21 February 2010
new directions
21/02/10 20:01
As a start some figures: The Leica Pocket Guide has been downloaded 12.000 times in two weeks, the leica lens compendium pdf version is now at a count of 55.000, with 5000 downloads in the last week.
An analysis of the image quality of the Zeiss Planat 1.4/85 on the Canon 5D II showed a strange behavior. The explanation proved to be very complex, involving the working and calibration of the AF mechanism, the effect of distance dependent focus shift, the possible influence of the EPROM in the lens in combination with the processing of the camera. On top of the classical optical qualities of a lens, we now have a plethora of measurements, generated by the several lens testing programs available over the internet and the ever increasing integration of lens parameters with the camera operation (AF as example), but also programs in the camera that adjust for known optical defects. Two competing trends we see: the analysis programs deliver an enormous amount of information, based on their own interpretation of optical performance of the lens and the camera/lens manufacturers use complex software algorithms (in lens and camera) to adjust and adapt the performance of the optical system and the perception of the resulting image. The upshot is a new level of insecurity: we do not know what we measure and we do not know how to explain what we assume we measure. This state of affairs is complicated by the tight coupling of software with the hardware and the opacity of the inner workings of the camera/lens/program interaction. All analysis programs only look at the end result of the imaging chain (the image file in Raw or JPG format), disregarding the working of the chain itself and there is no option for separating the constituent components of the chain. And when we need to interpret the end result we have lists of data and a stream of graphs to help us make sense of the data. So we have on the one hand many data delivered by quite sensitive equipment and on the other hand a lack of knowledge about how to interpret the data in a sensible way. The more we measure, the less we seem to know.
It is fair to assume that the final result is all we need to consider, but if we want to explain what we see and we claim to deliver intelligent assessments and comparisons of what we perceive, we need more than just wild guesses. This knowledge is simply lacking as no camera/lens manufacturer is telling you what goes on inside the black box.
This is the reason why I still test lenses in the classical way: with film and with MTF equipment. That is the only method to study the prime and pure characteristics of a lens and to make meaningful comparisons.
If you on the other hand want and can only test the full imaging chain in the digital workflow, you are at a loss as far as the lens qualities are concerned. The usual approach nowadays is to use a standard test chart that can be quantitatively evaluated by an analysis program and to produce some charts and figures that show differences in values. How meaningful these differences are is not discussed, nor is the fact how representative the results are. The performance of a lens is distance dependent, but most test situations assume only one distance form camera to target and in many cases this distance is not the most representative for the lens in question.
More data produce more insecurity and more information noise. A nice modern dilemma it surely is.
Long ago I wrote about the death of classical photography. That is the photography where the photographer uses film to capture a scene or moment that is recorded in a constellation of silver salts and carefully printed in the darkroom with the result a crafted unique copy of an image for use in a book or magazine or for art exhibition purposes. Modern digital photography is rapidly becoming a totally different medium. The single photograph is no longer the goal, but photography is converging to a mixed media show, where sound, still images, and moving images (video) are becoming combined to convey the message. As a sign of the times we may refer to the recent sale of a big portion of the photographs by Magnum to collect money for expanded photoreportage work by the members.Go to Magnum in Motion to see the modern multimedial digital show of Magnum members. A modern photographer does no longer shoot pictures, you need to develop yourself to an all-round artist, commanding visual and audio equipment or you become a journalist. Modern cameras support the multimedial approach: see the new Canon, NIkon, Panasonic and Olympus cameras. If magnum gets the message, why not the the rest of us?
Lens testing
An analysis of the image quality of the Zeiss Planat 1.4/85 on the Canon 5D II showed a strange behavior. The explanation proved to be very complex, involving the working and calibration of the AF mechanism, the effect of distance dependent focus shift, the possible influence of the EPROM in the lens in combination with the processing of the camera. On top of the classical optical qualities of a lens, we now have a plethora of measurements, generated by the several lens testing programs available over the internet and the ever increasing integration of lens parameters with the camera operation (AF as example), but also programs in the camera that adjust for known optical defects. Two competing trends we see: the analysis programs deliver an enormous amount of information, based on their own interpretation of optical performance of the lens and the camera/lens manufacturers use complex software algorithms (in lens and camera) to adjust and adapt the performance of the optical system and the perception of the resulting image. The upshot is a new level of insecurity: we do not know what we measure and we do not know how to explain what we assume we measure. This state of affairs is complicated by the tight coupling of software with the hardware and the opacity of the inner workings of the camera/lens/program interaction. All analysis programs only look at the end result of the imaging chain (the image file in Raw or JPG format), disregarding the working of the chain itself and there is no option for separating the constituent components of the chain. And when we need to interpret the end result we have lists of data and a stream of graphs to help us make sense of the data. So we have on the one hand many data delivered by quite sensitive equipment and on the other hand a lack of knowledge about how to interpret the data in a sensible way. The more we measure, the less we seem to know.
It is fair to assume that the final result is all we need to consider, but if we want to explain what we see and we claim to deliver intelligent assessments and comparisons of what we perceive, we need more than just wild guesses. This knowledge is simply lacking as no camera/lens manufacturer is telling you what goes on inside the black box.
This is the reason why I still test lenses in the classical way: with film and with MTF equipment. That is the only method to study the prime and pure characteristics of a lens and to make meaningful comparisons.
If you on the other hand want and can only test the full imaging chain in the digital workflow, you are at a loss as far as the lens qualities are concerned. The usual approach nowadays is to use a standard test chart that can be quantitatively evaluated by an analysis program and to produce some charts and figures that show differences in values. How meaningful these differences are is not discussed, nor is the fact how representative the results are. The performance of a lens is distance dependent, but most test situations assume only one distance form camera to target and in many cases this distance is not the most representative for the lens in question.
More data produce more insecurity and more information noise. A nice modern dilemma it surely is.
Modern photography.
Long ago I wrote about the death of classical photography. That is the photography where the photographer uses film to capture a scene or moment that is recorded in a constellation of silver salts and carefully printed in the darkroom with the result a crafted unique copy of an image for use in a book or magazine or for art exhibition purposes. Modern digital photography is rapidly becoming a totally different medium. The single photograph is no longer the goal, but photography is converging to a mixed media show, where sound, still images, and moving images (video) are becoming combined to convey the message. As a sign of the times we may refer to the recent sale of a big portion of the photographs by Magnum to collect money for expanded photoreportage work by the members.Go to Magnum in Motion to see the modern multimedial digital show of Magnum members. A modern photographer does no longer shoot pictures, you need to develop yourself to an all-round artist, commanding visual and audio equipment or you become a journalist. Modern cameras support the multimedial approach: see the new Canon, NIkon, Panasonic and Olympus cameras. If magnum gets the message, why not the the rest of us?