jaapv
RFF Sponsoring Member.
DOF: The reality of an illusion.
DOF is a subject that causes heated discussion in photographic circles. It is, of course, next to light and shape, one of the main photographic symbols to express ourselves.
There is a simple mathematical approach that is expressed in DOF scales on lenses and DOF tables in manuals, but, as always, that is not the whole story – by a fair margin.
DOF as a phenomenon is childishly simple. The human eye is a rather imperfect instrument for judging sharpness, so with a resolution of about 5 lp at 75 cm everything that is higher resolved appears sharp. So now the compications start. It readily confuses contrast with sharpness, the only reason that sharpening algorithms in postprocessing actually work.. So a photograph at noon at the beach will appear to have a deeper DOF than one on a misty morning. Of course, a photograph is, in reality sharp only in one plane, which is theoretically infinitely thin, but at least as thin as the state of correction of the lens and the quality of the receiving medium, be it film or sensor, allows. Lens manufacturers, in their quest for simplification and standardization have decided, in the 1920-ies, that an unsharpness of 0.03 mm on 35 mm film would be judged the measure of DOF. That leads us to the first set of complications.:
1. Without knowing the end enlargement of the photo one takes and without taking the contrast into consideration, judging the amount of DOF is actually rather hit and –mostly- miss.
2. As DOF is solely dependent on field of view, the “enlargement” of the focal length of the lens, which is responsible for the apparent deep DOF of wideangle-lenses and shallow DOF of long lenses gets into play, so the subsequent crop will influence the DOF in as much that if one crops a 28 mm shot down to the FOV of a 90 mm lens, the DOF will be exactly the same as that 90 mm lens would have produced.
3. Film is not without thickness. In reality a COC of 0.03 mm will act like a torch shining into a murky plate of soup. It will produce a cone, diffractions, reflections, if the light strikes the film at an angle it will turn into an oblong, etc., the net result being a larger diffuse spot. This is complicated by the fact that the films we have now are much thinner and higher resolving than we had in the 1920íes.
4. Digital sensors react far more like the ideal thin receiving medium than film, causing the COC’s to be even less diffused.
5. The net result is that the DOF produced now, and especially with modern lenses (of which I will write later) is more pronounced than it is historically. It is safe to assume that it is about 70% of the scale indicated on your lens. Btw. let’s not forget that it is not divided equally in foreground and background. The real division is, for simple mathematical reasons, 1/3-2/3, more or less, depending on subject distance.
All this caused me to call DOF in another context and another forum a RBU <rubber band unit>, which got me heavily flamed.
Then we get to the real controversial point, and that is the effect of individual lenses on DOF,
which relates to the elusive “boke”, which aptly translates to "chaos" or "confusion" I'm told, and to the rendering of out of focus picture elements.
In general the lens is corrected optimally for the plane of sharpness only, which means that aberrations like chromatic aberration and astigmatism increase quickly as sharpness decreases. Add this to my plate of soup effect and the magnitude of possibilities gets so large that only using the lens in practice will give any firm grasp of its (lack of) qualities.
The result is that, in extreme cases of not too well corrected lenses, there will be double contours, rings and general unpleasantness in the unsharp areas. That gives bad Boke. More elegantly, but still not optimally corrected lenses, and this applies to a large number of the older lenses used by RFF-ers, will produce generally soft and smoothly changing unsharp areas where the forms as such are undistorted. (did I mention geometrical distortion with the aberrations? This is the three-dimensional variant😉) That are lenses with a good boke. Then there are the newest, highly corrected lenses, like the Leica ASPH’s, APO’s etc. Those define the unsharp areas so well that they will break up the contours, giving rise to harsh boke.
Film will behave differently than sensors, as explained above. So a sweet lens on film may be unpredictably disappointing for digital and the other way around.
