M9- full frame, thought it was impossible?

[matt_m__toronto_]

<p>in the past i recall reading countless posts about how a full frame leica m rangefinder would be technically impossible. (due to such things as lens/film plane distance, and difraction? and other tech terms i didn't fully grasp but read with some interest) so...what 'impossibles' were overcome to make a full frame leica rangefinder a possibility?</p>

[Alex_Es]

<p>Quite a few "impossibilities" were overcome, Matt. The latest brochure explains it better than I can. Less than a decade ago a digital Leica M was considered impossible. </p>

[doug herr]

<p>One of the best ways to get something done is to tell an engineer that it's impossible :)</p>

<p>I suspect that one of the most important technological developments that made the M9 possible was Kodak's newly-developed technique for sandwiching the IR filter between the photo sites and the microlenses. Before this, IR filters were on top of the microlenses and were directly exposed to oblique light in the corners. The light is much closer to perpendicular after passing through the microlenses so the IR filter can be made more effective. It's still not 100% effective with all lenses because of the variation in angle of incidence, but this IR filter placement along with the coded lenses (or manual lens selection) gets the system within acceptable range.</p>

 

[jody_stowitts]

<p>Doug Herr: I've been seeing your beautiful pics for several months now, and I realize you are an avid proponent of Leica glass, especially your (resurrected) 280APO. My question(s) to you are: Have you completely given up film for your nature-wildlife shoots? And, is your digital rig of choice still R8,9 and DMR or something(s) else? And finally, when going digital, do you have to (in my own words) "monkey around" with them on one or more digital photo software programs, or can they come right out of the camera looking as good as your chromes of old?</p>

[aplumpton]

<p>"It's still not 100% effective with all lenses because of the variation in angle of incidence,..." </p>

<p>Douglas, that is a possible Achille's Heel of the new M9. How can one fixed microlens geometry conform to all the focal lengths available to the M9? And for those like me with some older lenses, or the less performing Leica ones, how will the off axis peformance be reconciled at full field, compared to the more limited feld imposed by the 1.33 X crop M8 series?</p>

<p>I hope these concerns are proven to be of minor importance. We have of course no tests yet of the M9 to get enthusiastic about.</p>

[thomas_hardy1]

 

 

 

<p>"It's still not 100% effective with all lenses because of the variation in angle of incidence,..."</p>

<blockquote>

<p>Douglas, that is a possible Achille's Heel of the new M9. How can one fixed microlens geometry conform to all the focal lengths available to the M9? And for those like me with some older lenses, or the less performing Leica ones, how will the off axis peformance be reconciled at full field, compared to the more limited feld imposed by the 1.33 X crop M8 series?</p>

</blockquote>

<blockquote>

 

</blockquote>

 

I think the same could be asked about the FF cameras from NIkon, Canon, and Sony. How do they perform with all of the lenses they have to contend with, especially the wide angle focal lengths?

 

 

[doug herr]

 

<p ><a href="../photodb/user?user_id=2347092">Arthur Plumpton</a> wrote: <strong>"How can one fixed microlens geometry conform to all the focal lengths available to the M9?"</strong></p>

 

<p > </p>

<p >The short answer is that it can't. The longer answer is that the microlens geometry is a pretty good compromise for the various lenses you can use on an M9, and the camera's firmware recognizes the properties of each lens and adjusts the image especially in the corners to compensate. The firmware has a database of M lenses (I don't know how complete it is) and unlike the M8 it allows the user to manually select the lens in use, for those whose lenses do not have the 6-bit optical code.</p>

<p > </p>

<p > </p>

<p > </p>

 

<p ><a href="../photodb/user?user_id=5211783">jody stowitts</a> <a href="../member-status-icons"></a>wrote: <strong>"I've been seeing your beautiful pics for several months now, and I realize you are an avid proponent of Leica glass, especially your (resurrected) 280APO. My question(s) to you are: Have you completely given up film for your nature-wildlife shoots? And, is your digital rig of choice still R8,9 and DMR or something(s) else? And finally, when going digital, do you have to (in my own words) "monkey around" with them on one or more digital photo software programs, or can they come right out of the camera looking as good as your chromes of old?"</strong></p>

