Digital MF SLR or Digital Back on View Camera

[q.g._de_bakker]

If lenses aren't good enough to provide fine enough detail for 80 MP backs, they are not good enough to provide fine enough detail for 22 MP backs that move a tiny bit to produce 88 MP.<br>The difference you could be seeing is not, cannot be, due to lenses.<br>Lenses are not incapable of dealing with ultra high resolution sensors at all. Sensors are still trying to catch up with what lenses are able to deliver.

[theodoros_fotometria]

<p>What you are saying is totally wrong O.G.... The 88mps of 22mp backs when shot in 16mp mode, <em>are exactly because each one of the shots made are of only 22mp (</em><em>i.e. require only 22mps) analysis of the sensor... the 88mp outcome is one of true 88mps but it's an (real) outcome of 16 shots (21 actually) that each one of them only records 22mps... </em>it's not interpolation either..., it's a mechanical process that:<br>

A. Records even in the space between pixels...<br>

B. Splits a pixel into four smaller ones that each one records different information... <em>IN TRUE COLOR (no Bayer pattern involved). </em><br>

This is all because a there is a mechanism in the camera (i.e. the MFDB) based on a piezzo electric sensor, which <em>moves each pixel by half its side length in four directions,</em> but still each shot is one of 22mps... and thus the requirement from the lens is to cope with that (the 22mp) analysis. Regards, Theodoros http://www.fotometria.gr</p>

[q.g._de_bakker]

Theodorus,<br>I know how this works. But have you thought about it? Have you thought about what you say too?<br>So think about it. What do these sensors record, if not detail in the image the lens projects on them? <i>"If lenses aren't good enough to provide fine enough detail for 80 MP backs, they are not good enough to provide fine enough detail for 22 MP backs that move a tiny bit to produce 88 MP."</i><br>Or in other words: <i>"[...] lenses (even the best ones) are incapable to cope with ultra high res sensors... if they where capable, then 80mp backs would be able to show the same detail as 16X done on an old 22mp sensor (=88mp)... which is far, far, ....too far from truth! "</i>, is nonsense. If there is a difference, it cannot be due to lenses.<br>Just think about it a bit.

[theodoros_fotometria]

<p>O.G.... I am not in a mood for an argument... and it's not helpful to Lobalodo's requirements, what I say, I have discussed with the makers of the (MS) backs themselves... so I suggest you think about it... Theodoros.</p>

[q.g._de_bakker]

You do not have to be in the mood for an argument. You just need to think about the things you say. You may have discussed them with the makers of backs, and if you said the same thing to them as you did say here, they too will have scratched their heads and politely asked you to think about what you say. Because it is incorrect, does not make any sense.<br>I'll explain it again: if a lens does not deliver enough information to capture, the way you capture what it does offer cannot exceed what is on offer to begin with. <i>"If lenses aren't good enough to provide fine enough detail for 80 MP backs, they are not good enough to provide fine enough detail for 22 MP backs that move a tiny bit to produce 88 MP."</i> where would that extra information the multishot backs capture, but cannot be delivered to one-shot backs by these lenses, come from?<br>It's not, cannot be, the lens, Theodorus.<br>Do think about it.

[theodoros_fotometria]

<p>What you miss is that because the lens is still.... and the image area is (almost) constant ...it doesn't mean that the same res will be captured per unit of image area, <em>because the image area moves in four directions!</em> (this means that the lens doesn't record the same information on the image area as it would at a single shot with an 80mp back... think about it! I am sorry, I won't reply back if you can't get it...<br>

If sensors where trying to catch up with lens resolution as you suggested earlier, then you would be able to see 4 times the resolution of a single shot 22mp sensor if using an 80mp back... which is far from truth... while with 16X <em>that's exactly what you get..., 4 times the resolution of a 22mp "fat pixel" MF sensor... </em>think about it.<em><br /></em></p>

[q.g._de_bakker]

How many times does it have to be said? Just stop and think, instead of continuing this nonsense.<br>A lens does not record information.<br><i>"If lenses aren't good enough to provide fine enough detail for 80 MP backs, they are not good enough to provide fine enough detail for 22 MP backs that move a tiny bit to produce 88 MP."</i> No matter how you choose to record what a lens projects, if it isn't there it's not there. If you manage to capture it anyway using MS, it has to be there to begin with. Easy enough, isn't it?<br><br>So, again, your <i>"discussion on MS"</i> does not <i>"prove that lenses (even the best ones) are incapable to cope with ultra high res sensors."</i> It proves that you probably haven't thought about, most certainly haven't understood, how things work.

