Retrofocus vs Gauss designs? (On "standard" focal lenghts)

<p>Hi, all. I`m wondering about the diferences between this designs. I know wide angle lenses use to favour the assimetrical Retrofocus design to save the mirrors, and Gauss designs use to be common from standard lenses to moderately long ones, but... what about the differences of both designs in "standard" focal lenghts?<br /> Mamiya offer both designs in their "standard" range of lenses (90, 110, 127, in different versions). As far as I know, the main differences could be:<br /> <strong>Advantages:</strong><br /> <em>Retrofocus:</em></p>

<ul>

<li>Even illumination across the frame</li>

</ul>

<p><em>Gauss:</em></p>

<ul>

<li>Higher center sharpness</li>

<li>Better aberration correction</li>

<li>Higher lens speed</li>

<li>Better close to infinity distance performance without floating elements</li>

</ul>

<p><strong>Disadvantages:</strong><br /> <em>Retrofocus:</em></p>

<ul>

<li>More pronounced field curvature</li>

<li>Higher distortion</li>

<li>Poor performance along the distance range</li>

</ul>

<p><em>Gauss:</em></p>

<ul>

<li>Vignetting</li>

</ul>

<p>---</p>

<ul>

</ul>

<p>Am I missing something? If not, the Gauss looks to be a clear winner. I wonder about the usefulness of the retrofocus design for lenses that barely are moderate wide angles, but has been classics in the Mamiya lineup and continuously updated. <br /> I also wonder if <strong>Hasselblad</strong> has similar options (like that 90 Retrofocus and 110 Gauss Mamiyas).<br /> Why they still list retrofocus standard lenses? Is the focal lenght the only attractive? What do you think?</p>

<p>Retrofocus has the light more normal/perpendicular to the sensor with digital; thus less weird corner issues.</p>

<p><br /> Some sensors prefer light to be Normal (high noon) that glancing ( ie sunset)</p>

<p>It makes sense but... one of them, the KL90/3.5L is a "recently" updated 90 (and improved) for the RB system (!). I`m intrigued...</p>

<blockquote>

<p>Why they still list retrofocus standard lenses</p>

</blockquote>

<p>No doubt because even "normal" focal length lenses can still be harder to design in such a way as to make sure of clearing a larger mirror. This gets tangled up with camera design, the effort to get closer to 100% size in the viewfinder, lens aperture, and many other factors. Lens design <em>per se</em> is not the only variable(s) that have to be looked at. Lens design is not altogether "pure" theory where one design can be said to be "better" in some universal sense</p>

<p>Note that a lens like the early "normal" "fast" Biotar (a double Gauss design) was built at a <strong>58mm</strong> focal length to achieve an f/2 aperture (very fast for the day) on the Contax S.</p>

<p>Even illumination depends on exit pupil position.<br>

Retro focus lenses have their exit pupils at a distance from the film greater than their focal length. But not greater than the exit pupil to film distance of double Gauss type lenses.<br>

So no advantage there.</p>

<p>Retro focus designs are used to create enough clearance between lens and film to allow putting a mirror between the two.<br>

As JDM says, the focal length of standard lenses is such that it is hard to make one that leaves enough room for the mirror and still be symmetrical. So though many standard lenses are of double Gauss design, many are a bit asymmetrical too, just because of that problem.<br>

They still are of a design that allows very good correction of most lens faults.</p>

<p>In practice, "retrofocus" design is a matter of degree rather than kind. The mounting flange on an Hasselblad is about 79mm from the film plane. Consequently any lens with a nominal focal length less than about 120mm is likely to be a retrofocus lens to some degree. Basically, the objective of a retrofocus lens has the effect of a negative element. In extreme cases, the exit pupil can lie completely outside the lens, between the last element and the focal plane. The 80mm Planar is right on the edge in this regard.</p>

<p>The reason is simple. The mirror must have room to swing upward at the moment of exposure. There is very little clearance behind the mounting flange for this to occur.</p>

<p>When the Planar is used in a Rolleiflex TLR, this clearance is not needed. Consequently the lense is smaller by about 1/3rd, shorter (fits deeper in the camera body) and, some say, sharper than its brethren designed for SLR cameras. All I can say is that the Rolleiflex version is very sharp, but so is the Hasselblad.</p>