<p > </p>

<p >Another short/long answer: I'm using the R8 with DMR almost exclusively now. I rarely have to do much to the DMR's raw files other than convert to .TIFF and .JPG, normalizing white point & black point, and to clean up any stray dust shadows. There have been a few photos where I had to do some trick masking and selective sharpening or density adjustments, but if these had been on film the images would have been tossed. The only software I use is Imacon FlexColor and Photoshop 6.</p>

<p > </p>

<p >I might not be the best judge of whether they're as good as the chromes from my much-loved Leicaflexes, so consider the response of Steve Appel, owner of Appel Gallery here in Sacramento (http://www.appelgallery.com) when I showed him a set of prints last month. The prints were a mix of film and DMR, two brands of camera, three brands of lenses with no identifying labels. Mr. Appel made a big fuss about the color quality and detail of some of the larger prints and asked what camera I was using (R8/DMR with 280mm f/4 APO). At the end of our meeting he selected several prints for display in the gallery and as it turned out all were made with the DMR, most with the 280mm f/4 APO.</p>

<p > </p>

<p >In a perfect world I'd be able to get DMR quality from a Leicaflex SL, or put a Leicaflex SL viewfinder in the R8. YMMV.</p>

 

 

[aplumpton]

<p>Douglas, thanks for your kind comment on the microlens.</p>

<p>"the camera's firmware recognizes the properties of each lens and adjusts the image especially in the corners to compensate"</p>

<p>This may correct vignetting and perhaps even colour shift, but I will be interested to see if it corrects other aspects of varied lens use with a fixed microlens geometry. </p>

[geoffs1]

<blockquote>

<p>I think the same could be asked about the FF cameras from NIkon, Canon, and Sony. How do they perform with all of the lenses they have to contend with, especially the wide angle focal lengths?</p>

</blockquote>

<p>Because of the reflex mirror in an SLR, lenses for those cameras have always had longer rear-element to film-plane distance (and less acute light-ray to film-plane angle). For example, wide-angle SLR lenses almost always use a <em>retro-focus</em> design.</p>

[leicaglow]

<p>Well, they didn't follow the <em>experts </em> here on Pnet.</p>

[kens]

<p>It was "impossible given the current technology" of the time. Today's new technologies = new solutions to old problems. I don't think many people thought it would be forever impossible - just a matter of time, persistence, and research $$$'s. Refinement of micro lens technologies and IR filter technologies, plus faster processors on smaller chips, were all important to finally make the M9 possible today.</p>

[joseph_wisniewski]

<blockquote>

<p>I suspect that one of the most important technological developments that made the M9 possible was Kodak's newly-developed technique for sandwiching the IR filter between the photo sites and the microlenses.</p>

</blockquote>

<p>The M9 has a fairly conventional 0.8mm glass IR blocking filter as the chip cover glass.</p>

[35mmdelux]

<p>"Well, they didn't follow the <em>experts </em>here on Pnet."<br>

Easy there, pilgrim.</p>

[joseph_wisniewski]

<blockquote>

<p>How can one fixed microlens geometry conform to all the focal lengths available to the M9?</p>

</blockquote>

<p>It's not that hard, really...<br>

The exit pupils of Leica lenses range from about 28mm from the focal plane (lenses wider than 35mm are normally slightly retrofocus, to about 90mm (for a 135mm telephoto, which is a Galilean layout). The FF sensor is 21.6mm from center to corner. That gives you a range of angles from exit pupil center to sensor corner of 37.7 degrees for the worst wides to 13.5 degrees for the longest teles...<br>

The average of those two angles is 25.6 degrees. (That's a bit much to offset microlenses. Hopefully, Kodak adopted Nikon style dual layer microlenses, and offset each of the two layers 12.8 degrees). Set the microlenses to this angle, and the worst error you have is 12.1 degrees. Same error, outward for a telephoto, inward for a wide.</p>

