Archived posting to the Leica Users Group, 2003/12/14
[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]Hi Henning, Olympus have now updated their marketing to refer to their lenses as "almost telecentric" rather than their previous claim about making the rays parallel. Clearly if the rays are parallel at the sensor they are in focus at infinity behind the camera, not on the sensor. Apparently the digital lenses are designed with the exit pupil further away from the sensor whereas with film cameras this dimension is of secondary importance, such that many film camera lenses are unsuitable for current digital cameras (such as,apparently, several M lenses). Making the exit pupil further from the sensor reduces the maximum angle of the rays focused on the sensor, but I have never seen a tolerance calculated as to what is a maximum acceptable angle - which would in any case also depend on the design and position of the anti aliasing and colour filters. My reasoning regarding the size and position of the rearmost element is a bit hairy ar*ed but it goes like this: the lens is the cheapest they can make work small elements are cheaper than large ones the rearmost element is the smallest that is required to achieve the maximum speed at full aperture the entire diameter of the rear element is used the rays focused on the sensor have come from the whole element including the periphery One would need some ray tracing to work out where the maximum angles are but they will not be 90 degrees (one could argue that 1 infinitely small ray is at 90 deg for all parts of the image). I am far from an expert on optics and it is easy to believe both that the accuracy of the output of a digital sensor is likely to reduce as the angle of the rays impinging on it increases, perhaps an angle tolerance similar to depth of field will emerge for designers - who knows. I could well have got it wrong but I think the above is about right - I did consider sitting down with Erwin Puts book and a protractor to measure the approximate max angle at the film (sensor) surface of the various M and R lenses to see which were the worst but I haven't had time! cheers Frank On Monday, December 15, 2003, at 03:00 am, Henning Wulff wrote: > At 9:26 AM +0000 12/14/03, Frank Dernie wrote: >> So much for all the bullsh*t about angle of incidence at the sensor >> then eh! If the last element is almost touching the sensor the angle >> of incidence will be very large unless, I suppose, the periphery of >> the lens element is not used but then why would it be there? >> Frank >> >> >> On Sunday, December 14, 2003, at 08:16 am, Eric Welch wrote: >> >>> A MUCH shorter focal length. And have you seen the diagram of the >>> internals of the Digilux2? The last element is almost touching the >>> sensor. With longer focal lengths, and a shutter in the way, you >>> have your answer. >>> >>> On Dec 13, 2003, at 9:56 AM, eric wrote: >>> >>>> The body and lens dimensions look so very M rangefinder. >>>> So again - Why can't they design a digital M body? >>>> I know it has been discussed at length about angles of incidence >>>> and sensors >>>> not up to the task - but what is so different about the digilux 2 >>>> that >>>> precludes designing a body to accept M lenses? >>> Eric >>> Carlsbad, CA > > Well Frank, the fact that the rear lens element is almost touching the > sensor has almost nothing to do with whether the angle of incidence is > high or not. With a lens designed for digital use, the design usually > attempts to include a rear element or group that 'straightens out' the > incident rays; such elements are often placed relatively far back in > the lens. I would think that this would be the case. The lens design > would include a collector group at the front, a zoom group possibly > combined with a focussing group, and then, at the back, a rectifier > group that will reduce the angle of incidence. This the way most > digicam lenses are constructed. > > Lenses designed for film cameras, especially rangefinder cameras, > don't need to have the light rays descend on the film perpendicularly, > and therefore aren't designed that way because doing so creates other > problems. Have you noticed the rather severe distortion that most > digicam lenses have? This is one of the consequences. In a couple of > years this will probably get worked out, but for now it is a problem. > > -- > * Henning J. Wulff > /|\ Wulff Photography & Design > /###\ mailto:henningw@archiphoto.com > |[ ]| http://www.archiphoto.com > -- > To unsubscribe, see http://mejac.palo-alto.ca.us/leica-users/unsub.html > - -- To unsubscribe, see http://mejac.palo-alto.ca.us/leica-users/unsub.html