<p>A similar design is nearly always used for lenses longer than 120mm, but in the opposite direction. A negative element is placed near the exit of the lens in order to move the exit pupil forward. This allows the lens to be shorter, consequently easier to handle and possibly lighter.</p>

<p>Lenses shorter than 80mm have an increasingly powerful retrofocus design. That said, the performance of Distagon lenses designed for 35mm cameras, specifically the ZF lenses for Nikon, perform much better than Nikon lenses about 1/2 the physical size (and cost 3x as much).</p>

 

<p>Thank you all for your comments... I`m afraid I was a bit dense this afternoon. Coming back home I recalled that the original "standard" Nikkor for the Nikon F was a "retro-modified" Gauss in 50mm. Faster ones started also at 58mm. I have checked that the flange distance on the Mamiyas is proportionally larger, it looks like, as you say, there is a technical limitation to made non Retrofocus 90mm lenses for Mamiya (specially for RB models which have an even larger flange).</p>

<p>Then, I wonder if there is any special attractive for that Retrofocus "standard" lenses, other that their focal lenght. Maybe what Kelly says, theoretically it could be a better performer on digital, but probably all are similar performers will all kind of corrections.</p>

<p>BTW, Nikon flange is 46,5mm > frame diameter is aprox. 43mm > shorter Gauss is 50mm<br /> RZ flange is 104mm > frame diameter is 87mm > shorter Gauss is 110mm<br /> RB flange is 111mm > frame diameter (I suspect) is 87mm > shorter Gauss is 127mm</p>

<p>I think the retrofocus design has been used pretty much for its obvious purpose - moving the glass forward to clear obsructions - and perhaps a bit more recently to give a more perpendicular ray path to digital sensors. Retrofocus "normal" lenses began as a means of getting a fast lens of less than 58mm onto an SLR; before retrofocus, you could get a 50mm Tessar up to f/2.8 or an f/2 lens at 58mm. I believe the first 50mm SLR lens of f/2 or faster for a 35mm SLR was made by Angenieux, who also introduced the retrofocus wide angle (and coined the word "Retrofocus" as a trade name). I have a 50mm f/1.5 Angenieux in M42 mount that dates from around 1953 .... it would be another decade before the rest of the world caught up with that.</p>

<p>If the lens is far enough forward in a straight Gauss formula, I can't think of any particular reason to build it as a retrofocus.</p>

<p>The Mamiya RB and RZ that you are asking about are "extreme" cases for MF SLRs, because of their rotating backs. So the "6x7 cm" RZ is really a 7x7 cm format body (in terms of its focusing screen size, mirror size, and longer flange distance to accomodate the mirror), while the 6x7/6x8 cm RB is really approaching an 8x8cm body.</p>

<p>Although they are also 6x7 cm SLRs, the Pentax 67 and Bronica GS1 only have to illuminate a 56 mm high frame, so their screen, mirror and flange distance can all be considerably shorter. This explains why the Pentax 90mm f2.8 lens can be an essentially double-Gauss design, while the same focal length on the RB67 (90mm f3.5 KL) must have a strong degree of retrofocus design.</p>

<p>There is one other interesting observation. Retrofocus design and faster apertures both increase off-axis aberrations, which is why you will rarely find the two in concert (unless aspherical elements are employed, as in some fast 35mm-format wideangles). To make their fastest standard lenses, MF SLR manufacturers went for slightly "longer than normal" (longer than the film diagonal) focal lengths and thus avoided retrofocus designs. Where two different standard lenses are offered, the faster one is always the longer focal length and the less retrofocus in design. Hence:<br>

Mamiya RZ67 - 90mm diagonal - 90/3.5 and 110/2.8 lenses<br>

Pentax 67 - 90mm diagonal - 90/2.8 and 105/2.4 lenses<br>

Hasselblad F - 80mm diagonal - 80/2.8 and 110/2 lenses<br>

Mamiya 645 - 70mm diagonal - 70/2.8 [and 80/2.8] and 80/1.9 lenses<br>

Contax 645 - 70mm diagonal - 80/2 lens</p>

<p>So why does the leaf-shutter Hasselblad break this "rule", with an 80/2.8 and a slower 100/3.5? That is more to do with the constraints of a smaller (size 0) leaf shutter within the lens. Of the cameras above, only the RB/RZ have leaf shutters, and those are larger (size 1) shutters. It was possible to squeeze a 110/2.8 into a size 1 shutter for the RZ67, but evidently a 127/2.8 was not possible for the RB67, so we just have a 127/3.5.</p>

