Poor scans - May explain why some switch to Digital

[StuartMoxham]

<p>Yep the blue rivers thing is quite strange.</p>

[rishij]

<p>Mark Smith: that's one of the softest 35mm scans I've ever seen in my entire life. Not at all representative.</p>

[Landrum Kelly]

<p>Oval smokestacks: There might be software solutions to such distortions. Maybe Scott will know.</p>

<p>--Lannie</p>

[mauro_franic]

<p>Stuart, I agree with your observations, the D2X produces nice prints indeed. The point was that even at 18" film (even 35mm) has a visible edge in quality. No need to think extravagant sizes are needed to see the difference.</p>

[Landrum Kelly]

<blockquote>

<p>"Mark Smith: that's one of the softest 35mm scans I've ever seen in my entire life. Not at all representative." --Rishi Sanyal</p>

 

</blockquote>

<p>Rishi, is that the result of a bad scan, or just a blown shot?</p>

<p>--Lannie</p>

[rishij]

<p>Here's a quick quiz for ya folks... here's a 10 megapixel image I found:<br>

 

<a href="http://staff.washington.edu/rjsanyal/Photography/FairyFalls_10MP.jpg">Link to Full-Size Image</a></p>

<p>Follow the link to the full-size image or download it. View it at 1:1.</p>

<p>Tell me what format you think it is. dSLR? 35mm? MF? Let's keep this objective.<br>

-Rishi</p>

[Landrum Kelly]

<blockquote>

<p>"Stuart, I agree with your observations, the D2X produces nice prints indeed. The point was that even at 18" film (even 35mm) has a visible edge in quality. No need to think extravagant sizes are needed to see the difference." --Mauro Franic</p>

</blockquote>

<p>Mauro, if you are talking about the first shots on the thread, wasn't that made with 6 x 7, not 35mm?</p>

<p>--Lannie</p>

[mauro_franic]

<p>David,<br>

35mm is not as good as 645 (don't know who said that). A 5DII is (to me) in the 35mm film range not 6x7 film (which is almost 5 times bigger). </p>

<p>I think you may be trying to convince yourself otherwise or just being disguised becaused you are limited to a 13" printer. </p>

<p>Also, if your film under the microscope doesn't show more detail than your Coolscan 8000 captures, you may consider reviewing your equipment. All the films I use clearly outresolve my Coolscan 9000.</p>

[Landrum Kelly]

<p>Rishi, that was made with a 10 MP DSLR.</p>

<p>Now where's my prize?</p>

<p>--Lannie</p>

[mauro_franic]

<p>Lannie, I was referring to the 35mm next to 40D example.</p>

[mauro_franic]

<p>Rishi, certainly not MF properly scanned.</p>

[Landrum Kelly]

<p>Sorry, Mauro. I misread.</p>

<p>--Lannie</p>

[david_littleboy__tokyo__ja]

<p>"DIGITAL AT 24mm x 36mm IS NOT ANYWHERE CLOSE TO MEDIUM FORMAT YET."</p>

<p>As someone who has been doing MF for a long time, 21MP looks to be an improvement over 645. My tests show 6x7 beating out both 645 and 12MP digital, but 21MP digital capturing the detail that 645 failed to. And it's not just the scans, it's what I'm seeing on the film with loupes. Lots of other people also find this (i.e. that 12MP is close to, 21MP is better than 645). This isn't news, or surprising.<br>

I'm _not_ saying "21MP is better than 6x7", I'm saying the 21MP is enough better than 645 to put it into the 6x7 ballpark. In my (and just about everyone else's) tests, 6x7 clearly beats out 12MP, but 21MP is so close that arguing is quibbling; 21MP really does capture details that 645 fails to. (And I was the one who first pointed out that the D2x lost the blue rivers[g].)</p>

<p>To get back to the original point, I don't blame bad scans. I blame physical reality. Most of the tests out there are quite good Nikon 9000 scans (other than the eggregious disasters at Luminous Landscape). I simply think that the physical reality is that film is gorgeous at 7x but shows infelicities that I don't want to see in my prints at 12x.</p>

[Landrum Kelly]

<p>David, leaving aside the 21 MP results for a moment, are you sure that 12 MP is outresolving 645 film? That just sounds counterintuitive. There is a lot of data on 645 film, 75% as much as there is on 6x6.</p>

