Earlier, I posted some examples of images made in two ways:
- Using the shift function of the GF 30/5.6 T/S lens with the sensor parallel to the target.
- Using no shifting, yawing the camera to put the target in the same place in the image, and using Photoshop in postproduction to make it seem as if the target had been photographs from straight on.
The relative results depended on f-stop. Sharpness at f/8 was about the same. Sharpness at f/5.6 was better in the non-shifted, Ps-manipulated image.
In this post, I’d like to discuss the advantages and disadvantages of each approach more generally.
The most obvious advantages of using shift is more accurate framing. What you see is what you get. It’s hard to predict what the framing will look like if you just pitch or yaw the camera and plan on correcting in post.
The most obvious advantage of using pitch or yaw is that you are using the center of the lens’ image circle more, and therefore there will be less sharpness lost than if you used more of the periphery of the lens’ image circle as would have been the case had you shifted.
To a first approximation, with a rectilinear lens and a rectilinear final image, there will be the same amount of stretching of the subject. If the shift or pitch is upwards, that stretching will be most apparent at the top of the frame. It follows that, if sharpness is limited by the lens itself rather than the sampling performed by the sensor, that the paragraph above this one implies that the result might be sharper if you tilt or yaw than if you shift. To a first approximation, sharpness will be about the same in both cases if the lens sharpness is diffraction limited, as would be the case with the GF 30/5.6 T/S lens stopped down to f/11 or narrower. As we’ve seen above, for that lens, sharpness is about the same at f/8.
However, at wide apertures, it is more likely that the periphery of the lens image circle is softer than more central parts of the projected image, and thus avoiding that part of the image circle can produce a sharper result, as we saw with the GF 30 T/S lens at f/5.6.
If part of the image sharpness is limited by the sampling frequency of the sensor, then shifting has an advantage in that part of the image, since pitch or yaw would require resampling that part of the image at a higher spatial frequency than was used for the capture.
Leave a Reply