<p>Quote: "So why does the leaf-shutter Hasselblad break this "rule", with an 80/2.8 and a slower 100/3.5? That is more to do with the constraints of a smaller (size 0) leaf shutter within the lens."<br>

<br /><br>

That makes sense, and I'm sure it's right. Also, the 100mm Planar is said to be intended for "scientific applications. " In my casual, mostly hand-held work, I have not noticed (yet) any difference in sharpness or distortion between my 80mm and 100mm Planar. But probably the 100 Planar's smaller aperture allowed the designers additional freedom in maximizing quality.<br>

<br /><br>

I imagine that the need to avoid the retrofocus design, with its inherent distortion, explains the choice of 100mm over 80mm for a lens of such high quality. There is an old thread in the archives in which I had a dialog with Dr. Fleischer, then head of lens applications at Zeiss. Fleischer said that the 80mm Planar is "slightly" or "mildly" retrofocus. I believe the 100mm Planar is meant as a near-perfect alternative normal lens for applications that require it.</p>

<p>Dr. Hans Sauer (the man who designed the lens) once explained that the 100 mm Planar was to be designed free from any constraints that things like lens flange to film distance impose.<br />A thing (lens flange to film distance) that did indeed make the 80 mm a bit 'retrofocus', and still needed a too short mirror in the earlier Hasselblads.</p>

<p>Interestingly though, the resulting lens, the f/3.5 100 mm Planar, also is a bit retrofocus.</p>

<p>The thing is not as much whether a lens is retrofocus as such, but what they had to do to make a lens retrofocus.<br />The Planars, 80 mm and 100 mm alike, do sit a bit further away from the film than the focal length would suggest, and as such could be called retrofocus lenses. But they are still of a design that allows very good correction of just about all lens faults.<br />For shorter focal lengths, getting the lens to sit far away enough from the film to allow reflex viewing demands an entirely different design. A design that makes it much harder to correct all lens faults.</p>

<p>The term 'retrofocus' should perhaps be reserved for lenses that need to be of a 'special' design to move them away from the film enough.<br />And not be used for lenses that are a bit assymetrical (and which lens isn't?) in a way that happens to move them a bit away from the film plane.</p>

<p><em>In my casual, mostly hand-held work, I have not noticed (yet) any difference in sharpness or distortion between my 80mm and 100mm Planar. </em></p>

<p>If you compare pictures taken without a tripod, you are unlikely to find an Hasselblad any better than a Kodak Hawkeye.</p>

<p>And that doesn't have to be bad. See what some people can do with a Hawkeye (<a href="http://westfordcomp.com/classics/hawkeye/index.html">link</a>)</p>

<p>Very, very interesting observations. Things are much clearer to me now. It makes me think that Mamiyas certainly have "modern", generous dimensions although still tight in its design (specially the RZ).<br /> I don`t know about Hasselblads but also wondered about the history of the current sliding mirror mechanism. I need to have a look on this.<br /> As I`m a bit paranoid about photography, I now feel the need of testing by myself both lens designs (say, 80&100 Planars, 90&110 Sekors). I`ll try to cool this momentum a bit. I appreciate so much your comments and ideas.</p>

<p>In a nutshell:<br />- A first version of a mechanism that allowed use of a large enough mirrror appeared in the 2000 FC in 1977. All later 2000-series camera of course also had that mirror mechanism.<br />- A second was used in the 500 ELX and 553 ELXmodels, first introduced in 1984.<br />- A third in 1991 in the first of the 200-series, the 205 TCC (and then in all other 200-series models too).<br />- Then it appeared in the 500 C(...) line, first in the 503 CW in 1996. (So despite the great attention that was given to this "sliding mirror', it arrived late, almost 20 years after the first larger mirror mechanism was introduced). Later also in the 501 CM, in 1997.<br />- The 555 ELD of 1998 perhaps has the sturdiest version of the mechanism.</p>