<p>The Kodak 6.9 micron cell with standard microlenses is a half stop down at 29 degrees from perpendicular, a full stop down at 34 degrees from the perpendicular. Assuming Kodak didn't screw up the offset microlenses (and Kodak is pretty good, my sources say the 0.5mm Kyocera BS7 filter debacle on M8 was Leica's doing) that means that there's no noticeable attenuation of "chief rays" (rays directly from the center of the exit pupil) and you have to get an additional 17 degrees away from the exit pupil before you have half a stop of falloff on the worst edge of the exit pupil (and no falloff on the opposite edge). f2.0 lenses have marginal rays +/- 14 degrees from their chief ray, you shouldn't see vignetting with any f2 lens over the entire focal length range.</p>

<p>50mm lenses tend to be the fastest, but they're in the catbird seat. If they're relatively symmetrical and have an exit pupil at 50mm, the angle from pupil center to sensor corner is 23 degrees, just 3 degrees (OK, 2.6 degrees) from the microlens offset angle. That's a perfect optical interface, and even a 50mm f1.0 should be fine.</p>

<p>So, basically, the only way Leica could be in trouble is if Kodak only used a single microlens layer and had trouble achieving 25.6 degree offset.</p>

 

<blockquote>

<p>And for those like me with some older lenses, or the less performing Leica ones, how will the off axis peformance be reconciled at full field, compared to the more limited feld imposed by the 1.33 X crop M8 series?</p>

</blockquote>

<p>Hopefully, M9 will outperform M8, which uses a single microlens layer and (if memory serves) an 8 degree offset, when it's optimal offset would be closer to 18 degrees.</p>

[andy_piper2]

<p>It's interesting to note (as one who's actually looked at the M9 sensor from 6 inches away) that the M9 sensor looks a bit different from most sensors, and certainly the M8's. It has less of the green tint normally visible, and surface reflections have a diffraction rainbow striping, like the reflections off the recording side of a CD or DVD.</p>

<p>It is also interesting to note that my shots with a 135 Tele-Elmar (an uncoded lens) showed a very slight gradual darkening towards the corners when used on the M9 "straight" - which was eliminated when I identified the lens in the manual list in the camera's menu. So the camera is doing some vignetting corrections for telephotos as well as wides. The only unknown is whether that is just normal optical vignetting being corrected, or something induced by the sensor.</p>

<p>What Leica originally said (over 5 years ago) was that a digital M camera was impossible with current sensor architecture. Once it became clear that they were in a life-or-death situation, they moved (with help from many others, notably Kodak) to change the architecture rather than waiting for it to be spoon-fed to them. The sensor changes were subtle - the other factor was image processing to correct the remaining vignetting and color shifts from different lenses (thus the coding requirement).</p>

<p>In 1958 it was not technologically possible to put a man on the moon. In 1968 it was. Times change.</p>

[mike dixon]

<p><em>If they're relatively symmetrical and have an exit pupil at 50mm, the angle from pupil center to sensor corner is 23 degrees, just 3 degrees (OK, 2.6 degrees) from the microlens offset angle. That's a perfect optical interface, and even a 50mm f1.0 should be fine.</em></p>

<p>The exit pupil on my 50 Summilux is about 30mm from the film plane(when focused at infinity); the 50 DR Summicron is similar.</p>

[aplumpton]

<p>"In 1958 it was not technologically possible to put a man on the moon. In 1968 it was. Times change."</p>

<p>In fact it wasn't even thought possible in the mid 50s to put a small Sputnik into orbit.</p>

<p>Interesting point about the 135 TE, as well as Joseph's data on microlenses and the compromises that must be made with sensor and microlens design that are virtually absent with film use. On the other hand, numerical correction of vignetting alone is a good thing, especially with the very fast lenses.</p>

<p> </p>

[david j.lee]

<p>so i guess what comes next is the full frame,replacement-do it yourself digital back for my M3 and M6. it will include, of course, a device that will let you change from digital to film and back instantly. ( it will be available in black only )</p>

[kelly_flanigan1]

