Benjamin Marks
Veteran
Dave: I'll give it another try: An ideal lens at a particular aperture would produce an image where one plane (think of it as a slice of reality) is in the sharpest focus. The "thickness" of this slice varies with both the aperture and the distance from your subject. For instance, when you are focused at "infinity" on your lens, everything past that point will be in focus (a very thick slice of reality). At the minimum focus distance, the slice can quite thin (for instance with a long lens, wide open it could be mere millimeters). You with me so far?
You probably know too that as you close a lens down, the depth of this zone of focus increases (so, for a given lens, more DOF at f:16 than at f:1.4). But with certain lenses the zone of focus, this "slice" of the world that is forming an image on your film or on the sensor, in addition to increasing in depth as you stop down, can also move a small amount forwards or back as you stop down your lens AND without your having changed the focus of the lens. This is focus shift.
For all camera systems in which you are not actually viewing the image-forming light at the film/sensor plane and at the taking aperture (for example, cameras with focusing systems other than a mirrorless SLR, a view camera, or an SLR focused using a depth-of-view preview lever) the camera's focusing system exists to convey accurately to you where you have chosen to put the point of best focus. In an otherwise well-calibrated system, when the indicated zone of best focus does not include the point that you have chosen to be in focus using the cameras indicators (SLR focusing screen, RF camera's focusing patch), this is focus shift. It does not manifest in a quadrant of an image, it manifests in the whole slice of reality that the lens captures. In the museum picture above, it is noticeable in the corner because there is something there. What the poster of that picture has told us is that he focused on the man in the corner. If we assume his system is properly calibrated and then see this result when the photographer focused on that subject, focus-shift is a good candidate for explaining why the point of best focus appears to be behind the man. Of course, it could also be that the photographer moved when he composed the picture, or leaned forward unconsciously trying to get the framing right . . . but let's assume for the sake of explanation that the photographer had good technique when making the exposure and that he didn't understand the behavior of this lens used at this aperture and distance. Under these conditions, if the photographer had tried to take a picture of something at this aperture with this lens that existed in a plane (say, the canvas on the wall) the actual plane of best focus might be located behind that indicated by his camera's focusing system, resulting in an image that would not appear critically sharp.
So: the manifestation is this: photographer with good technique focuses on the eye of the man in the chair. In the resulting image, the man's eye appears out of focus and the frame of the painting 6 inches behind his eye appears in focus. Unless compensated for (in this case by leaning back in response to the camera's information that the eye was, in fact, in focus), this camera, this lens, used at this aperture, for a subject at this distance will ALWAYS produce an image that is less than critically sharp.
Ben
[Edit: how you deal with focus shift depends on what kind of camera you are using. With a mirrorless EVF, you would stop down the lens first and then focus. Now the point of best focus has been chosen by the photographer with the focus shift "baked in" (because it is the act of stopping down the lens that causes focus shift). On a view camera, you would do the same thing: you compose with the lens wide open for convenience, but before you take the picture you also view the image on the groundglass with the lens stopped down. With a good SLR and enough light you can use a depth-of-field preview lever (the hint is in the name) and check what is in focus and what isn't before taking a picture. With an RF camera, there is no way of doing any of this since you are not focusing with image-forming light, you are focusing based on a mechanical approximation (the camera's rangefinder and the lens's cam) that is usually good enough. With an RF camera, if your images are not sharp enough for you, you must learn which of your lenses exhibit this behavior and then, using trial and error, figure out how to compensate. In the picture above, you would compensate by leaning back a little.]
You probably know too that as you close a lens down, the depth of this zone of focus increases (so, for a given lens, more DOF at f:16 than at f:1.4). But with certain lenses the zone of focus, this "slice" of the world that is forming an image on your film or on the sensor, in addition to increasing in depth as you stop down, can also move a small amount forwards or back as you stop down your lens AND without your having changed the focus of the lens. This is focus shift.
For all camera systems in which you are not actually viewing the image-forming light at the film/sensor plane and at the taking aperture (for example, cameras with focusing systems other than a mirrorless SLR, a view camera, or an SLR focused using a depth-of-view preview lever) the camera's focusing system exists to convey accurately to you where you have chosen to put the point of best focus. In an otherwise well-calibrated system, when the indicated zone of best focus does not include the point that you have chosen to be in focus using the cameras indicators (SLR focusing screen, RF camera's focusing patch), this is focus shift. It does not manifest in a quadrant of an image, it manifests in the whole slice of reality that the lens captures. In the museum picture above, it is noticeable in the corner because there is something there. What the poster of that picture has told us is that he focused on the man in the corner. If we assume his system is properly calibrated and then see this result when the photographer focused on that subject, focus-shift is a good candidate for explaining why the point of best focus appears to be behind the man. Of course, it could also be that the photographer moved when he composed the picture, or leaned forward unconsciously trying to get the framing right . . . but let's assume for the sake of explanation that the photographer had good technique when making the exposure and that he didn't understand the behavior of this lens used at this aperture and distance. Under these conditions, if the photographer had tried to take a picture of something at this aperture with this lens that existed in a plane (say, the canvas on the wall) the actual plane of best focus might be located behind that indicated by his camera's focusing system, resulting in an image that would not appear critically sharp.
So: the manifestation is this: photographer with good technique focuses on the eye of the man in the chair. In the resulting image, the man's eye appears out of focus and the frame of the painting 6 inches behind his eye appears in focus. Unless compensated for (in this case by leaning back in response to the camera's information that the eye was, in fact, in focus), this camera, this lens, used at this aperture, for a subject at this distance will ALWAYS produce an image that is less than critically sharp.
Ben
[Edit: how you deal with focus shift depends on what kind of camera you are using. With a mirrorless EVF, you would stop down the lens first and then focus. Now the point of best focus has been chosen by the photographer with the focus shift "baked in" (because it is the act of stopping down the lens that causes focus shift). On a view camera, you would do the same thing: you compose with the lens wide open for convenience, but before you take the picture you also view the image on the groundglass with the lens stopped down. With a good SLR and enough light you can use a depth-of-field preview lever (the hint is in the name) and check what is in focus and what isn't before taking a picture. With an RF camera, there is no way of doing any of this since you are not focusing with image-forming light, you are focusing based on a mechanical approximation (the camera's rangefinder and the lens's cam) that is usually good enough. With an RF camera, if your images are not sharp enough for you, you must learn which of your lenses exhibit this behavior and then, using trial and error, figure out how to compensate. In the picture above, you would compensate by leaning back a little.]
Last edited: