It all comes down to the print

[15sunrises]

<p>And the circle begins again. </p>

<p>I'm happy that the type of photography I do doesn't require me to conduct such in-depth tests of the quality of my equipment. This truly seems to be a nearly pointless endeavour that could take up endless amounts of time. Why not just shoot what you enjoy and if you find your medium limiting, then do some research and find out why. Not quite sure what photos of maps are supposed to contribute to the photographic community beyond creating a point of debate.</p>

<p>Also, the shots posted at Apr 18, 2010; 05:04 p.m. by Mauro seem to be improperly focused on the 40D. My 40D far outresolves that with a 3rd party lens.</p>

[mauro_franic]

<p>Dave T,</p>

<p>The 40D shots are perfectly focused and were taken with a 60mm macro - the sharpest lens available - (as you can tell with resolution in the chart turning into moire).</p>

<p>The were just enlarged to match re resolution of the scan.</p>

<p>100%:</p><div></div>

[15sunrises]

<p>Still seems blurrier and more jaggy than the results that I get with poorer lenses.</p>

[mauro_franic]

<p>That is because you assume the camera will behave in a certain way. But it doesn't. You are welcome to try and post an example. For this test, just shoot the resolution chart with the chart fitting aprox 9 times in the height of the picture.</p>

<p>What you see is resolution extinction not bluriness. i.e. there isn't enough detail to resolve the letter, numbers, colors or lines further. This happens below the nominal pixel resolution.</p>

<p>Jaggy and blurrier are friends that move in opposite direction.</p>

<p>You see jagginess and moire because the AA filter could not blur input enough to avoid them.</p>

[mauro_franic]

<p>"Not quite sure what photos of maps are supposed to contribute"</p>

<p>They help you understand the limitations of your equipment.</p>

<p>For example, if you printed the test page which is the object of this discussion, you probably would not use the 40D for a shot that you may decide to sell or put on a wall at larger than 11x4 or 16x20 (I own/use a 40D as well).</p>

<p>On the other hand you know you can print 30x40 or larger from MF without any compromise (I own/use MF as well).</p>

<p>This test have direct impact to me and others. e.g. I would still use the 40D when it is convenient especially if I will not print larger than 11x14.</p>

[mike_hitchen]

<p>MF beats APS-C when printed at 40x30 inches.<br>

I may haev missed the object of this thread but is there anyone out there who expected anything different?</p>

[15sunrises]

<p>The point is to get people photographing maps and focus charts rather than using their cameras outside apparently... it worked.</p>

[mauro_franic]

<p>ha ha... the government must be behind it.</p>

[matt_needham]

<p>I go back to the simplicity of the post header. It all comes down to the print. If you have experience in both processes it should be easy and cheap enough to hang large prints next to each other, and see which you prefer. My prediction is that experience will influence print quality quite a bit. Folks who've been studying one process and not the other are going to find they make good prints from the process they are familiar with, and lousy ones from the process they don't know. It's easy enough to go out and see mediocre prints from all processes. On the other hand it is also possible to find stunning examples from both film and digital. If even one person can pull it off it suggests to me that the supposed weaknesses are less with the tools and more with the tool users. When I bought my first DSLR my prints from digital didn't look as good as the prints from medium format film that had been coming out of my darkroom for the previous decade. Six years later and my darkroom is shut down. I've been living with large prints from both for years now, looking at them everyday, and have no regrets as to my choice of process. </p>

[Ed_Ingold]

<p>In the interest of an apples-apples comparision, I ran Smart Sharpen (1.0/150%)* on <strong>both </strong>the Velvia and 7D panels. The 7D panel (right) still looks better.</p>

<p>* This is more sharpening than I would normally use. Something on the order of 0.7/150 would produce fewer artifacts. However, I attempted to match the effects posed by Mauro.</p><div></div>

[Ed_Ingold]

<p>It is accepted practice to resample digital images, scanned or original, to between 200 and 300 ppi for printing so that the individual pixels do not show. The granularity of a digital image enlarged 4x or so without resampling does not fairly represent how the image would appear. The train was moving far to fast to attach a resolution chart ;-)</p>

<p>This example shows a 400% crop along with an unsharpened resampled version of that crop. The actual resolution is not changed, but the appearance is greatly improved. The original picture was taken with a Nikon D2x, and the example is equivalent a 40x60 inch print examined with a 4x loupe.</p><div></div>

[mauro_franic]

<p>Thank you Edward.</p>

<p>The optimal sharpening level seems to either be higher for the film scan or the 7D had some sharpening applied to it by the camera. Matching the resulting sharpening make a better comparison.<br>

Still, to me, the 7D does not outresolve this jpeg from Velvia 35mm I sharpened.</p>

<p>I'm also waiting for Daniel's raw to look at it first hand. Les will send me his full scan of Velvia.<br>

It will be clear to see the differences once I have the uncompressed unprocessed files.</p>

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

[mauro_franic]

<p>I won't make any definite statements until I have the sources from Daniel and Les - hopefully soon, but from the shape of the stars in the jpegs, the 7D does not outresolve the 35mm scan.</p>

<p><br />I will post the comparisons of the originals as soon as I receive them.</p>

[mauro_franic]

<p>Edward, what resolution do I have to print the chain you just posted to equal a 40x60 print.</p>

<p>Please let me know, I'd like to print it now to see what it looks like - no loupe.</p>

[craig_big]

Defending digital on the basis of resolution is sort of foolish. Defending either system on an emotional basis is just as foolish. If people would get over these issues, the argument would have been over 10 years ago.

 

Mauro makes a good point when he says that most people adopted digital workflow back when the resolution was much worse than it is today. Most people bought their first digital cameras with resolutions like 2MP or 3MP. It was widely accepted at the time that digital didn't have the same resolution as film, but the improved workflow made up for the difference. Digital cameras also removed much of the mystery from photography for much of the general public. For the first time, casual users could see the REAL resolution of their photographs. With 1-hour film labs and 4x6 prints, beginners were never able to realize that before. Now, with better digital sensors, we've seen the gap narrow to the point where a test has to be very precise in order to prove what is still the reality: film flat out resolves better than any digital sensor yet to be invented. But the only people who notice anymore are medium format shooters. In 35mm at least, the gap is awfully narrow, and a bad film scan will skew the test in the wrong direction.

[sebastianmoran]

<p>Mauro, you wrote:</p>

<blockquote>

<p>The optimal sharpening level seems to either be higher for the film scan or the 7D had some sharpening applied to it by the camera</p>

</blockquote>

<p>Yes, the digital capture had some sharpening applied in the process of converting the Bayer pattern to pixels. <br>

<br /><br>

Because they came from different sources, all the processing would be different, including the sharpening, to produce comparable results. In my experience, any film scan needs noise reduction as well.</p>

[Ed_Ingold]

<p>The uncertainty (blur) in film scans is nearly Gaussian, and can stand significantly more sharpening than a DSLR image. I generally stay below 1 pixel in width, but may go as high as 250% in the amount (USM), taking care to mask out broad areas without detail.</p>

<p>The blur is largely due to optical diffusion (scattering) in the film and chemical diffusion in the development process. It does not appear to be a limitation in the (Nikon) scanner. First rate lenses, combined with good technique, are generally not limiting either. The effect extends 4-6 pixels wide in an otherwise sharp picture, whereas the reproduction of physical defect, like a scratch, is sharp to within 1-2 pixels. This is consistent with the rolloff at 20-30 lp/mm seen in published MTF curves for film.</p>

[mauro_franic]

<p>Craig's statement is a very good summary.<br>

Particularly regarding to the fact that film's workflow (scanning and processing especially), if done incorrectly or with sub-par equipment, can negatively impact the outcome of 35mm film to a point it can be inferior than a DSLR's. This, combined with the added convenience of digital for mass producing shots, made DSLRs a preferable solution for many consumers and commercial photographers early on.</p>

<p> </p>

[craig_big]

Let's not ignore Mauro's post from Apr 18, 2010; 05:17 p.m. He demonstrates how much more clearly an Imacon scanner can pull detail from film vs a Nikon Coolscan. The Nikon is considered top-of-the-line for amateur and non-professional use, but produces unacceptable results for professional scanning. Really critical scans are done on a wet-mounted drum-scanner, which takes the Imacon scan 1 notch better. So yes, blur is a limitation in the (Nikon) scanner.

[Ed_Ingold]

<p><em>Edward, what resolution do I have to print the chain you just posted to equal a 40x60 print.</em></p>

<p>Print it at 280 ppi without resampling. The example is a 400% crop of the original, which would be 10x15 inches at 284 ppi before resampling. Use a small sheet of paper - the print will only be about 3/4 x 1-1/2 inches.</p>

[Ed_Ingold]

<p><em>The Nikon is considered top-of-the-line for amateur and non-professional use, but produces unacceptable results for professional scanning.</em></p>

<p>We are really quibbling about minutia - the Imacon has perhaps 25% better resolution than the Nikon. You could also argue that home printing can never replace large prints by professional shops. In either case, you are wrong.</p>

<p>The fact is most people do not have a Nikon scanner, so the alternatives are using an affordable flatbed or having scans done professionally. In that case, most service bureaus use what they have - drum scanners. Likewise, few people have a 24" (or wider) printer at home. Those that do (e.g., Mauro and Michael Reichmann, <a href="http://www.Luminous-Landscape.com">www.Luminous-Landscape.com</a>) would argue otherwise.</p>

<p>The fact is, if you want big enlargements (over 16x20") not limited by the source, you need to use larger film (or MF digital), representing a vanishingly small portion of the community of photographers.</p>

<p>What this thread demonstrates is what you can expect when printing from different photographic media, and has not degraded into the typical "either/or" responses. Many times "good enough" is simply good enough. That applies when building bridges and airplanes too, where the consequences real problems are more newsworthy.</p>

[craig_big]

There have always been two different arguments happening at the same time: "What can film actually do?" and "What does it do for me?". The second question is the one that brings up all the emotional responses. It is the first question that begs to be answered with the most scientific testing on the best equipment, and requires that the minutia be analyzed, namely the differences between scanners.

 

I've already seen excellent test results comparing optical prints from 35mm film to digital prints from a 24MP full frame. There is a clear winner. The problem is that when everyone abandoned film, we all abandoned the optical printers, too. Now we're comparing film SCANS to digital capture. This is an impossible battle, because not only does digital have an advantage in this realm, but the film workflow has been impossibly encumbered to an unusable degree.

 

What I see in reality is that the resolving potential of digital is only now beginning to match the resolution achieved by film in a fully optical process 50 years ago. Now all the digital shooters feel vindicated for their now 10 years of devotion to an inferior medium. I hate to ring KR's bell, but Outdoor Photographer still shoots their covers on film. Hollywood still shoots movies on film, and there's a procedure that stands to save millions when they convert to an all-digital workflow.

 

Personally, I shoot both. I enjoy the digital workflow, but when I shoot film I just know the results will be perfect. And even when I botch something, I can recover alot more from a 35mm negative than from a digital RAW file. We used to call this "latitude".

 

When it comes down to the prints, however, I must admit that prints from the latest generation of digital cameras are really getting good. Probably just as good as film prints. I only extend this concession to the LATEST generation of digitals, like from 2008 onward. Everything I saw and used before that still had major shortcomings.

[Ed_Ingold]

<p><em>Hollywood still shoots movies on film...</em></p>

<p>Don't be so sure. By some estimates, nearly 70% of feature films are shot with digital cameras, generally 4K (or less). There is a minor industry making "glamour" filters for these cameras so that brutally sharp digital video looks more filmlike (q.v., Schneider Optics), standard issue for the Washington press corps as well. Nearly 100% of editing is done in digital, even if the original was shot on film, then printed on film for distibution. When enough theaters have HD digital projectors, that will change too.</p>

[craig_big]

I really don't mean to be contrary, but even according to your statistics, without checking first for their veracity, that still means 30% of all feature films to be released in 2010 will be shot on film stock. That's a staggering number in the digital age. According to opinions I've read from AFI directors, film is still regarded widely as being "excellent", and many more directors would use film if they always had the budget for it. My numbers may be off, because I'm just pulling statistics from memory, but I believe it costs somewhere around $100/ft to purchase, shoot, process, telecine, and color-correct film stock. I think they'll usually shoot something up to 250,000 feet, so the film cost alone is in the $10-25 million range for a feature film. There is undoubtedly alot of savings to be realized by using digital cameras instead, so there are some VERY dedicated producers out there who are sticking with film for its aesthetic quality. As far as I know, photographers do the same. The main reason to "switch" to digital back in 2001-2003 was to save time and money, not quality. Given the high cost of digital cameras, I'm not sure that most photographers actually saved the money. In 2010, they are almost on par for quality, and given the rising cost of film due to scarcity and development, there's a stronger argument for cost savings.

[sebastianmoran]

<p>Craig, really appreciate your last few postings. </p>

<p>Found this very interesting article addressing image quality in movies, film vs. direct digital production. It's a Sony technical presentation by Laurence J. Thorpe, referencing work by Otto Schade Sr. at RCA:<br>

<a href="http://pro.sony.com/bbsccms/ext/cinealta/docs/24PTechnicalSeminar2.pdf">http://pro.sony.com/bbsccms/ext/cinealta/docs/24PTechnicalSeminar2.pdf</a></p>

<p>Answers a question I've been wondering about: Prints from digital cameras look better than you would expect based purely on the resolution achieved. In the early days of digital, I made some pretty good 8x10's from a 1.5MPx camera (Fuji MX-70).</p>

<p>Why? Thorpe (and Schade) answer... Because the human viewer is more sensitive to high MTF than to ultimate resolution, and digital tends to have higher MTF until you run out of sensor sites. Schade's proposed measure is the area under the SQUARE of the MTF. If this is correct, MTF50 would be a better measure of image quality than extinction resolution. Any reactions to the article?</p>

<p>What does this mean for shooting film? Would we do anything differently if we were seeking the widest range of high MTF? Versus maximizing ultimate resolution? (No gripes about anyone's comments... Just trying to understand what going on in the image. Mauro, I enjoy the discussions you start!)</p>

<p> </p>