<p>The 35mm cartridge with a digital sensor was something that was thought about back in 1992; when early digital cameras were being used by tiny fraction of folks; thats 17 years ago. Later we had the Silicon Film attempts; which ended about 1.3 megapixel. The 24x36mm film aperture/film plane is different with respect to the cartridge almost every 35mm camera ever made; thus there cannot be a drop in gizmo unless it has a flex cable deal to the sensor; and a mess of variations/versions for each camera brand and model. Maybe once any Patents and legal issues are gone; one might get a Chinese digital adapter gizmo in a bubble pack if we are lucky; vary lucky. This "dream" is ancient; I heard about this at a converence in Reno back in 1992; at photo/printing trade show. Back then in 1992 a VGA camera was several grand and only pros used them; a few megs of ram could be a grand; Photoshop was then more expense than today in real dollars and only a Tiny fraction in ones cost of shooting digital; ie in the noise.</p>

[kelly_flanigan1]

<p>The M9's full frame impossiblities must of been overcome via a newer sensor; more tolerant of light way non-perpendicular; ie the corners. Its probably a compromise; hopefully good enough.<br>

If one is "an observer" at the corners of a 24x36mm frame; one "sees" the light more straight up; ie high noon with my 16mm Nikon F nikkor; than my VC 15mm lens. The 16mm is a retrofocus optical design to clear the Nikon slr's mirror; the 15mm is not. "An observer" ie little man at the corners of a 24x36mm frame "sees" the light less straight up; ie more grazing with the non retrofocus 15mm VC lens.<br>

It was impossible in the past; they didnt have a sensor more tolerant of less than normal light bundles.<br>

M8's and RD-1's and M9's production volumes are puny ; little compared to say a Canon dRebel; thus one as a camera maker has little leverage with a sensor maker for a low run high buck camera.<br>

On hundreds of past threads some folks with zero manufacturing experience seem to think that production volumes do not matter; thus their brain wants a 500 buck M9; thinking somehow the tooth fairy will magically pay for the tooling and development costs.<br>

There is a reason one can buy a Harbor Freight 4 1/2" angle grinder on sale for 17 dollars; and a M9 costs many thousands. The grinder is basically a mature design; with layout defined 40 plus years ago; it is sold in massive volumes of millions; it is built with low cost labor. The M8/RD-1/M9 are new designs; sold in puny volumes; built with labor that costs radically more.<br>

With a low volume non mature product one has to "get back" the costs to design/tool/build the gizmo; its "alot per unit" since there will never be alot of sales volumes. The M9 buyer is less tolerant of a micro scratch; blem; cosmetic error/flaw than the 17 buck Harbor Freight grinder buyer.<br>

One chap is concerned with looks over function; another is worried about sharpening his lawn mower's blade. The bitching, whining, warranty costs to hand hold M9 buyers flaws will be a higher percentage of the units sales price; than a dumb grinder.</p>

[les_lammers]

<p>It is impossible. The world is flat too. In 1974 I borrowed $100 from the Navy Credit Union to get a TI calculator that did little more than a $2 unit today. Technology..ain't it great?<br>

To all of you getting a M9...enjoy!</p>

[g-man1]

If they had told us it was possible then fewer people would've bought the M8

[stephen_asprey2]

<p>With purchasing digital, one has to put a stake in the ground at a point of time. Then buy what's available now to do the job. And tomorrow technology will have moved a bit. One has to learn to be satisfied. With film, technology changes occurred about every 5-10 years. With digital its about 2-3 years. Moaning about it won't change it.</p>

<p>Gone now are the days where new technology also benefits past purchasing decisions. The saving grace is that at least some camera manufacturers keep their lens mount the same for decades so that what fits a new camera today will also fit older ones. That at least is some measure of investment protection.</p>

[joseph_wisniewski]

<blockquote>

<p>VC 15mm lens. The 16mm is a retrofocus optical design to clear the Nikon slr's mirror; the 15mm is not.</p>

</blockquote>

<p>The 15mm VC is retrofocus with about a 28mm exit pupil.</p>

[kelly_flanigan1]

<p>Joseph; when I got my 15mm VC lens 9 years ago it was called a non retrofocus design. Camera Quest calls both the 12mm adn 15mm VC lenses a non retrofocus len design.<br>

<br /> What is your references that say that the 15mm design is now changed?<br /> <br /> Steven Gandy also markets the 12mm and 15mm VC versions for the Nikon F series; one has to use mirror lockup. He describes botht of these a 12mm and 15mm lenses as a non retrofocus lens design too.</p>