I have only scratched the surface of this subject here. Please add to this and contradict me as you please. If the thread does as well as the sensor size one, I’ll be happy 🙂
DOF is a subject that causes heated discussion in photographic circles. It is, of course, next to light and shape, one of the main photographic symbols to express ourselves.
There is a simple mathematical approach that is expressed in DOF scales on lenses and DOF tables in manuals, but, as always, that is not the whole story – by a fair margin.
DOF as a phenomenon is childishly simple. The human eye is a rather imperfect instrument for judging sharpness, so with a resolution of about 5 lp at 75 cm everything that is higher resolved appears sharp. So now the compications start. It readily confuses contrast with sharpness, the only reason that sharpening algorithms in postprocessing actually work.. So a photograph at noon at the beach will appear to have a deeper DOF than one on a misty morning. Of course, a photograph is, in reality sharp only in one plane, which is theoretically infinitely thin, but at least as thin as the state of correction of the lens and the quality of the receiving medium, be it film or sensor, allows. Lens manufacturers, in their quest for simplification and standardization have decided, in the 1920-ies, that an unsharpness of 0.03 mm on 35 mm film would be judged the measure of DOF. That leads us to the first set of complications.:
1. Without knowing the end enlargement of the photo one takes and without taking the contrast into consideration, judging the amount of DOF is actually rather hit and –mostly- miss.
2. As DOF is solely dependent on field of view, the “enlargement” of the focal length of the lens, which is responsible for the apparent deep DOF of wideangle-lenses and shallow DOF of long lenses gets into play, so the subsequent crop will influence the DOF in as much that if one crops a 28 mm shot down to the FOV of a 90 mm lens, the DOF will be exactly the same as that 90 mm lens would have produced.
3. Film is not without thickness. In reality a COC of 0.03 mm will act like a torch shining into a murky plate of soup. It will produce a cone, diffractions, reflections, if the light strikes the film at an angle it will turn into an oblong, etc., the net result being a larger diffuse spot. This is complicated by the fact that the films we have now are much thinner and higher resolving than we had in the 1920íes.
4. Digital sensors react far more like the ideal thin receiving medium than film, causing the COC’s to be even less diffused.
5. The net result is that the DOF produced now, and especially with modern lenses (of which I will write later) is more pronounced than it is historically. It is safe to assume that it is about 70% of the scale indicated on your lens. Btw. let’s not forget that it is not divided equally in foreground and background. The real division is, for simple mathematical reasons, 1/3-2/3, more or less, depending on subject distance.
All this caused me to call DOF in another context and another forum a RBU <rubber band unit>, which got me heavily flamed.
Then we get to the real controversial point, and that is the effect of individual lenses on DOF,
which relates to the elusive “boke”, which aptly translates to "chaos" or "confusion" I'm told, and to the rendering of out of focus picture elements.
In general the lens is corrected optimally for the plane of sharpness only, which means that aberrations like chromatic aberration and astigmatism increase quickly as sharpness decreases. Add this to my plate of soup effect and the magnitude of possibilities gets so large that only using the lens in practice will give any firm grasp of its (lack of) qualities.
The result is that, in extreme cases of not too well corrected lenses, there will be double contours, rings and general unpleasantness in the unsharp areas. That gives bad Boke. More elegantly, but still not optimally corrected lenses, and this applies to a large number of the older lenses used by RFF-ers, will produce generally soft and smoothly changing unsharp areas where the forms as such are undistorted. (did I mention geometrical distortion with the aberrations? This is the three-dimensional variant😉) That are lenses with a good boke. Then there are the newest, highly corrected lenses, like the Leica ASPH’s, APO’s etc. Those define the unsharp areas so well that they will break up the contours, giving rise to harsh boke.
Film will behave differently than sensors, as explained above. So a sweet lens on film may be unpredictably disappointing for digital and the other way around.
I have only scratched the surface of this subject here. Please add to this and contradict me as you please. If the thread does as well as the sensor size one, I’ll be happy 🙂
Last edited:
