Leica M full frame CCD sensor camera DIY prototype - with pictures!

Discussion in 'Leica and Rangefinders' started by yongfei, Dec 12, 2008.

  1. I have been interested in Leica M camera for serveral years and I want to see a full frame Leica M digital camera. Most people would say that it can't be done under the current CCD sensor technology. But I want to do my own experiment to find out the possibility of a full frame Leica M camera. Now I have acquired a 24x36MM digital back, so I used it as the back end. First, I tried to use the Leica 35mm/1.4 Summilux as the lens. But found the lens end protruding too much. So I switched to Industar 55mm screw mount lens as the lesn for the testing of my prototype camera. I am so excited that I have got crystal clear pictures tonight!
    So my assumption is: a full frame sensor on M camera is technically possible, but it requires the redesign of some of the Leica M lenses. Also, for wide angle, maybe SLR type design will be needed to avoid dark corner vignett. But I think some people would be happy to sacrifice a little quality on the wide angle side, in return for a full frame Leica M camera.
    I will post some pictures here soon.
    00RmbG-97245584.jpg
     
  2. I am still processing a few test pictures and will post them soon.
     
  3. Another picture
    00RmbQ-97247584.jpg
     
  4. This is amazing and ambitious. That you would even attempt to make it is impressive by itself. I would love to see some images. I'm no design expert, but It looks like you might be able to squeeze that into an M5 sized body and not an M6. That might be the holdup for Leica though likely not for the user. Where did you get all the required electronics to work with the CCD? Is that also from the same DSLR? Very impressive.
     
  5. Orville, thanks for your encouragement. My real intention is to use the internet to show Leica AG that it is technically possible to develop a full frame M camera. Yes, it would much look like a M5 or even bigger.
    My main components are readily available: Eyelike M11 digital back in Hasselblad V mount, Copal #1 shutter, and a lens plate made of three 3" floppy disks. The shutter has a 40mm thread. So I need to hold that 39mm SM lens tight to take pictures.
    Therefore if I can do it, I am sure Leica company will also be able to do it. But whether it is financially feasible is another matter. But I belive that a full frame CCD sensor M camera can be made under current CCD technology. And my mission is to prove it.
    00Rmct-97257584.jpg
     
  6. This is one of the full frame test pictures taken by the Industar. If you look closely, you can find some dark corners. It is at about F8.
    00RmdY-97261584.jpg
     
  7. Here is another test shot this evening. I will take a few outdoor pictures tomorrow. If someone is wondering how a Noctilux 50/F0.95 will perform on the full frame, he can send the lens to me and I will test it for free:)
    To be practical, I will focus on tesing this screw mount lens first, then figure out a way to mount the M bayonet lens, for which the short flange distance is really a challenge for adaption...
    00RmeB-97267584.jpg
     
  8. Like many of you, I have been wondering how the wonderful 35MM/1.4 Sumilux will work under full frame. Since I cann't mount it yet, I took the risk of scratching CCD surface or the lens end glass, and handhold them together to take a few test pictures. Because the lens is about an extra 5mm flange distance away from the CCD, it only alows me to take pictures at very close distance.
    00RmeR-97269684.jpg
     
  9. That's cool, but is it a Leica? It may be a 'Linca.' ;)
     
  10. The front view of my prototype camera
    00Rmeb-97273684.jpg
     
  11. I used to make my own telescopes and once made a pinhole camera a long long time ago, but this is way beyond that. Once again, all the best.
     
  12. I can't tell the scale, but it doesn't exactly look like something that would fit into a Leica. What are you using as a shutter?
     
  13. Seriously, seriously cool. Nice job.
    Too bad more people don't hack cameras. We have all this wonderful technology and what do people do? Nothing! ;-) I'd love a Nikon FX sensor in an M. Can someone arrange that? :p
     
  14. Definitely a nice job of optical breadboarding!
    But as you say, wide-angles (which are the bread-and-butter of the Leica M rangefinder) are the fly in the ointment. A 50mm or longer is going to play better with a digital sensor than a 21 or 24 M lens. And so would redesigned retrofocus, SLR-like 21s and 24s (and that may be the direction Leica has to take, although not everyone will want to shell out $15,000 for a new set of Leica wides).
    Leica has already redesigned their 21 and 24 as (biggish) f/1.4 lenses, and I'll bet the optical designs are more accomodating to the needs of digital sensors, as well as being fast, than the old designed-for-film f/2.8s.
     
  15. In what way would the new lenses be more accommodating to digital sensors Andy? Don't forget they still make the M7 & MP. At any rate, lenses made 10 and 20 years ago already perform very very well with the M8.
     
  16. Get a beat up M body and slam the back on it. Pretty cool, let's see more pictures.
     
  17. Yongfei Lin --
    I am greatly impressed by your ability to assemble an operating digital camera from a variety of components and make it work properly. It clearly required considerable technical knowledge, mechanical skill, and initiative to do this, and the result is admirable. What you have done is comparable, in certain ways, to Oskar Barnack's development of the prototype of the original Leica, the so-called Ur-Leica. The quality of the sample photo you took with the Industar is quite good, considering the lens and camera used to make it.
    A full-frame digital Leica M body would be of interest to many Leica photographers. The M8 is good, but the crop factor is one of the things (along with price, color-shift and reliability issues) that may be keeping it from wider acceptance.
    There are, however, some characteristics of the camera you have developed thus far that might possibly explain why Leica has not yet provided a full-frame Leica digital M, and has instead provided the M8 in its current crop-sensor design. This is not intended as a criticism of your excellent effort, which is truly groundbreaking, but rather as speculation about the possible reasons why Leica's engineers might have gone a different route in developing the M8. These include the following.
    • As you note, the setup you have developed will not allow the use of existing Leica wide-angle lenses in M-mount, without some sort of adjustment for the protrusion of the rear elements of these lenses into the camera body. (You got a 35mm to work with it after a fashion, but apparently not in the same way as the 50mm.) Since wide-angle lenses are quite popular with Leica users, one of Leica's design goals may have been accommodating existing wide-angle lenses on a digital M body without modification. Leica's designers apparently had difficulty designing an M8 digital sensor that would clear the back of wide-angle lenses but still sense the light from them well enough to avoid unacceptable light fall-off in the corners of the image.
    • The distance from the rear of the lens mount to the rear of the sensor body on your prototype appears to be deeper than the front to back depth of a traditional Leica M film camera body. One of Leica's design goals may have been to produce a digital M body with the same approximate physical dimensions, including front to back depth, as existing Leica M film camera bodies. Leica's designers may have had in mind the degree of customer resistance that the firm encountered in the 1970s when it tried modifying the size and shape of the traditional Leica M body to accommodate the exposure metering technology then available, resulting in the M5, a camera larger than an M4 that was not as successful as the firm had hoped. The M5 was enough of a flop in the marketplace that Leica almost abandoned the camera business before resuming production of the M4. This dimensional constraint may have been one of the reasons why Leica's designers had a hard time getting the M8's sensor to clear the rear components of wide-angle lenses.
    • It is not clear how expensive the full-frame CCD device would be, or how expensive re-engineering the Leica M8 body to accommodate it would be, in terms of manufacturing a marketable commercial product as opposed to a prototype. I do not know, one way or the other, whether it would be more expensive than the current M8. Leicas in general, and the Leica M8 in particular, are already expensive enough to be luxury items that not all photographers can afford. It is possible that Leica's engineers concluded that a full-frame digital Leica M was technically possible, but would be too expensive to sell at a price affordable to consumers.
    • For the time being, the Leica firm appears to have directed the focus and work efforts of its engineers into the development of the new S-system digital SLR, rather than into further development of the M8 as a digital Leica M rangefinder camera. As the majority of professional photographers now use digital SLRs, the S-system may be a make-or-break product for Leica, especially in today's difficult economic environment. If the firm can market the S-system successfully to professional photographers, though, that might make it easier to pursue further development of a full-frame digital M.
    My speculation about those issues does not detract in any way from my admiration for your accomplishment in assembling a working digital camera as an individual inventor, without the resources of a large industrial company behind you. That is quite an accomplishment. I hope that it encourages Leica to take a closer look at whether a full-frame digital Leica M body is not only technically feasible, but economically possible. You are providing an incentive for Leica to continue examining these issues in a way that few if any of the rest of us could. I encourage you to continue working on refining your prototype, and look forward to hearing more about your efforts.
     
  18. Today I am quite excited. Because I know if I ever successfully capture an full frame image under daylight condition, that would the the first strong prove to every Leica lovers that a full frame Leica is indeed possible. Ok, I did it with a $20 russian screw mount lens, but I will continue further if I have the resources to custom made a few adapters. Right now, all my components are non destructive and resuable in nature.
    So the first thing I do is pointing my camera to the street, and I get the following picture. Seems the lens distance is too far away from CCD, so the camera couldn't focus to infinity. But what a beautiful out of focus blur! It wouldn't win any photo contest. But for a test shot, it is decent enough to prove that I do have a functional full frame M39 mount camera! And as you can see, there are no visible darkness in the four corners. Also, the color from the MF digital back is very neutral. I only applied very little curve this image.
    00Rn3b-97421684.jpg
     
  19. So I made the lens plate a little thinner then took a few more pictures. I was really nervous because shutter retainer ring is so close to the CCD, it could scratch it.
    Now I got almost perfect focus in infinite. On the lower right side, there is a visible dark corner, but I think that is due to some extra duct tape.
    I am happy with the testing result under daylight condition. I would like some comments to see if you agree this is a good full frame camera.
    00Rn4N-97429784.jpg
     
  20. very impressive yongfei. do keep it up and i wil be checking back regularly for updates! :)
     
  21. Hi Peter, thanks for your compliment. You've raised a lot of valid points. I agree with most of your comments. In deed, building a prototype is one thing, but commercial success is a totally differnt game, especially for Leica M which has such a religious following.
    My experimental objective is purely on the technical side: to capature photos using Leica M lenses (or equivalent) on a CCD ful frame sensor, with ability to focus into infinity.
    As for the 35MM/1.4, I still haven't managed to get a picture to evaluate the real world photo quality under full frame. I use a Copal leaf shutter which will cause the lens end to hit the shutter blade. Because a Leica M style body has a shutter near the camera back, this lens will not hit Leica M's shutter at all.
    So if I can mount the shutter on its lens shade, and adjust the thickness of the lens plate, I should be able to get some decent images for evaluation purpose - with a real Leica lens!
    I will start to work this evening towards this objective. As I know people will be less interested if photos were not taken by a leica lens:) Hopefully I will be able to modify my prototype, and post some pictures taken by 35mm/1.4 tomorrow.
    To John: Yes, a beat up M body will be a good idea. Or even a M mount Voigtlander. But there is a problem: if the CCD mouting to the camera is not precise or firm, any tiny movement will easily scratch my CCD filter. My current approach is less destructive, and provides equivalent result for evaluation purpose. After all, camera is just a black box, although Leica's box is legendary, it is still just a box. What Leica can do, is to minimize that digital back, then build a mounting interface like Horseman's Digiflex camera which accepts Nikon lens.
     
  22. Those photos show reasonable results for any camera of moderate cost, and most impressive results for a home-built prototype, especially given the lens that you are currently using with it. The images uniform and reasonably sharp right out to the edges. You've made it quite clear that this camera works, and works fairly well.
    It's fascinating to be able to watch online as you make adjustments on the fly, such as making the lens plate thinner to allow the lens to reach focus at infinity, doing some quick test photos and posting the results. It's a bit like standing in Steve Jobs's driveway and watching while he tests the prototype of the first Apple computer in his garage. It's evident that you have a talent for this.
    I didn't see a viewfinder, rangefinder or viewing screen in your photos of your prototype, but you ought to be able to rig up some sort of external accessory viewfinder to improve ease of use. It wouldn't have to be precise, just close enough to give you a reasonable approximation of the angle of view of the lens. A quick and dirty approach might be to buy a disposable or inexpensive point and shoot camera at a local discount drugstore, hacksaw the viewfinder off it, and attach it to your prototype with duct tape or superglue. A more elegant, but slower and more expensive approach, might be to buy a Voightlander accessory viewfinder from Camera Quest and make the equivalent of a flash shoe to attach it to the camera body.
    While that wouldn't offer precise focusing, using zone focus and taking advantage of depth of field at smaller apertures ought to provide fairly good results for daylight shooting, as shown by your test shots.
    If you are looking for a further demonstration of what your prototype can do, you might try using it to take a variety of photos -- a portrait of a family member or friend, a landscape in the park or a nearby scenic area, a street photo from a busy part of downtown, maybe a photo of a sports car that fills most of the image frame. They wouldn't have to be anything exotic, just a range of different things to show that your prototype has the same degree of photographic flexibility as a basic point-and-shoot film or digital camera with a fixed focal length lens.
    Good luck with all this -- you've come up with the most interesting project I've seen on Photo.Net in years.
     
  23. I am simply out of words.
    This is really amazing. I will continue to follow this thread.
    Good work Yongfei.
     
  24. Hello everyone, thanks for all your kind words.
    This evening, I have started to test the 35mm/1.4 Summilux. There is a critical flaw in my test process, but I still would like to share with everyone about my thought process.
    I decided to increase the thickness of my lens plate, and move the shutter to the front. Here is how I did it:
    00RnEE-97481584.jpg
     
  25. Final assembly of lens plate:
    00RnEF-97481684.jpg
     
  26. Shutter in front of 35mm/1.4 lens
    00RnEH-97481784.jpg
     
  27. Here is the result. I think my #1 Copal shutter is the main reason for causing the circled dark corners. But it is still not clear if the lens itself is also causing the vignett. Because the shutter caused vignett is obviously the dominate factor here.
    Conclusion: I can't do any further test with a #1 Copal shutter. A No.3 sized shutter may solve this problem and to continue the test, but unfortunately I don't have a No. 3 shutter.
    So please don't the 35/1.4 coverage based on this photo. I think most likely t is caused by shutter!
    00RnEP-97483584.jpg
     
  28. So I moved on to test 50mm Summicron with the same approach:
    00RnET-97483684.jpg
     
  29. Same dark corners. But much better as it is a normal lens. I also tested the Industar again with the shutter in front of lens approach. It is further better than Summicron's as expected, because it is a 55mm lens.
    But again, this is purely because the shutter's openning circle is too small. I think 50mm Summicron will fully work. Tomorrow, I may change to shutter behind the lens configuration and do some outdoor testing under daylight condition.
    00RnEW-97485684.jpg
     
  30. Now comes the fun part. I have a PK Sigma 24mm/2.8 lens that will have an almost correct flange distance (for Pentax) if the shutter is mounted behind. So I tried this setup for test the retrofocus wide angle coverage on Leica M:
    00RnEb-97487584.jpg
     
  31. Surprise! This lens has the full coverage on Leica M!
    So it somewhat proves that if we can adapt some Nikon (Zeiss ZF) or Pentax lenses with an adaptor, we will have decent coverage on a full frame Leica M.
    If the new Leica wide angle lenses are refrofocus SLR type design, then Leica should be able to design a full frame Leica M. Otherwise, we can at least use some SLR type lens with a rangefinder focus coupled adaptor. But you need to manually transfer the distance value to the lens. Those adaptors were used to be available at Cameraquest.
    Tomorrow, I will do some outdoor testing with this Sigma 24mm and Leica 50mm Summicron. I truely appreciate your conitnued interest and support, which give me the motivation to carry on further at the risk of dropping my dimond-like lenses or scratching the jewelry-like CCD:)
    00RnEm-97489584.jpg
     
  32. Keep it up, this is a very good idea and it has a lot of potential! The way I see this developing is by removing the rear door on an M5 (or any other M) and mounting the sensor right where the film would be. With some precise testing and adjustment I could see this working nicely and it would alow you to mount lens using the mount and even focus them regularly. The trick would be in mounting your sensor properly behind the shutter and then turning it on before the shutter opens (maybe using the flash sync?)
    Good Luck
     
  33. Dear Peter, thanks for your poetic and encouraging words. Actually I am just a junior software programmer who spent most of his free time on photo websites in stead of IT forums. So I am more worried about my own daily job rather than trying to immitate that great Jobs:)
    For all those years, I have been reading all kinds of photo newsgroups without any real contribution or purpose. There have been so many outstanding photographers, so many rich collectors etc. But tonight, I feel very happy and fullfilled, because at last I am able to contribute with an important experiment on the feasibility of full frame Leica M - one of the hottest on-going topics here.
    As this prototype goes, my major objective is to have some idea on the Leica lens coverage. I have a Hasselblad camera that is what this back was bought for. And I will get a Horseman Digiflex soon. Because this is a thethered back, it is not suitable for Leica M style photography. But my long term objective is to get a Hasselblad CFV back. Once I have that, I will improve this prototype with viewfinder and a fotoman style external manual rangefinder etc. That will become an Apla style high end camera. But it will be a very portable solution, with the advantage of being able to easily change the lenses on site.
     
  34. Best news ever! Mission accomplished!
    After writing my above reply to Peter, I turned off computer and went to bed. It's been a long night for me, and I was happy with what I have accomplished so far, and was thinking about test plans for tomorrow. Suddenly with a strike of genius, I realized that I don't even need a #3 shutter to test the 35/1.4. I can turn off the light and test it in darkness, with shutter used only to trigger the back!
    So I got up from bed, hooked up everything and took this picture: Leica 35mm/F1.4 Summilux ASPH at F11. I have to say, the result is astonding, much better than I have hoped for! Now this is Leica lens quality, this is the full frame 35 wide angle result you've been waiting for so many years. It's here, it's here!
    00RnHT-97517584.jpg
     
  35. The above picture has no color calibration whatsoever, straight from the camera. With my limited photoshop skill, I did a little adjustment and made the following 100% crop. The image is of high quality and very solid.
    00RnHV-97517684.jpg
     
  36. It is a long time since I was on photonet but now I flabbergasted and need to express my respect for your little expiriment...lets just say "wow" that must be so much fun to do that I'm tempted to also buy a random digital back..
     
  37. With the above picture, I thought I have beautifully finished my Leica 35mm/1.4 testing, and detached my back. But suddenly I realized that I need to test this lens wide open at F1.4! Otherwise, you guys will bug me everyday for that photo. So here it is:)
    It has some purple coloration, but I think it can be easily fixed in software, or even automatically by using a 6 bit profile specific for this lens. (note: another possible reason for the purple color could be the flash light is striking directly into my lens. Notice how the wall above the cabinet is acting like a mirror in this photo. So the result at F1.4 could be better!)
    Now I am done with the Leica 35mm/1.4 testing. Please ask a lot of questions to validate my test procedure, but please don't ask me to test it under daylight. It is a too dangerous setup. During the above testing, I wished I had three hands to hold everything together, and a forth hand to catch something that might drop on ground. I was reallly risking with my Leica lens and digital back for nothing:)
    Any comments? If my test is valid, we need a volunteer who has a good PR skill to talk to Leica company about those "individual" findings, and hear their comments:)
    00RnHi-97521684.jpg
     
  38. Yongfei, your are definitely on to something here.
    The only reason that Leica won't/can't do FF is that the traditional wide angles would suffer too much (corner vignetting) to be usable wide open. Or so it seems from your experiments. But normals work!
    But as you say, there would be no problem with retrofocus designs. If Leica purists would swallow that is debatable, but it COULD work.
     
  39. Ronald, if my above testing has no major flaw, it proves that this 35mm F1.4, which could be the most important lens in Leica M family, works almost perfectly. What more can the Leica purists ask for other than this lens mounted on a full frame Leica M?
    I have no other Leica/Zeiss ZM wide angle lenses to experiment to approve or disapprove my assumption about the image coverage of existing wide angle lenses. So Leicaphiles, now it's the time to donate your Leica/Zeiss lenses to me for some further testing:)
     
  40. "... if my above testing has no major flaw, it proves that this 35mm F1.4, which could be the most important lens in Leica M family, works almost perfectly"

    Sorry, the above photo demonstrates quite the opposite to me.
     
  41. I concur with Doug.
     
  42. Doug, Vivek - doesn't the 35/1.4 vignette that much at f/1.4 on film?
     
  43. Ok, I might be getting a little too excited after "working" more than 12 hours on this experiment. I presented the photo as it was shot, knowing that they can be improved in the software.
    The F11 shot was obviously under exposed by at least 1 stop. Here is an adjusted picture with level and white balance. I have no complaints on this photo. Much beyond my expectation after the failure of the above #1 shutter before the lens test approach.
    00RnSK-97603684.jpg
     
  44. Now for F1.4 shot, there were three variable factors. One is my monolight flash on the left side in a 3x5 foot softbox that could overpower the shooting table. The other factor was the strong lighting was directly reflected from the wall into the lens (I didn't put on the lens shade, and had an uv filter on it), like shooting towards the sun. Also, this one was taken further away than the F11 shot, making the flash lighting to be more uneven.
    Now I would agree that the "raw quality" at F1.4 might not be acceptable to someone who use it all the time. But I am still not sure unless I have a safe way to shoot it under natural lighting condition which is Leica M mostly used for.
    Here is the white balanced F1.4 photo with the vignetting adjustment in Photoshop. Acceptability is in the eyes of beholders. But again, unless I have a chance to test under available natural light, I won't judge F1.4's performance based on this single picture.
    00RnSo-97607684.jpg
     
  45. Hello Yongfei, thanks for sharing the DIY project! One quick question for the f/1.4 photo, where did you focus on?
     
  46. Hi Bruce, this is a very interesting question:)
    Actually this is a prototype camera built with floppy disks and duct tapes. So currently I had no control of focus at all, because the focus scale is not accurate.
    But I did show that my previous F11 picture was focused on the cabinet. Because I moved back a little bit, so I expect the focus would be somewhere a little bit closer. Sure enough, the focus is on the copier paper! Here is a 100% of the above F1.4 photo.
    00RnXE-97635684.jpg
     
  47. Wow! I am truly impressed with what you are doing. You just don't hear about people doing this kind of thing anymore. You may very well end up starting your own line of cameras. I think it is very cool and retro that you used floppy disks to make the lensboard. Keep it up, I love this sort of experimental DIY stuff. It shows that creativity (as far as electronics go) is not dead. I myself am developing a cost-effective and useful LED lamphouse for my Beseler 23C enlarger.
     
  48. As planned, today I start to test the Sigma 24/2.8 first. Again, it was a lot of work to get it into some kind of correct focus.
    I feel that the color presentation of this digital back is very neutral.
    00Rnhe-97705584.jpg
     
  49. 100% crop of the left side tree.
    00Rnho-97707584.jpg
     
  50. Now I need to test 50mm Summicron. Because I have a shutter after the lens, I can't get it to focus into infinite. What do I test then? Well, that famous out of focus bokeh:)
    I think the lower dark corners are due to my lens plate openning is not big enough. So it is a little out of alignment with the back. No big deal.
    With these pictures, I have finished all my Leica M full frame camera prototype testing. In the past years, I've been reading many discussions about this full frame sensor topic, but have seen zero pictures. As people say, one picture is worth thousand words.
    Hopefully it is just a marketing decision that Leica doesn't want to pursue the full frame M. I believe the situation was similar to where Nikon was a few years ago, when Nikon executives insisted that DX format should be enough for professional use. This experiement suggests that, maybe, Leica M full frame is indeed achievable using the current CCD/CMOS technology together with some customized post processing software.
    00RniB-97711684.jpg
     
  51. Yongfei, this is an amazingly ambitious poject! I wish you luck!
    Question- can you shoot teathered to a PC with your back and get a live view image to focus correctly? Would probably be easier than guestimating focus, and with a sharp image, we could really see waht this thing can do!
     
  52. Since you are concerned about risking damage to your Leica 35mm f/1.4 (which seems reasonable), you might want to get a less expensive wide-angle to experiment with. While the Russian 35mm f/2.8 Jupiter-12 is available inexpensively (one recently sold for $80 on a well-known auction site), it has a large, exposed, convex rear element, so it would not appear particularly well-suited for your experiments. Stephen Gandy (www.cameraquest.com) has a Voightlander 35mm f/2.5 Color Skopar C LTM in black finish available for $229. He also offers the same lens in M mount, designated the 35mm f/2.5 Color Skopar P Type II, for $319. A lot less to worry about than your Leica 35mm f/1.4 if the shutter hits the rear element, or the lens drops off your 3.5" diskette lensboard while taking pictures outdoors...
     
  53. Michael, thanks for your good wishes. My struggle is not about guessing the focus distance, but about making it to focus at infinity. If you look at this Summicron-on-shutter picture, you will see that very long and sharp screw threads on the shutter end. So basically I have three choices: 1) cut it off to make it shorter; 2) reduce the lens plate further with a risk of scratching my CCD filter; 3) leave it as it is, and pretending that I don't even need a sharp picture to prove my point. My current choice is option 3;-)
    Nathan said it is retro, but I say it is really primitive. If I have free access to SK Grimes CNC machines, I will definetely use it:)
    00Rnis-97721684.jpg
     
  54. Right now, digtal back and "body" are just taped. In future, after I upgrade to a non-thethered back, I will get one of those custom made Hasseblad plate, and have a piece of thin wood or metal plate as the body. Then my last challenge is to adapt or make a Leica bayonet mount on the shutter. The 39mm screen mount thread is not a problem at all, as the #1 shutter has a 40mm thread.
     
  55. Here is the interface plate that I am planning to use in future
    00Rnk6-97729984.jpg
     
  56. Peter, as you've suggested, I may just use the scew mount 35mm to continue the testing. It is much easier to do. But I think people here are more interested in the 35mm/1.4 pictures. I may buy a No 3 or even larger shutter on eBay to continue with this image coverage test. But it could take a while to find a cheap but functional shutter.
     
  57. You might be able to hack-saw the thread end of the shutter if you place a 1" wooden dowel into the hole. The wood needs to be tight enough on the threads to support the shutter so you don't bend anything, and all the metal bits need to be carefully vacuumed so they do not enter the shutter and damage it or scratch the CCD. I suspect the black corners on some of the photos to be shadows cast by the shutter, they do not look like the edges of a light circle of a lens. Another thing, many of those fuzzy photos look like camera shake... you need to find a way to attach the camera to a tripod mount. There is actually a company offering a professional version of your machine here, using common SLR mounts:
    http://www.kapturegroup.com/true/wide.html
     
  58. The Sinar Copal shutter with a proper Leica M-> Sinar adapter board will be easier to do, since I saw your digital back already has a Sinar mount adapter. The Sinar Copal shutter has big enough opening so you can do it behind the lens, and the lens and CCD can be farily close to each yet safely apart.
     
  59. Hi Bruce, this Sinar board is not mine:)
    Anyway, the CCD needs be to less than 28mm from the M lens mount. The shutter itself takes 20mm. If I need 5mm for a safe CCD distance, then I have about 3 mm to make adaptions. Or even less if you consider the true CCD surface is under a glass filter.
    Anyway, I have learned a lot with this project and may just put the back to where it belongs: I have a Hartblei camera that can synch under 1/30 using this back. So I already have too much fun with the Carl Zeiss Jena MF lenses. This Leica M project is just icing on the cake.
    00RnrX-97799584.jpg
     
  60. jbm

    jbm

    Hey...great work! This is totally cool and I hope as a new rangefinder convert that the emergence of a FF leica M is someday a reality.
    One question: what is this eyelike digital back and do they make them for Mamiya?
    Keep it up...love to see you going MacGyver on your available resources.
    Cheers,
    Jay
     
  61. Great effort. Put it in a sexy black case with a red dot and you'll be able to sell it for a fortune.
     
  62. btw
    I do hope you're not the first one that puth M glass on a FF sensor
    There is an adapter for canon that allows you to connect it to the 5D etc
    ;-)
     
  63. Brilliant work. I think the next step, as has been suggested, is to get an M mount camera, take off the back and get a sensor in where the film would be. As for wide angles, I can not see why, with the current low light sensitivity of CCD that the programming could be made such that the centre is at one sensitivity and the edges higher. That is to say, if you get a 2 stop difference from centre to edge then allow the edge to be two stops more sensitive. If the relative performance was calculated for each lens and programmed into the CPU then the current Leica lens coding used by the M8 would allow for much better wide angle performance with FF digital sensor whilst still allowing normal lenses to utilise the full frame at equal ISO. This might mean a maximum relative ISO of say 800, with the edges for a wide angle at an equivalent 3200 for example, but wouldn't it be worth it to use wide angle Leica glass as real wide angles? This is not beyond the realms of what any serious camera company could achieve today - so come on Leica (and if they don't then I might).
     
  64. Jay, hope Leica can have the resource to further minimize its Imacon-R digital back and make it full frame. The Eyelike is similar to Sinar backs. It is an open platform back. You may use most of the MF or view cameras via dedicated adatptors (which is expensive or hard to find though). I like this open source idea.
    Alan, I would be happy to sell my stack of floppie disks for $100 which is a fortune for what they are. Then the collector can send them to Leica factory to paint an authentic red dot.
    Reinier, I am not sure if I am the first one who puts M glass on FF and took some pictures or not. I regret that I don't have enough resource to make it more accurate. One needs very little money and time to make a 80% functional camera. But to sequeeze the last 20%, one needs to spend a lot of time and a fortune.
    I believe the Canon adapter you are talking about is for Leica R, not M. Therefore, I claim myself to be the first (corporate or personal wise) in the world who has obtained Leica M lens image on a full frame sensor . Now I am waiting for Leica optical lab to show us otherwise, if they ever bothered to make some full frame prototype images to support their "currently full frame is not possible on M"claim. Also, don't forget to wait for Zeiss Ikon digital lab's response;)
    Stephen, I agree with your software enhancement approach, as shown on my Leica 35/1.4 test shots, a lot of things can be done by software.
     
  65. Reinier, I am not sure if I am the first one who puts M glass on FF and took some pictures or not.​
    I can assure you that Leica has full-frame M prototypes with sensors from a number of different manufacturers. A friend who is a Leica representative was in Solms earlier this year....Leica is testing but they are not satisfied with the results.
    Therefore, I claim myself to be the first (corporate or personal wise) in the world who has obtained Leica M lens image on a full frame sensor.​
    Unfortunately for you - Leica has the full frame prototypes...
    Now I am waiting for Leica optical lab to show us otherwise, if they ever bothered to make some full frame prototype images to support their "currently full frame is not possible on M"claim.​
    You're misconstrueing what Leica is saying. What they have said is that with the currently available sensors, they cannot provide a product that meets their performance requirements. There is a difference from providing a rudimentary image and making a camera that will work most all M lenses.
     
  66. Steve, thanks for your insider information. It's good to hear that Leica is finding ways to build a full frame M.
    Yes, I agree that putting together a prototype camera and get some images is one easy thing, but to get high performance is a totally different game.
     
  67. It seems many people would prefer to see a digital back that is mounted to a real M camera.
    I MAY have a way to continue this project with a Voigtlander T or other M mount cameras. I prefer Voigtland because it is easier to modify, and also because it has 1/125 flash synch.
    But first I need to find a functional, used M mount camera. Destroying a brand new camera is not an environmental friendly behavior.
    If someone can help me buy a cheap one, please post your information here.
     
  68. I agree that to make a full frame M camera that will work with most of all EXISTING M lenses may be a mission impossible. But technically, as my experiment shows, you can use Zeiss ZF equivalent lenses on the wide angle side and (relatively speaking) easily build a M with the current 16 bit full frame CCD techonolgy.
    Yes, this is a compromise. But to some users, this is less an compromise in nature than a cropped size CCD. Besides, all Nikon, Canon, Hasselblad and other manufactuers are using retrofocus design in their wide angle lenses. Leica can beat them with high precision, all metal design lenses. Besides, I think the current M8 is actually using 8 bit final raw file, which is against Leica's own claim of its high performance standard.
     
  69. One of the problems with slapping this back on a real Leica is that the sensor itself is behind a glass protective barrier. In film cameras with a focal plane shutter, the shutter is placed as close to the film as physically possible... in order to maximize space behind the lens. On interchangeable back medium format cameras, the film plane is behind the mount by a few millimeters to allow for the dark slide and the mount itself. This will make it impossible to mount a medium format back to a 35mm focal plane shutter camera and still attain infinity focus. What really needs to be available is a focal plane shutter not attached to a camera, with mounts on both sides. I think there would be a market for such a device, especially in 6x6 size, as it could be used specifically for mounting 35mm camera lenses to 6x6 backs, as well as for mounting on LF cameras directly behind the lens. I've often thought it would be useful to strip down a Pentacon for just such a purpose... but it would be an expensive affair. Another thing you may consider is finding some Visoflex lenses, which will have a much longer registration distance.
     
  70. Yongfei Lin, I just discovered this thread and quickly read through it. All I can say is that I offer you my encouragement and applause for the work you are doing. Keep it up, my friend!
     
  71. What a great bit of fun! Thanks for sharing your adventures in camera building.
    Godfrey
     
  72. Patrick, really appreciate your thoughtful concern and understanding. However, most people here would like to see a digital back put on a real M camera, with nothing more or nothing less. So I have two choices here: 1) stop right here as it is today, believing that my box is conceptually an equivlant to a Leica M. 2) do as other people's wish (as well as my inner devil's wish) to put the CCD filter as close as possible (within 1mm) to the shutter. I should be able to get some decent images before the filter glass gets noticeably scratched. Does any insurance company provide an affordable CCD filter replacement insurance? I may need it;-)
    If you look at my M11's picture, you can see that it has a "naked" CCD frame if the Hasselblad interface is taken away. Beautiful, isn't it? That's why I like the idea of open platform: buy a digital back for once, and play with it on all platforms and DIY projects. I think it is risky but possible to mount it on a Voigtlander M body. My secret plan will be: 1) cut a 24x36mm hole on my CCD's alminimum protective cover (unshown here, but it is something of the size of a match box); 2) glue this cover to those two film rails outside the shutter. 3) push back's CCD inside the protective cover as far as possible, right before the filter glass touches the shutter.
    So it is easy for me to come up with this interfacing approach, but it really requires some high degree of precision to implement it.
    BTW, Horseman's Digiflex can connect a Hasselblad V digital back, with Nikon lens mount. It seems that it was custom build on some Voigtlander's Bessa body parts. But like you said, they must have recessed the shutter inside the body to compensate for the increased distance caused by the back.
    Godfrey, nice to see you here! I still remember reading your shared experience about your Rolleiflex cameras a few years ago. Then you moved on to some Sony or Olympus digital gear. Then I lost track of your adventure. But hope you didn't get lost in this digital jungle yourself;-)
     
  73. No, that adapter is for leica M
    but as I understand there are issues with focussing at infinity...or better that is not possible apperently ;-)
    So at least your the first that gets "sharp at infinity" images with a FF and Leica M glass :D
     
  74. Reinier, good to know that such an adapter exists. Yes, people can adapt normal Leica M lens to as many SLR bodies as they wish, as long as they abandon the hope for infinity focus;-)
    BTW, I've found a way to further reduce the lens to CCD distance by another 1mm. And I will try to mount it on a tripod. I will retest the Summicron 50mm and a few other SLR lenses this weekend. Hope to get some sharp images this time. Stay tuned...
     
  75. Yongfei, thank you for sharing your experiments! I thought following page may be of interest to you:
    http://www.16-9.net/lens_tests/15mm_test1.html
    The Voigtländer lens tested there is optically identical to 15/4,5 in leica mount, and it behaves surprisingly well n front of Canon's full-frame sensor. Maybe this lens would take your central shutter too, if you could saw off its rear threads?
     
  76. The above link is very impressive, especially the attention-to-details during the testing. I need to learn from that for my own testing procedure.
    Lens selection wise, I am more interested to know the image circle coverage of each Leica M compatible lenses. Especially, Zeiss ZM wide angle makes me very curious. It would be very interesting to know their performance under full frame.
     
  77. Probably stupid questions, but...
    Mounting a full frame CCD sensor behind the lens in the traditional way doesn't allow for enough room when using lenses on the Leica M mount, right? So, I wonder if there is way to mount a thin mirror where the usual film would be and then bounce the light to a location where the full frame CCD sensor could be located.
    And...Why can't manufactures just the make the camera body "thick" enough to accommodate a full frame sensor and the Leica M mount?
     
  78. Bob, maybe you can take a look at this link: Camera Mounts Sorted by Register http://www.graphics.cornell.edu/~westin/misc/mounts-by-register.html
    Ideally, there should be just one universal mount specification for all 135 cameras. But the manufacturers just don't want their cheese to be moved away.
    The practical value of my testing approach is: if you can use this method to test all your expensive Leica lenses, you will be able to know which lenses work well under full frame and which ones don't. So if the Leica or Zeiss Ikon full frame camera ever comes out and becomes popular, you may want to sell off those not working lenses ASAP, as their second hand value may go down big time.
     
  79. Couldn't you, just to test the 35mm in daylight, remove the shutter, put some NDs in front and use a lenscap or black cloth as shutter?
     
  80. Hi Yongfei! I am really enjoying your project! I think you should start a weblog about it. I will certainly follow what you're doing.
    Maybe an idea to test your 1.4 in daylight: there are manuals from the early 20th centrury on how to make a shutter out of wood (yes, wood!) with a fixed shutter speed of around 1/30 s. They are meant to go in front of the lens. In combination with a strong ND filter (because the daylight will be too bright for 1/30 at f/1.4) this could be the way to go.
    Or if your ND filter is strong enough: no shutter at all, use a lenscap as a shutter. But make sure you can mount your camera on a tripod.
    Keep up the good work!
    Willem
     
  81. As promised, I tried to do some more testings on 50mm Summicron and 35mm Summilux. It is a very difficult task. It took me a total of four hours to get a few usable pictures. Focusing is purely a matter of luck. Last time, I got sharp pictures with 35mm even at 1.4, but the lighting was too strong. This time, I got very good luck with 50MM, but not so with 35mm.
    Sunpak 120J was used as flash (on the left side). Shutter was only used to trigger the digital back. Shooting needs to be in total darkness for the back to come into "ready to shoot" status. So I have to restart the Eyelike capture application almost after each shot.
    00RqLj-98863584.jpg
     
  82. As you can see, the depth of field is very shallow. That could be because the lens is not totally evenly "mounted". So it effectively becomes a tilt lens. Here is the 100%.
    00RqLn-98865584.jpg
     
  83. Another 50mm, forgot the F stop. Should be below F5.6
    00RqLr-98865684.jpg
     
  84. Here is the 100% crop
    00RqLs-98865784.jpg
     
  85. 35mm is a total miss. But fortunately, I got a sharp picture last time. I just post here a F5.6 result. It is still out of focus.
    With those pictures, I think I have completed my Leica M lens full frame testing effort. It has been a fun experience. And I am glad that I can share those findings with a wonderful group of people who are passionate Leica lovers.
    00RqLv-98867584.jpg
     
  86. Sorry I took so long to respond - offline on the high seas for a week.
    "In what way would the new lenses be more accommodating to digital sensors Andy?"


    First design them more like an SLR 14 or 20 or 24 - with lots of back focus space. Then increase the diameter of the rear elements to be larger than the sensor diagonal. And then make the rear element(s) something of a transfer or corrector lens that redirects the image to project straight backwards onto the sensor.
    Doing each or all of these make the light rays travel more directly into the face of the sensor, rather than outward at an angle from some central pupil-point in the lens. This lets the light travel straight through any microlenses into the "wells" of the silicon that are light sensitive, instead of being offset and hitting the walls of the wells (Think of a city street shaded by tall buildings in winter, but brightly lit at noon on a summer day).
    "Digital" lenses would still work on film, so there's no reason they can't perform on an M7. Their drawback, if any, would be that they are generally quite big compared to Leica's film lenses, for the same focal length and aperture. (compare a Canon 20mm f/2.8 with its 72mm filter size to a Leica 21mm f/2.8 that only requires 55mm diameter, or the really old Superangulon f/3.4 that used 49mm filters). And that they require an extra bending of the light rays in back to push the wide-angle image back to parallel rays, which likely means additional aberrations that then ust corrected themselves.
    Yes, Leica's lens range generally works well on the CROPPED M8 sensor. But we're discussing a larger sensor in these experiments.
    Note that the DISTANCE from the lens to the sensor, in and of itself, is not the problem - there are lenses in small digicams (including the Leica Digilux 2) that practically touch the sensor at all "focal lengths" But they are designed to catch the incoming image and point it towards the silicon perpendicularly, whereas legacy film lenses tend to spread out the image.
    Leica's new 21 and 24 designs tend towards those attributes that favor digital (without discounting film) - larger rear elements to physically cover the whole image area from in front (goes along with the f/1.4 apertures) and longish tubular "light pipe" internal layouts to keep the light generally moving parallel in the first place.
     
  87. Yongfei, what atr those amps you're using w/your hifi?
     
  88. On my Hi-Fi system, the amps are Stratos http://www.odysseyaudio.com/ and Audio Note Kit-1. And I am using Lavry DA-10 D/A converter with volume control as the pre-amp. The loudspeaker is a DIY design after Dynaudio Contour 1.3.
     
  89. Thanks Andy.
     
  90. There have been a few discussions about this project on some internet BBS . Obviously, some of the guys didn't read the whole thread. So I'd like to post some of my conclusions here:
    - There are multiple purposes associated with this experiment. At the beginning, I want to put things together just for fun. Later on, I became quite curious about the performance of my Leica 35mm/1.4 on a full frame CCD. Because if the image quality is not good, I will sell it before its resale value starts to go down.
    - I tested Leica 50mm F 2.0. There is no obvious problem with this lens on a full frame CCD.
    - I don't have any Leica wide angle lenses, so I tested a Sigma 24MM SLR lens. The full frame result is quite good. So I can't say if the Leica/Zeiss wide angle lenses will work or not on a full frame CCD.
    - In the current configuration, I use a Copal shutter behind the lens. So sometimes the total distance is to far for infinity focus. The real M type shutter will not run into this problem. But to put a digital back on the M camera at the film plane rail, is very easy to scratch the CCD.
     

Share This Page