David: On Thu, October 15, 2009 01:38, karl shah-jenner wrote: > Has no one else has encountered the claim that digital lenses output > parallel rays of light to the sensor? >Is it possible that you've latched onto "telecentric" when they're actually using different techniques, with different side-effects? There are usually several ways to accomplish a given goal. it's the description often used - that of light playing on the sensor running parallel to the optical axis I've latched onto, something that to me, never made sense knowing what I do of lenses. Then the term 'telecentric' started popping up (or "near-telecentric") in descriptions and manufacturers explanations. So I dug and found an explanation of what telecentric lenses actually are, and while they do as described - they do not in any way appear to be used for standard photography. These are the only lenses I have ever heard of which produce image forming light parallel to the optical axis. There may be others I don't know, but telecentric lenses are a whole world of different for me and I cannot imagine (nor have I found) anything to suggest another lens configuration that can pull off this magic What I suspect is the manufacturers were having the marketing guys write up whatever they liked to convince buyers that they needed new lenses - aside from the obvious problems of using lenses designed for 35mm on what is often essentially a competitor to 110 - and this is no different from the issues one would experience using a Pentax MF lens on a 35mm camera. Of course the cameras are different and an adapter is needed - The tricky bit is they've manufactured the cameras to integrate to a small extent with the 35mm systems available. I'm again guessing, but this was to make it easy for people to make the switch .. marketing telling us that the transition is almost seamless - "hey people, they even LOOK the same!". The fact being it's really not. "Not only have you bought a whole new camera body, you now need to buy new lenses (!) You can still use some of the old ones, but there will be problems" But how to make people buy the new lens range? Spin some story about how the light behaves differently.. >I won't for a moment disagree that the designation "digital" on lenses has been thrown around somewhat randomly and hasn't always meant anything, and that it has no very specific definition. However, I do think that as we've learned more about digital photography, we've found some issues with how image-forming light coming out a lens interacts with film vs. digital sensors, and that some changes to lenses and some changes to sensors (or the microlens arrays over the sensors) have been made to make things work better. These issues include how the sensor reacts to light at shallow angles (considerably differently from film), and the fact that the sensor is shiny and thus can make sharp reflections, in addition to the fact that many DSLR cameras use smaller-size sensors than 24x36mm Shiny is another one that gets me. Some film surfaces have been *very* shiny, and atop that, some very pale - reflecting a *lot* of light around inside cameras. If you recall many years back me describing here how I built an infrared-lit film changing cabinet with a CCD camera inside? it made loading troublesome cameras, spools and sheet film holders very easy and was extra handy if I wasn't sure if a double dark was loaded or not ;) As I was doing a lot of IR photography I popped a bunch of cameras and lenses under the IR to have a peek and see how reflective the inner workings of the cameras were.. which is when I discovered deep black anodised aluminium reflects IR light like chromework - at the end of the day I found old British lenses with black painted interiors reflected the least and mirrored what I had been experiencing when using various lenses - a much sharper, contrasty image than with any other lenses. And another surprise was the cameras - the Canon T70 ( a dog by anyone's definition) was by far and away the darkest camera internally for IR light. It was also the *only* camera with a dark film pressure plate. Since Kodak weren't using an antihalation layer in their IR film, the T70 was the best camera to use for IR if one didn't want the dreamy, flared results we were used to seeing in pictorial photography. - Adding black paper from 120 roll film to cover the pressure plate of other cameras helped to a degree, but the internal reflections were still there.. My point - reflections occurred and most people didn't notice a jot. The Olympus cameras equipped with off the film metering were a lovely design, and for 90% of users performed faultlessly - but there were still that 10% who found a niggling problem - and for those using photography for critical applications, they came to realise the levels of light reflected from differing films caused differing exposures. Again, few people noticed anything. Another gadget I built was intended for testing film speeds - a panel of ND squares I could photograph ensuring a full range of exposures in 1 stop intervals across a single frame of film :) Except it didn't work. The reason as I reported here was that different lenses produced quite different contrasts. What I didn't report was that when using the same lens and switching films, different films had different contrasts. This was not just a development issue as one would expect, as I had controlled for that by establishing reference development times from direct exposed and developed film. This again suggested refection of the film surface played some part. mhey - I gave all this experimenting up and worked with what I had ( I decided while it was a nice mental exercise, *I* didn't really care either ;) Digital sensors - there is some interesting mechanics at play there! microlenses. Firstly we know it's not a straight sensor by sensor conversion to image. We don't *need* super highly resolved light falling on each individual sensor, in fact it's undesirable as it causes all manner of problems like moiré - subsequently an antialiasing filter is added to 'blur' the light. We use 3 or more sensors to determine the colour of the light then we interpolate that back into three separate pixel colours and locations. Microlenses are needed though to gather the light hitting that part of the sensor and focus it tightly onto the phototransistor it's self. Now we all know a bit about lenses, and we all know how a dome shaped hemispherical lens sitting above something will work - pretty much whatever angle the light strikes the hemisphere will be focussed pretty much dead centre - where the phototransistor sits ..just as it's designed to do to prevent the light falloff from reflection that would occur if the dome were not there and the surface were in fact flat. ..(if you didn't know this, think how a wide angle works) So we have microlenses and a fuzzing filter, we have algorithms playing with the data collected from the photosites - we even know now that fringing is correctable by firmware (the Sony F828 was widely criticized as having poor optics due to purple fringing - the fact is such fringing was a reality even for film but few people ever enlarged film images to the extent they do when simply viewing digital pictures and subsequently never saw it) When we have all this, what need is there for anything faintly resembling a requirement for light rays to be collimated, even partially? Maybe an argument could be presented for such a thing but then where are the real world examples? I'd imagine that in its infancy when DSLRs were first hitting the market and sensors were still being refined we'd have seen an uproar from people buying into these systems who used their old film camera lenses only to find vignetted dark corners, blurry images and colour fringing. It didn't happen. So what the heck is all this in the manufacturers blurb about telecentric or near telecentric lenses, parallel light, etc? Again, I'd love to hear from anyone who owns a 'digital' lens made by a manufacturer who claims any such thing who can look through the lens and see if the aperture look *really* far away (like infinity?) I suspect it's all bunk. Always did, I just never had a firm grip on what it was they were claiming before.. karl
