Best Scanner for Medium/Large Format Film?

[timlayton]

<p>Stuart, you asked for more examples from the D3X and film. If you go to my Flickr stream I have a lot out there. The link is: <a href="http://www.flickr.com/photos/timlaytonsr">http://www.flickr.com/photos/timlaytonsr</a><br>

I have separated the film from the rest, but if you look at the file names in the description it is easy to tell the difference between my D3S, D3X and 5D Mark II and 1D Mark IV files. I am getting ready to post a bunch more to the film set very soon because I finally got my scanning workflow where it needs to be. This is the whole reason I started this thread. I have been going through this journey of figuring out the limitations, benefits, etc with scanning my MF film and I think I am finally at a place where I have a good handle on it and can actually leverage it wtihin my digital workflow. <br>

You asked for my personal conclusions so I will tell you what I think and keep in mind this is just my opinion based on my own circumstances and variables as well as my my own personal biases and possibly limitations. <br>

I have been a photographer for about 25 years and from the mid 80's till about 2001 I shot MF exclusively. I own 6x45 and 6x7 systems, but almost always shoot 6x7 unless I want to just enjoy my 6x45 rig or know I won't need a big print. For my studio and macro work I use the RZ67 Pro II with a full line up of glass and for urban and some landscape I use the Mamiya 7 Rangefinder. On the digital side I switched over to Nikon a couple years ago and have the D3S and the D3X as well as some old 35mm bodies that still work great and definitely enjoy from time to time, but I just don't use them for professional work. I had owned a Canon 1D Mark IV and a 5D Mark II for a couple years and shot about 25,000 images with those systems. <br>

Here are my personal observations in random order:<br>

1.) Now that I know the technical limitations of the Epson V750 and have learned the Silverfast AI tool I can successfully produce extremely sharp and professional images for my customers from my MF transparencies (typically FujiChrome). I concur with others the limitation is about 2400dpi on the Epson V750 which produces a file that is about 6600 pixels by 5250. I've done several tests at 48-bit and 24-bit scans and I effectively can't see a difference in my prints. The largest print I've tested that with has been 24". So, I use the 24-bit and it almost cuts the file size in half to about 94Mb. Using the 300 dpi rule for prints you can easily get 22" prints and in many cases 240dpi is very acceptable and will easily produce 27"+ prints. <br>

2.) If I need, or think I need, more detail, broader dynamic range and higher resolution I just take to my local pro lab for true 4000dpi scans which produce files well over 10,000 pixels on each side. That will produce 36" prints and larger with no problem. For me the end result is always the quality of the print and if my customer is happy. I've found that it really depends on the image if it really reveals or warrants a 4000dpi scan for the smaller prints or not. My pro lab is about 45 minutes away from me making this a 3 hour process each time I need high res scans so I really make sure I need these because it involved two trips and a lot of time. If I could find a new Nikon 9000ED that wasn't over list price I would probably just buy it and be done with it. My time is worth more to me than the time lost doing these trips. To be honest if I had realized the Nikon 9000 was compatible with Snow Leopard I would have just bought it out of the gate, but in an odd way by starting wit the V750 I have learned a lot of valuable lessons and workflow that will still apply to the Nikon. It turns out that the V750 is capable of producing very good quality in my opinion. By the way I did find out that by using the after market holders and ANR inserts the quality went up. I got them from betterscanning.com. <br>

3.) I approach all of this from two perspectives: artistic and utilitarian. Many times a camera is simply a tool to me that helps me create what I have in my mind. Digital or film in some cases does not matter. When I do college sports or wildlife the optimum tool is the D3S with my 300 f/2.8 and set of the TC's. When I do my botanical fine art work in the studio it becomes a matter of preference at this point. For example, if you go look at my Flickr stream you will see a lot of my botanical work. The MF film just "looks" different than the digital images from the D3X and D3S. When I want that look that I can only get from film then I go with it. If I need a very clean and almost sterile look I go with digital. Since I have T/S lenses for my Nikon's I tend to use digital for architecture when needed and if I can use film I will. Almost all of my landscape work is film. I have even recently expanded into 4x5 for landscapes. In my studio and macro work I don't mind the slower pace of film and the added steps for the quality and product that I am able to produce. In fact, I find that it is setting me apart from others because of that different look and feel as opposed to digital. <br>

As far as one being "better" than the other is not really how I think about it. I have a long history with film and absolutely love the slower pace, more thoughtful approach to photography, and is now becoming a different look from the norm (digital). I couldn't do my job without my DSLR equipment because I couldn't do my college sports or wildlife with my MF gear. When I am riffling off 9 fps with my D3S and have a wild animal squared up in my viewfinder my heart is pounding 100mph. When I produce a piece of custom fine art for my client via my MF systems I feel a sense of pride and happiness that is unparalleled. But the bottom line and what I think you might be looking for, is when I produce a 24" or 30" print or piece of fine art, the quality I get from my D3X or from my MF film systems are both stunning, but in different ways. I recently spent a day shooting a new Hasselblad H4D-40 and the new H4D-31. For my studio and Macro work it definitely produces "cleaner" files than my scanning methods via film, but for a $15k price of admission I find myself very happy with scanning my MF film. Also, because it is digital it has that different look that I talked about, which is not a bad thing at all, just not my personal preference if I have an option. I suspect my next purchase will be a Nikon 9000 so I can eliminate the trips to my pro lab. In the end there are no right or wrong answers, to me if you or your customer is happy then the camera or the workflow is just a process to create your product. I will say that I fully acknowledge the world and workplace is clearly gone digital and that is one reason why I continue to evolve my digital gear, but there is also an independent part of me from the creative side that continues to produce fine art using MF film because I don't have to play by those rules in this scenario. <br>

Tim</p>

<p> </p>

[StuartMoxham]

<p>Tim that was a very thorough answer thanks for sharing your thoughts and observations with us. Very good to hear that you have found a place for both film and digital and that you appreciate the different qualities that they have to offer. Thanks again Tim for taking the time.</p>

[sebastianmoran]

<p>Tim, very nice images posted here and on your site.</p>

<p>First, about images from flatbed scans... I did some testing with a V500. In my tests, I get 1300 ppi on one axis and 2000 ppi on the other. It's better at resolving lines that run parallel to the scan track. See test target scan in this thread: <a rel="nofollow" href="../digital-darkroom-forum/00UZJA">http://www.photo.net/digital-darkroom-forum/00UZJA</a></p>

<p>So, what can you do with a V500? I recently used V500 scans of a 6x9 shot on color negative film to produce prints at 8x12 and 12x18. I like prints with enough detail to stand up to close inspection, and I judge these as plenty sharp. You are welcome to take a look, print the image or a sample, and draw your own conclusions.</p>

<p>Thread: <a rel="nofollow" href="../film-and-processing-forum/00W7Rk">http://www.photo.net/film-and-processing-forum/00W7Rk</a></p>

<p>Large sample file 3600x5400 pixels ready for print at 12x18, shot on 6x9 film and scanned with my V500:<a rel="nofollow" href="http://2under.net/images/100201-Mamiya-100-f28-Cheers-Img6-v500-12x18.jpg" target="_blank">http://2under.net/images/100201-Mamiya-100-f28-Cheers-Img6-v500-12x18.jpg</a></p>

<p>I have not tested larger prints. My result is consistent with you getting good 20"+ prints from V750 scans of 6x7.</p>

 

[sebastianmoran]

<p>Tim, you asked about digital pixels vs. pixels from scanned film.</p>

<p>I did several film vs. digital side-by-side comparisons and concluded: In terms of image quality in a print at 12x18", 6x9 scan of Ektar 100 on a v500... is about equal to a good 12MPx DSLR, and about equal toCoolscan V scan of Ektar 35mm. (Tim, I think this is pretty close to your estimate that V750 scans of 6x7 equal a 24MPx DSLR.) </p>

<p>So, I find that 4000 x 6000 pixel scans from film produce images of the same quality, in my subjective appraisal, as 10-12 MPx DSLR images. It seems that digital pixels are better than pixels from scanned film. Why?</p>

<p>I found a reference that I found interesting, "Sony 24p Technical Seminar #2" about sharpness and resolution of digital vs. film for movies. I would love to hear reactions of others to this.</p>