[fotografz]

<p>Hi Lobalobo,</p>

<p>Regarding your questions about differences in backs ... and about precision in order to get the best out of what-ever you decide on getting.</p>

<p>The back differences are an endless debate depending on who you talk to. I agree with Theodoros that in many cases the processing software and how skilled you are using it plays a pivotal role. Heck, for less critical work, I even processed H4D/60 shots in Lightroom because Hasselblad provided Adobe with all the profiles and DAC lens corrections ... I do this when I may be shooting with 2 or 3 different cameras and want it all in order in a LR catalog so I can synergistically process the files. Phocus still produces better color rendition, and is a must have for fast and sure tethered work. I've never gotten along with C1 as many times as I have tried ... but one opinion shouldn't preclude it at all, since it is a powerful and well featured program that gets better with every issue ... and many swear by it.</p>

<p>As to the backs themselves, it depends on what is compared to what. As far as IQ is concerned they are all good. The 39 meg backs from Hasselblad and Phase One use the same Kodak sensor, where the Aptus 33 meg backs use a Dalsa sensor that some prefer for its (debatable) slightly more film like image qualities. </p>

<p>How the different backs operate differs: the Leaf Aptus is touch screen, the others are more analog buttons. Only you can determine which you like. The key difference with the Phase backs is that the battery is integrated into the back, and a Aptus, Hasselblad CF39 or CFV backs, require a battery be clipped on the bottom. The backs from any integrated Hasselblad camera (H2D, H3D or H3D-II, H4D) have no provision for an on-back power source once removed from the H camera ... requiring either computer tethered operation, or tethered use of the Hasselblad Image Bank-II, which also powers the backs up to the H4D (but excluding the H4D/60). For the more streamlined battery reason alone, the Phase One P45+ would be the better choice for any field/tech camera IMHO. Without a doubt, the later Phase One backs are superior in operation for your applications like architectural or landscape work with camera movements. </p>

<p><strong>Precision:</strong> that is what I meant by having a "tinkerer's personality" when it comes to the tech cameras. It can and often does require precise shimming of the back in relation to the camera body. I recall an in-depth tutorial from Alpa (or one of their dealers) showing how to do it ... and frankly, it would try my patience. I once had to shim a Schneider 120 APO macro digital lens because, while all my other view lenses were dead on, I couldn't achieve critial focus with the 120 to save my life ... my tech savvy dealer came by and took the lens apart and began removing wafer thin shims inside ... while going back and forth using the Hasslblad audio feed back focus feature in the Phocus software, until we got it just right. He has a LOT more patience than I do. However, it IS a different hands-on tactile experience that some actually enjoy, and IMO when you get it all dead on, nothing can beat it.</p>

<p>Using a back on a reflex camera introduces all sorts of debates ... some earlier backs on earlier cameras often required shimming. If it was off, the digital back maker would blame the camera company, and the camera company would blame the back maker ... sounds funny now, wasn't funny then. However, it does seem that using one back on multiple camera bodies requires incredibly precise manufacturing tolerances for both elements. Yet, my Aptus 75s actually wasn't all bad on the Mamiya AF-III. Unfortunately, I simply never got on with that camera in any way positive ... in fact I hated it and all of the lenses compared to the Contax 645 that I previously used. I then just used the Aptus on the RZ which I did like a lot despite its elephantine demeanor.</p>