<p>Not all of the versions of the mechanism that move a larger mirror back before/while it swings up are the same. In fact, i don't think any of the versions identified above are, but i have to check that.<br />But though they perhaps not all <em>are</em> the same, they all <em>do</em> the same.</p>

<p>"If you compare pictures taken without a tripod, you are unlikely to find an Hasselblad any better than a Kodak Hawkeye."</p>

<p>@ Edward Ingold: My remarks were intended to contribute useful information, not to invite irresponsible wisecracks from those who have nothing of value to say. As a long-time contributor to this forum, I prefer that it be frequented only by those who have something of value to contribute, and to those sincerely seeking information. I don't expect everyone to be perfect or 100% accurate. We all get things wrong occasionally. I do expect sincerity.</p>

<p>@ Q.G.: Thanks for adding this fresh info! I never knew the designer's name. And thanks for explaining that the effect of a retrofocus design vs. non-retrofocus on performance is a matter of degrees, not just a bad/good distinction. So much has been learned by the engineers in recent years! My 40 and 50mm Distagon CF FLE lenses give me excellent results, with nothing at all to complain about. Even my 60mm Distagon CT* is very good, giving me no excuse to trade for a more current version.</p>

<blockquote>

<p>As a long-time contributor to this forum, I prefer that it be frequented only by those who have something of value to contribute, and to those sincerely seeking information.</p>

</blockquote>

<p>That's nice that <em>you</em> prefer that, but the forums, while moderated, are forums still. One feature of the beast is freedom of speech and the right to post even wise-cracks if you please. Nor is humor "insincere" when addressing the issue of lens comparisons for hand-held pictures. Because some one doesn't agree with a point does not mean it has no "value."</p>

<p>Sure, there's room for humor, and lots of room for contrasting opinions. I just felt that the remark about a Kodak Hawkeye (I had one when I was about nine; wish I still had it) was not only unrealistically exaggerated, but also intended as a condescending put-down. The poster might have said that the slight unsharpness caused by hand-holding could be enough to take the fine edge off the 100mm Planar. I knew that, of course: that's why I pointed out that some of my shots are hand-held. Others are taken using my Bogen/Manfrotto 3221 tripod with a heavy ball head. Of course, those are at least a little better! But as I get older and realize I don't know how much time I have left, I no longer look at negatives with a microscope. I prefer to spend my time shooting and printing.</p>

<p>This is only a hobby for some of us. We shouldn't take these posts too seriously. At the same time, we should be respectful, and we should pause before writing, to question whether what we are about to write is accurate and helpful. Sarcastic, exaggerated, hyperbolic, or condescending remarks are unhelpful and not commensurate with the goals of the forum.</p>

<p> </p>

<p>Q.G. (Quinten?), your info is time saver to me. Thank you very much. Now I wonder about the mirror in the original designs; I suspect it was not different than other cameras, where the screen appear darkened if used with longer lenses. Looks like it was not the main designers issue in the past. But that`s certainly another topic.<br /> Ray, you are right, that`s exactly what I was lucubrating about. Your comments are so clarifying.<br>

Thank you all very much!</p>

<p>I don't know whether it was a general issue.<br>

But on the Hasselblad, a large enough mirror would hit the rear of the 80 mm lens when that was set to infinity.<br>

The earlier 1000-series Hasselblads did not have that issue, because they had a longer flange to film distance, so more room for a larger mirror. (I don't know why they shortened it for the 500-series.)</p>

<p>The too short mirror is not a problem when the exit pupil of the lens that is put on the camera is close enough to the mirror. The mirror will catch all the light coming from the lens and deflect it up to the focussing screen.<br>

But when the distance increases, part of the light eminating from the rear of the lens will find its way past the lower edge of the mirror, creating the vignetted top of the viewfinder image.</p>

<p>This, by the way, doesn't depend on the focal length as such but again on the design of the lens.<br>