<p>The gist of the article is about contrast vs. detail (MTF vs. spatial frequency). Film contrast declines continuously from low detail to extinction (max resolved detail). Digital is different; digital maintains higher contrast up to the ppi where it all falls apart. If we measure both systems in terms of max resolved detail (extinction resolution), then film shows greater ultimate resolution, but digital has higher contrast in the middle. </p>

<p><img src="http://2under.net/images/MTFGraphFilmDigital.jpg" alt="" /> </p>

<p>Then we need to overlay human vision on top of this technical performance. What gives a human viewer the perception of sharpness? Otto Schade, Sr. at RCA, in research back in the television era, found human viewers call it "sharp" when they see high contrast in medium detail. </p>

<p>To me this is an explanation that makes sense: Digital images look better for their pixel count than film images because the digital pixels carry higher contrast (MTF) in ranges of detail that matter to human observers.</p>

<p>If all this is correct, then MTF-50, rather than ultimate resolution, would be a better measure for evaluating imaging systems. </p>

<p>The Sony paper: <a href="http://bssc.sel.sony.com/Professional/production/productsite/files/24PTechnicalSeminar2.pdf">http://bssc.sel.sony.com/Professional/production/productsite/files/24PTechnicalSeminar2.pdf</a></p>

<p>In case the link breaks, Google "Sony 24p technical seminar 2"</p>

<p>This is all outside my expertise, but I want to understand it better. Would love to know what experts think? Does anyone know the credentials of the author, Thorpe?</p>

[danielleetaylor]

<p>Tim Layton - Before this became another film vs. digital thread Rodeo Joe gave you the best answer you're going to get if you want high quality at low cost: "scan" your film using a DSLR, a good macro lens, and a light box. For larger formats you simply stitch multiple shots, each shot being a subsection of the total film area. Otherwise for MF the Nikon CoolScan 9000 is your best bet.</p>

<p>As to the thread topic diversion: if you want to shoot digital, shoot digital. If you want to shoot film, shoot film. The resolution comparisons and debates have been done to death. Top tier DSLRs comfortably exceed 35mm in absolute IQ, and are comparable to MF at print sizes up to about 24". A 3 frame stitch from a top tier DSLR will comfortably match or exceed the best 6x7 scans. MF backs are comparable to 4x5 at this point. That gives you a pretty good idea of options and where things stand.</p>

[danielleetaylor]

<p>Philip Wilson - there was obviously something wrong with your first 5D2 crop. My gut instinct says vibration. The second one was much better. It would be interesting to see that file properly enlarged to GX680 scan dimensions and compared at 100% and in print. My guess is the GX680 would still show some additional detail, but that in print up to about 24" or maybe 30" the 5D2 would be comparable.</p>

<p>I do not agree that lenses are the limiting factor in resolution for small format digital. Plenty of lenses can out resolve the 5D2 and even 7D sensors, especially now that manufacturers are updating lenses to meet the demands of modern sensors. At the 7D's pixel density a FF sensor would offer 45 MP. I would love to see such a sensor as it would match or out resolve 6x7 scans.</p>

<p>I also just can't let the comment about the 7D go. I won't post samples because it's too far off topic, but your statement has been disproven by many people in many places. If your 7D is not performing to within a hair of your 5D2 in terms of IQ, at low to mid ISO, then there is something wrong.</p>

[alex_iwonttell]

<p>Not sure if this has already been mentioned, but there are 3rd party wet mount adapters with variable height available for Epson 750: http://www.betterscanning.com/ - I've seen some examples (not on their web site) which have shown a tremendous improvement. Best value for the buck, IMHO.</p>

[timlayton]

<p>Based on Daniel Taylor's feedback I just went and did a simple test. <br>

The first image was photographed with the RZ67 Pro II and 140mm Macro lens. It was scanned with the Epson V750 at 24bit, 2400dpi.<br>

<img src="http://i605.photobucket.com/albums/tt140/timlayton/RZ-1-24k-24bit_6533x5260-72dpi.jpg" alt="" width="1023" height="824" /><br>

This second image I placed the transparency on my light table and used my Nikon D3S with the Nikkor 105mm Macro lens at f/32. <br>

<img src="http://i605.photobucket.com/albums/tt140/timlayton/_DSC1547.jpg" alt="" /><br>

Tim</p>

 

[AlanKlein]

<p>I read main thrust of the Sony paper. They're comparing video vs 35mm film processes to create movies and why resolution and sharpness are different between film and digital. In fact the title of the paper is<em> Sony HD Cam Picture Sharpness - Issues of Image Resolution</em>. I don't see how that proves anything one way or the other since the way we create final prints is a different process. Moving images are observed different by the human eye and brain. Do you know of a study that campares these issues more directly to what we do?</p>

[sebastianmoran]

<p>Alan, not trying to prove anything, just looking for explanations for what I see in images. The article is about digital vs. film capture and how to think about image quality, sharpness, and resolution. </p>

<p>It does promote Sony digital gear, so it's not exactly unbiased.</p>

<p>In my images, I think that digital compare more favorably to film than the ultimate resolution tests would indicate. I think digital pixels are "better" than scanned film pixels. I'd like to understand why.</p>

[mjferron]

<p>This is a 35mm scan from the lowley V700. I usually don't push 35mm past 8x10 but i made an impressive 11x14 from this negative. The film must be flat and the emultion side facing up. I keep my adjusters as close to the glass as possible. I also scan medium format with excellent results. Sometimes I don't think folks who trash this scanner have spent a lot of time with it.</p><div></div>

[StuartMoxham]

<p>Look at it this way Richard thousands of photographers went over to digital capture. Many photographers gave up Hasselbalds and Bronicas and started using DSLRs and we are not talking 20mp full frame either. There must be a reason beyond just material costs and convenience.</p>

<p>Most photographers I know care about the quality of their work and how it looks. Most of them did not take lightly to making the change. Many used digital alongside their film cameras and compared the prints and files and drew rather similar conclusions to what you I and many others see.</p>

<p>I'm not pro film or pro digital I am quite happy to use both. To me does not matter about resolution charts or if I can read all the text on a tiny distant sign when viewing a 100% crop if the prints will never be large enough see the text on the distant sign.</p>

<p>What does matter is how the prints look though. I do notice that with the smaller prints I make the hand enlarged B&W prints seem to have more details than similar sized digi minilab prints. When we resize or the lab resizes our digital camera files to 6x4 inches at 300ppi we end up with a file around 2mp alot of the pixels were thrown away. Personaly I feel that the smaller B&W hand enlarged 3 1/2 x 5 to 5x7 inch prints have more details and better tonality than the ones I get back from the digilab of couse I am talking about the prints from sharp negs.</p>

[StuartMoxham]

<p>Michael nice old Cadilac in that shot. Here's a Buick.</p><div></div>

[uk]

<p>Stuart,<br>

and here's the back of a Buick Riviera just like that last image.</p>

<p>This time, a 35mm Tri-X frame scanned on my Imacon 848.</p>

[StuartMoxham]

<p>Thanks Gary. Red Mustang from the same show. Film was Agfa Portrait 160, 645 format.</p><div></div>

[AlanKlein]

<p>I've been using the Epson V600 for 35mm and 6x7. It doesn't do 4x5's. I'm not a pro but it seems pretty good. The shadows tend to be choked up a litttle especially on 35mm and I do PP to sharpen up the images. I don't process during the scan except for ICE sometimes if the negatives are dirty and I can't blow off the dust enough. It takes a lot longer to scan with ICE. </p>

<p>You can check my gallery for both 35mm and 6x7's. I tried enlarging the 6x7 lighthouse to 17x22" print (I printed 1/4 of the image onto my Canon 8 1/2x11" printer) and I thought the picture looked pretty good if you didn't get too close to eyeball. The three egret and stream shots were cropped to a 1/4 of the original picture so those should give an idea of equivalent enlargements. Hope these samples help you make a decision. Alan</p>

<p><a href="../photos/AlanKlein">http://www.photo.net/photos/AlanKlein</a></p>

[sebastianmoran]

<p>Hi Stuart -- The car photos are great; I like the red Mustang. </p>

<p>Replying to your comments on the quality of digital: I was quite happy with the IQ I got from my 6MPx camera (D70). I got a few nice 8x10 prints and many nice screen images from my 1.5MPx camera (Fuji MX-700), which really amazed me. There could be a lot to talk about in here, but following the OP's (Tim's) questions, the one on my mind is why does digital look as good as it does in print? That's why I was interested in the Sony article.</p>

