a given focal length), so either (1) the real lens are mostly close to the model, they fit well (give good corrections), so the fit should be similar in FX and DX calibration, with a possibly measurable but not enormous advantage to the FX calibration (2) the real lens are far from the model, the calibrations would be different, but they would both be bad fit anyway (give not so good correction). So, all in all I wouldn't expect a big difference: if the lens is suitable for correction it should be okay anyway. Have you tried and looked at the result ? Ultimately, the photographer's eye is judge, isn't it ? Regards, -- Stéphane Gourichon ------------------------------------------------------------------------------ Learn Graph Databases - Download FREE O'Reilly Book "Graph Databases" is the definitive new guide to graph databases and their applications. Written by three acclaimed leaders in the field, this first edition is now available. Download your free book today! http://p.sf.net/sfu/NeoTech _______________________________________________ Darktable-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/darktable-users