Which is exemplified by the 120 mm Makro-Planar showing more viewfinde vignetting than the longer 150 mm Sonnar. The exit pupil position shows why: 135.3 mm for the 120 mm Makro-Planar, but only 112.5 mm for the 150 mm Sonnar.<br>

The exit pupil position of the even shorter 100 mm Planar is 115.8 mm, so even that lens will show vignetting sooner than the 150 mm.</p>

<p>Perhaps other brands had to content with this phenomenon causing viewfinder vignetting with longer lenses too. I really don't know.<br>

But what i do know is that the focal length of the Fresnel lens, the thing that is used to get an evenly illuminated focusing screen, has to match (roughly) the exit pupil position of the lens. So you would need a different focal length Fresnel lens with longer lenses, else the corners darken.<br>

I suspect that this is more likely to be what is behind the darkened screens with longer lenses thing you mentioned. Not a too short mirror.</p>

 

<p>No, it`s exactly what you mentioned above:<br /> <em>"... part of the light eminating from the rear of the lens will find its way past the lower edge of the mirror, creating the vignetted top of the viewfinder image."</em><br /> Actually the effect is that defined darkened strip.<br />As you say, I understand that the fresnel cutting angles should be made to match that pupil positions for optimal performance, something that doesn`t use to happen, I guess. It`s more likely "one for all".<br /> BTW, Hasselblad certainly has a very interesting documented background, something I enjoy sooo much.</p>

<blockquote>

<p>Perhaps other brands had to content with this phenomenon causing viewfinder vignetting with longer lenses too. I really don't know.</p>

</blockquote>

<p>I can confirm that it was also an issue with a Kiev/Arax 60 I had, and the longer lenses - 300mm Sonnar and 500mm Prakticar.</p>

 

<blockquote>

<p>But what i do know is that the focal length of the Fresnel lens, the thing that is used to get an evenly illuminated focusing screen, has to match (roughly) the exit pupil position of the lens. So you would need a different focal length Fresnel lens with longer lenses, else the corners darken.</p>

</blockquote>

<p>Q.G., cheers! You have explained something that I'd noticed recently on my Mamiya 645 AFD. The screen corners do indeed darken by about a stop when I have the 200mm APO mounted, whereas the screen is uniformly bright with shorter lenses like the 80mm. I wondered about this, since it was still visible on the screen even when the lens was stopped down a bit and no falloff would be visible on the recorded image. Another thing: I just checked and I cannot really see this effect on my older M645 1000s; it must have a differently optimized fresnel.</p>

<p>Man, it was nice to look through my M645 1000s again. The world just looks better through a waist level finder.</p>

<p> </p>

<p>@Edward Ingold<br>

<em> "If you compare pictures taken without a tripod, you are unlikely to find an Hasselblad any better than a Kodak Hawkeye."</em><br>

Edward, such a flippant response really doesn't help newer users get the most out of their equipment. Yes there can (as in CAN) be gains by using a good support system, but like anything else in life, it's a give and take. On the one hand a good tripod can drastically help, a less than adequate one can harm. Some cameras (the RB67 in particular) have an extremely well dampened mirror, having been intended to be used handheld.<em><br /></em></p>

<p>There's a fundamental contradiction for me in this discussion. How can I criticize others for presuming to dictate the content of forum responses without being guilty of the same sin myself?</p>

<p>This is not the beginner's forum for "newer users", and I am not even sure that Edward's comment would have to be moderated 'out' even if it were. It's really a pretty inoffensive little overstatement, perhaps. Frankly, I think some persons who take so much offense at it, might profit from taking a shot or two with a nice <a href="http://mattdentonphoto.com/cameras/kodak_hawkeye.html">Brownie Hawkeye</a> simple meniscus lens.</p>

<p>I, as a longtime contributor to this and other forums, find posts trying to dictate the direction of other people's discussion to be offensive, <strong><em>including this one</em></strong>.</p>

<p>And I really don't give a damn whether people hand hold their medium format cameras or not, but the issue of sharpness or lack of it from camera motion is a legitimate question.</p>

<p>JDM; Edward's comment above is not one that would normally be edited out by Moderators here at PN. Opinions are legitimate even if we feel they are sometimes wrong. Personal attacks are another matter. Quid pro quo exchanges filled with venom are not legitimate.</p>