<p>Can you link us to some comparisons showing us what you have seen?</p>

<p>--Lannie</p>

[david_littleboy__tokyo__ja]

<p>"you sure that 12 MP is outresolving 645 film?"</p>

<p>Sorry, I didn't intend to say that. I think that 12MP and 645 are close enough to be functionally equivalent for pictorial photographic purposes. Again, both 21MP and 6x7 are superior to 12MP/645, IMHO. (Did I reverse digits somewhere???)</p>

<p>"A lot of data on 645": I compare along the horizontal (i.e. with compositions matched on the long axis) and ignore 6x6. This shortchanges film for 11x14 and 16x20, but gets it right for 12x18 and 20x30. But, IMHO, these should be ballpark estimates.</p>

[robert lee]

<blockquote>

<p>You're taking of aligning image elements, not simply pixels, correct? ... Also, yeah you could do it for film. Man it'd be tedious to do that plus AEB for slide film. Of course, if you did it at the level of scanning, you wouldn't get rid of film grain.</p>

</blockquote>

<p>Think of it as shooting multiple frames of the same thing with the same framing. The alignment involves identifying fixed key points across the frames, then using these mile posts to translate, rotate, and perhaps scale against the chosen reference frame.</p>

<p>The technique is just an (optional) step in the contrast blending process. It's done to gain tonal precision and increase dynamic range. By itself it is inferior to super resolution since resolution gain is entirely absent.</p>

<p>When used with multiple film frames, film grain is reduced just as with the digitally originated images. In the limit, I suppose one could make Neopan 1600 look like a 100ISO shot out of the 5D at 100ISO.</p>

<p>For a film workflow, contrast blending can potentially be useful for digitizing dense reversal film. The idea is to scan the same frame using successively higher analog gain settings. The alignment step fixes registration issues frame to frame. The enfuse stacking step gets at dense shadow details unavailable with low gain settings, and retains the highlight details lost at high gain settings due to sensor saturation and bloom.</p>

[don_e]

<p>"Tell me what format you think it is. dSLR? 35mm? MF? Let's keep this objective."</p>

<p>Kit zoom lens? That suggests dslr.</p>

 

[roman_p]

<p >Thanks Jeff for your support. But I have a couple words to Scott too. </p>

<p > </p>

<p >Scott:</p>

<p ><em >Nikon and other camera manufactures are concentrating on the digital market because there is virtually no sales of new film cameras.</em></p>

<p >Scott, seems to me that you, as many digital shooters, are just looking for excuses why you switched to digital and trying to convince yourself you did a right thing. It’s typical when people suddenly recognize they have been fooled. “no sales of film cameras” we hear it for about a decade. But there are no such thing like “no sale”. They just don’t want to make them.</p>

<p > </p>

<p >To support this point lets imagine such situation. The last scanner(s) from Nikon had been introduced in 2003 and the last 35mm camera - in 2004 (F6). In same time Nikon introduced “the state of art, the camera of the future” D70 for $1200 USD (correct me if I’m wrong). 4 ½ years later this “camera of the future” already dead (no much future though). But we still are using our scanners, they cost as much as 4 yr ago, and look at Mauro test – they produce better image than another “camera of the future” – D2X! And if Nikon had stopped developing digital cameras in 2004, how much sales do you think your D70 would have now for the price of $1200? Let me guess – nil. That means the digital would be dead instantly. And don’t expect that D700 or D3(X, whatever) will have better destiny. Scott, I don’t think you’re not aware of it; seems to me you are afraid to confess that this is right.</p>

<p > </p>

<p >And I don’t think that the idea to eliminate art of film photography from our life came out from Nikon or Canon. However, there money talks. Apparently this idea was widely supported by Microsoft, Sony and other financial giants, who can control not just the market, but our mind as well and, using their financial leverage, may form any public opinion favorable to them. And unfortunately our “free” public media has no much to resist them.</p>

<p > </p>

<p >I don’t know how many folks participating in this discussion really understand that the digital technology is quite outdated and based on primitive binary digits theory to represent all numbers and data (which means “pixels”). This theory has been introduced in 30’s last century and hasn’t been changed much since. This somehow may explain why digital camera market is not stable, and digital cameras are dying that fast. </p>

<p > </p>

<p ><em >you are rather out of touch with what has been going on the last 10 years or so.</em></p>

<p > </p>

<p >No, Scott, I’m pretty much in touch with what I’m doing having my three 35mm Nikon bodies. And I’m sure that I’m going to use them for many years to come. But those who made themselves addictive to this so called “advanced digital technology” will face hard time.</p>

<p > </p>

<p >Rishi:</p>

<p >I guess it’s DSLR. Too “unrealistic” to be film (well, at least what I’m using).</p>

[rishij]

<p>Don E:</p>

<blockquote>

<p>Kit zoom lens? That suggests dslr.</p>

</blockquote>

<p>Huh? What's that supposed to mean?<br>

-Rishi</p>

[rishij]

<p>Robert:</p>

<blockquote>

<p>For a film workflow, contrast blending can potentially be useful for digitizing dense reversal film</p>

</blockquote>

<p>Yup. Some of the patents from the 90's regarding this sort of technique amaze me in the sense that they were granted. They're entirely obvious & self-evident methods. For example, ND filter on one row of the CCD while nothing on the other.</p>

<p>Multiple scans, definitely with multi-sampling on the higher analog gain ones that are specifically aimed at extracting information from shadows, would certainly be useful. One could even merge in software like PhotoAcute and/or Photomatix. Additionally, modifying the light source and/or adding a Peltier element onto the back of the CCD would be worthwhile in such a workflow.</p>

<p>Ahh the endless possibilities, if one has the time & effort :)<br>

-Rishi</p>

[don_e]

<p>Huh? What's that supposed to mean?<br /> -Rishi</p>

<p> </p><div></div>

[Landrum Kelly]

<p>Kit lens? I don't know about that. It's not terrible in the corners.</p>

<p>When are you going to tell us what it was shot with, Rishi.</p>

<p>--Lannie</p>

[Landrum Kelly]

<blockquote>

<p>I don’t know how many folks participating in this discussion really understand that the digital technology is quite outdated and based on primitive binary digits theory to represent all numbers and data (which means “pixels”). This theory has been introduced in 30’s last century and hasn’t been changed much since. <strong>This somehow may explain why digital camera market is not stable, and digital cameras are dying that fast. </strong> --Roman Ponomarjov</p>

</blockquote>

<p>Uh, no. They are not dying, merely becoming obsolete because the technology is advancing so rapidly. There's no conspiracy here. The manufacturers simply do not see a long-term future for film scanning, nor a great profit potential in it either, compared to what they can get from selling more and more impressive DSLRs.</p>

<p>Is digital scanning approaching the end of its life? People still have slides and negatives to scan, and art schools still are using film technology. Even so, at some point, the time costs of scanning have to be weighed against the benefits, and as digital sensors improve (along with in-camera processing), the benefits of scanning grow less and less by comparison.</p>

<p>We also keep saying that surely we are approaching the limits of how many pixels may be packed onto a 24 x 36 sensor, but we might be in for a surprise. Quantum effects aside (and those are indeterminate for me, to say the least), perhaps some new technology will allow more and more pixels to be packed onto such small surfaces, without interference and thus noise--never say never, since a technological and theoretical paradigm shift might be around the corner. (Do we ever see them coming?)</p>

<p>By the way, Roman, there is nothing "primitive" about binary digits. One might as well say that base ten is primitive because it has been around for so long. There is nothing imminently obsolete about 1 or 0, yes or no. Women have employed those choices for millions of years quite successfully, leading us around with rings through our noses even better than the camera companies have.</p>

<p>--Lannie</p>

[leo_papandreou1]

<blockquote>

<p>Tell me what format you think it is.</p>

</blockquote>

<p>Velvia, 35mm, overcast day.</p>

[roman_p]

<p >Alright, Lannie, may be I went over board calling binary digit theory primitive. As a design enginner I’m using computers lot. It really helps up with all our calculations and analysis. But it’s technical things. However I consider it’s “primitive” regarding visual art. How much creative you would be if your brain can recognize only two digits -1 and 0? </p>

<p > </p>

<p ><em >They are not dying, merely becoming obsolete because the technology is advancing so rapidly.</em></p>

<p ><em > </em></p>

<p >I have to disagree with you. This technology had stopped advancing long time ago. The difference between modern computer and 20 yrs ago is only the speed of data processing and transferring and memory volume. So are digital cameras. It’s not advancing, it’s retrofitting. </p>