<p>The much debated move by Hasselblad to integrate the H camera (and close it to other backs) was in part to eliminate the variables. Each camera was/is individually calibrated at the factory and all the settings stored for that serial number ... and as Hasselblad added features like true color, micro focus adjustments as you stopped down to avoid focus shift, DAC feed-back, then True Focus, that integration seemed to pay off ... but, whether it was really necessary is highly questionable and beyond my scope of understanding. All I know is that after 9 different H cameras over the years, all of them were precise right out of the box.</p>

<p>Hope this helps a bit,</p>

<p>-Marc</p>

<p> </p>

[lobalobo]

<p>Thanks so much, Marc. Your answers have been responsive and clear and very helpful. One last question (and I apologize if I've already asked this and forgotten): If the precision of the back position is not an advantage of, say, a shimmed Alpa camera, over an integrated system such as the H4 or H5, do you think that the lenses available for a technical camera like the Alpa would produce a superior image compared to the H lenses? I know that the range and precision of movements available on the Alpa compared to the (very expensive H tilt-shift adapter) are a plus, as is the ability to create movements without a tele factor (using short-barreled lenses), but I wonder whether there is also an IQ advantage. (Luminous Landscape articles say yes, but I've grown skeptical of these.) Thanks again.</p>

[fotografz]

<p>Lobalobo, </p>

<p>Frankly, I simply do not have direct experience with the more recent Schneider and Rodenstock lenses. This is why I suggested GetDpi forum as an additional resource. Very nice and knowledgable people there ... a good amount of tech camera users with direct experience with most all of the lenses and cameras. </p>

<p>From what I can gather, the consensus is that there is a definite IQ advantage.</p>

<p>That said, two of my favorite landscape shooters use regular MFD reflex cameras and system lenses ... which is saying a lot, because I'm not even slightly interested in landscape shooting unless there are people in the frame ... then it becomes environmental portraiture : -)</p>

[lobalobo]

<p>Great; thanks to all.</p>

 

[joseph_wisniewski]

<p>Awesome. Theodoros Fotometria has decided to subject photo.net to the same sort of behavior that got him banned three times from dpReview.</p>

<p>I'll go make the popcorn.</p>

[ondebanks]

<blockquote>

<p>Awesome. Theodoros Fotometria has decided to subject photo.net to the same sort of behavior that got him banned three times from dpReview.</p>

</blockquote>

<p>Isn't getting yourself banned from dpreview - of all places! - a bit like Satan kicking you out of hell for just being <em>too </em>evil? :)</p>

<p>Well anyway, Joseph, thanks to your post pointing out that Theodoros 'has form', I just remembered who he is and where I (and a few others) previously ran up against him. It was this LuLa thread: http://www.luminous-landscape.com/forum/index.php?topic=58085.0 My handle is <em>ondebanks</em>, his is <em>fotometria gr. </em>It's a good read, especially if you can stay with it to page 3. My favourite quote from him in that thread was "<strong>I...don't care much to analyze the technical origin of the solution given from manufacturers nor the scientific theory behind it for that matter</strong>". So while asserting on the one hand that he doesn't care about the science, he was adamant on the other hand that our evidence-supported science was wrong!<br>

<em><br /></em>Theodoros, here in this thread you are (again) digging a hole because of clinging to a basic technical error...Q.G. is quite correct in rebutting your assertions about 88MP multi-sampled vs. 80MP single-shot. It's clear that you are going <em>way</em> beyond the mild advantage touted by manufacturers of not requiring de-Bayering interpolation with the multi-shot (rendered even milder by the sensor size differences - 48x36 mm vs 54x40 mm), and that you don't understand how shifted and static sensors interact with the focal plane of the lens. Withdraw with some grace and do some research on point-spread functions (PSF), pixel response functions (PRF), and sampling. Check out the concept of "drizzling" with the Hubble Space Telescope dithered-mode images (analogous to your multi-shot): http://www.adass.org/adass/proceedings/adass99/O6-02/<br>

I happened to be at that ADASS conference in 1999 and I spoke to Richard Hook about that paper he presented with Andy Fruchter. Pay particular attention to where they say "Finally, as with all linear reconstruction methods, drizzling makes no attempt to reduce the loss of resolution resulting from the convolution with either the PSF or the PRF." The multi-sampling DBs uses a linear reconstruction method to go from 22MP to 88MP or whatever. Therefore the multi-sampled image still has resolution losses due to the PRF. It can't magic them away. The smaller pixels of the 80MP one-shot actually have the advantage of a correspondingly smaller PRF, but that advantage is eroded by the de-Bayering interpolation.</p>

<p>Let's go to a reductio ad absurdem - always a great way to explore whether a hotly debated contention makes sense. If, as you say, multisampling can take a 22MP sensor past the performance of an 80MP sensor (of similar dimensions), why stop there? Why not develop multisampling sensors with the same dimensions, but far fewer and far larger pixels than that? A 1 MP sensor with 42 micron pixels, and hundreds of sampling shift positions? A single-pixel sensor with 42 mm pixels, and millions of sampling shift positions? Do you still think that you will get ~80MP from such approaches? The PRF says "no".</p>

[bradleygibson]

<p>Lobalobo,<br /> Great question. I currently own a MF digital system, and recently purchased an Arca/Swiss M-Line Two to put in front of my digital back.</p>

<p>The verdict?</p>

<p>There is no contest between the IQ of the Rodenstock Sironar Digital lenses I have and the Schneider and Zeiss glass I own and have owned in medium format. The Rodenstock IQ is markedly superior. (Note that I've owned most digital brands of medium format--Hasselblad V, H, Contax, Phamiya and Rollei--a good, digital-tolerance-capable tech/view camera with the right lenses are defifnitely a step up in IQ from any of these.)</p>

<p><br /> On the other hand, working with the view camera is very, very slow, by comparison. If you don't mind the workflow, the decision is easy. But do consider the inconvenience of the workflow and your personal style.</p>

<p>I have written a bit about my experiences and provided raw files in this posting: http://www.getdpi.com/forum/medium-format-systems-digital-backs/45161-s-m2-rodenstock-sironar-digital-afi-ii-10-%3D-holy-%40.html<br /><br />I hope that is helpful!<br /> -bRad</p>

<p>P.S. A P45+ on an Alpa would be a fine combination. Enjoy!</p>

[lobalobo]

<p>Thanks, Bradley. I don't mind slow shooting; do it now with 4x5 film and a focusing screen. And I'm not sure that I'd often want to hand-hold a medium-format camera, which would, I take it, remove much of the advantage of the large sensor. So the question becomes whether there would be times that I'd have a camera on a tripod and still want to shoot quickly; I could see a fashion pro needing this feature, but as an amateur shooting landscape and architecture mostly, I don't think this will come up very often. And for the money I'd spend on a digital back and view/technical system such as the Alpa, it would not be much more to buy a used Hasselblad V camera and a couple of used V lenses for the few occasions I would want to shoot more quickly; no auto-focus in this plan, and I wouldn't be able to match the best lenses to the quicker shooting, but it won't come up that much that I'd want or need to do this anyway. So your answer has been truly helpful. Now I'll just have to see whether my wife will be merely shocked, or threaten to kill me, when I mention what I plan to test and maybe buy.</p>

[mag_miksch]

<p><em>"I'd go w/ the sports car</em>."<br>

+1^^<br /></p>

[joseph_wisniewski]

<p>Ray, yeah, that's classic Theodoros.<br>

He's dropping into classic behavior here, too, I see. There's another thread that he hijacked to beat his favorite dad horse, the "demand" for a D700 style body with a D4 sensor.</p>

<blockquote>

<p>a bit like Satan kicking you out of hell for just being <em>too </em>evil?</p>

</blockquote>

<p>Well, Theo sure doesn't live up to his Holy name. I've seen him lie outright (thou shall not bear false witness, lol) and swear. He's prone to randomly attacking anyone, and even attacks people who agree with him, for not agreeing "properly". </p>