08-06-2011, 04:54 PM
[quote name='Brightcolours' timestamp='1312648309' post='10433']
You stretch the image while "correcting", and clip the excess off (ending up with similar pixel resolution as the sensor generates. So you lose lens resolution mostly on the edges and corners, but only once... not twice.
[/quote]
I agree. Not twice. I just wanted to emphasize that also the center is affected. If you just cut the corners, the image size will be reduced from 4000*3000 to maybe 3850*2900 (4:3 12MP example). If you now save the image, you will not have a 12 MP image. But the cameras and also Adobe Lightroom do not do that. They save a full 12 MP (4000*3000) image. Therefore, it is necessary to extrapolate the entire, cropped image. I am wondering about the impact on IQ (depending on the extrapolation algorithm). In any case, I do not like the idea.
You stretch the image while "correcting", and clip the excess off (ending up with similar pixel resolution as the sensor generates. So you lose lens resolution mostly on the edges and corners, but only once... not twice.
[/quote]
I agree. Not twice. I just wanted to emphasize that also the center is affected. If you just cut the corners, the image size will be reduced from 4000*3000 to maybe 3850*2900 (4:3 12MP example). If you now save the image, you will not have a 12 MP image. But the cameras and also Adobe Lightroom do not do that. They save a full 12 MP (4000*3000) image. Therefore, it is necessary to extrapolate the entire, cropped image. I am wondering about the impact on IQ (depending on the extrapolation algorithm). In any case, I do not like the idea.