On samedi 20 avril 2019 21:41:19 CEST Florian Pressler wrote: (...) > > Further investigation shows that darktable reports the ARW file with > dimensions 6048/4024, while it "should" have 6000/4000. Are the extra > pixels in the ARW-file kind of a buffer for in-body-stabilization? Where do > they come from, and why do they disappear in the converted DNG-file?
The extra pixels are simply present on the sensor, and stored in the raw file. As for why they are there: In-body stabilization would indeed need a few pixels on each side. Also, demosaicing has to treat the edge pixels differently from the others, so if you want to use all pixels, your algorithm gets more complicated. Producing a slightly smaller image lets you treat all the relevant pixels the same way. And, as Michael mentioned, the extra pixels give some room for in-camera lens correction. Note that this all only refers to in-camera treatment, I doubt Sony would take into account the wishes/needs of external (darktable) devs, perhaps unless they work for Adobe. And keep in mind that a given sensor can be used in a lot of different camera models, of different brands. Remco ____________________________________________________________________________ darktable user mailing list to unsubscribe send a mail to [email protected]