[StuartMoxham]

<p>Richard the way that old Agfa portrait did bright reds was really something special. For a film that had a muted color palete it had really great reds. The old optical prints I had from this roll had great reds. I just wish I knew where the prints were.</p>

<p>I've got some really nice 8x12inch prints from my D70 and D80, at this size you get everything in the print that a 6-10mp DSLR has to offer but you don't really have any grain, the prints are exceptionaly clean, color and saturation is exactly where I want it, there are enough details for the print to look detailed and there is a great appearance of sharpness. With color 35mm neg films that I had been using an 8x12 would just start showing grain and that grain I find adds a texture that is not there in real life the details are likely the same maybe a touch better but I always felt that 8x12 was as big as I wanted to go with 35mm film.</p>

[dan_south]

Interesting discussion! When using a DSLR and a macro lens to re-photograph a chrome on a light table, would it be

possible to capture different sections of the chrome in separate shots and combine them with stitching techniques? If that

would work, resolution and aspect ratio mismatches would no longer be problematic. I'm wondering whether anyone has

tried this approach. I have a lot of chromes in my files that are dying to see the light of day.

[dan_south]

Adding to the comment above, my concern would be that some distortion would be added since the camera and lens

position changes for each section of the chrome. Would that pose a problem or would the stitching software compensate

for this?

[Landrum Kelly]

<p>For what it's worth, the Nikon Coolscan 9000 is suddenly available at J&R Music World, or at least it was five minutes ago. The fact that they have been rather hard to come by lately is the only reason that I announce that here.</p>

<p>I am not suggesting that it is the best, simply that it is a choice somwhere between the Epson and Imacon in terms of both quality and price.</p>

<p>I wonder how long they will stay in stock.</p>

<p>--Lannie</p>

[timlayton]

<p>Dan, the short answer to your question is yes, however I am not sure how practical that application would be for more than a few images from time to time. To give you a little insight, I use that technique with my Nikon DSLR's and my T/S lens when I need a cosistent DOF across my subject. Typically this is something like three flowers or a similar subject. I setup in manual mode dialing in my exposure and aperture and for the first of three images and don't touch anything again. I shift left to cover the left 1/3rd of the image. Then shift to center overlapping about 1/3rd of the left image and then shift to the right overlapping about 1/3rd again. I do a photo merge in Photoshop CS5 and it works flawlessly. I've never tried this technique with a regular or macro lens probably because the T/S does such a fantastic job. </p>

<p>Tim</p>

 

[f_ph1]

<p>Michael Ferron:<br /> <br />Your image from 35mm scanned on the v700 (impressive, by the way) proves an inconvenient point. Notice how many of us would rather keep this ridiculous argument going, despite glaring evidence in favor of the quality the Epson scanner is capable of?<br /> Again, I can't help but conclude that it is due to two primary facts: many photography enthusiasts want to justify the copious sums of money they spent on big ticket photo gear, and/or they want to substitute empirical data (e.g. lp/mm, etc.) for creativity and talent as the benchmark for the making of compelling (quality) photographs.<br /> The argument that some "settle" for less quality, which implies that those who don't tremble in the shadow of the Coolscan or Imacon have far lower standards, is bullsh#t. I love how the photograph you posted flies in the face of this.</p>

<p>The emperor is naked. Some think he should be wearing Armani or nothing at all, when a good pair of Levis will suffice.</p>

[StuartMoxham]

<p>One more from the v500 35mm kodak gold 100.</p><div></div>

[rob_piontek]

<p>Just want to say that I shot two frames of a 6x6 neg with my 12MP DSLR and stitched them to get a 16MP image. I wasn't even being very careful and it came out fine. Certainly if you are careful with alignment and film flatness more frames are no problem. The problem is time - how long are you willing to spend per image? And in any case will you actually print them large enough to warrant more than 12MP?</p>

<p>On the other hand, the article I linked to above, comes to the conclusion that if you have a large collection to scan, scanning with your DSLR (with 1 shot per neg) can be much faster than a flatbed and the quality can be very good.</p>