Soft 35mm scans - do I need a better scanner?

[randall_pukalo]

Here is a sample Costco scan from negative film. Click on the "All Sizes" button above the picture to see the full rez image.

[randall_pukalo]

Another Costco sample.

[randall_pukalo]

[robert lee]

<blockquote>

<p>... especially if I want to scan more than a single frame in a batch. That is one reason why I like the V700 ...</p>

</blockquote>

<p>This is the reason why the Nikon 5000 is the best available for newly shot (uncut) film. A fairly straightforward tweak will enable the film strip feeder to scan up to 40 continuous frames without intervention. For 16bit scans at 4000dpi, it takes just about one hour to digitize 36 exposures. The cool part is that you don't have to baby sit the process.</p>

<p>I'm not up to date on pricing, but the last I looked the 5000 was basically double the price of the V700. If you intend to shoot solely 135 format film, the 5000 is worthwhile. The resolution advantage is welcome, of course, given how area challenged 135 is. However, don't underestimate the usabiulity advantage of the much faster workflow.</p>

 

<blockquote>

<p>I am playing around with film myself as an amateur photographer, but I would like to scan old family photos as well. That will take insane amounts of time with a dedicated film scanner?</p>

</blockquote>

<p>Yes, it will take a horrendous amounts of time even with a dedicated scanner. The Nikon 9000 is probably the best of what's available for this purpose. The carrier for 135 format film will hold two strips of 6 frames.</p>

 

[phialv]

<p>Hi Peter, I think you have too many variables under consideration. You want to know if your scanner accurately and sensitively captures image data, as compared with whatever the 'gold standard' is. In order to do this, the test samples must be identical. In this case, you should obtain two print copies of one or more standards (you must find a source - IEEE, NIST, or somebody beyond reproach, and perhaps more than one test to look at different things) and scan each more than one time (to assure reproducibility) at two or more levels of scanning quality (ideally you'd like to know the quality at the 'mid-point'). Send copies of the same prints to a lab (I don't know if you want a drum or a flat-bed scanner) and repeat the test EXACTLY as you did it. They should NOT apply additional optimization. They should also report to you the precise conditions of the scan. You should then compare defined, discrete areas on the scanned image: dark areas, light areas, smooth & sharp transitions, etc.<br>

Note: None of this says anything about which image or method is 'better' or which process yields a scanned image you like more. It just shows the ability of your system to objectively scan images, as compared to the current best methods available.<br>

Good luck,Phil</p>

[peter_haagerup]

<p>Thanks for all the replies. I am learning a lot about film scanning :-)</p>

<p>It gives me headaches to figure out how to scan those (insert a very bad word) photos with reasonable quality. I am not expecting the film scans to look like a DSLR photo. Film is some grainy, quite low resolution dirty stuff unless you use really low ISO film and carry a tripod around all the time. Am I right?</p>

<p>Also, using a Olympus OM-10 is not easy. No autofocus, the light meter is freaking me out because I have to turn it on for almost every frame. The cat-eye focus plane in the viewfinder is probably not that precise, so a lot of my pictures are a little out of focus. In bright sunny weather, one could step down the aperture at f/8.0 or even more - pictures shot on a distance will look fine.</p>

<p>So, I admit that the sample picture I supplied in this thread is not a reference picture but I think of it as a "normal" picture. It could be sharper but it is not very unsharp. I still think that the scanner is blurring it up a lot. The professional scans you guys uploaded looks a little over sharpened to me, but it clearly shows up the grain and the sharpness is what I would expect for most of my own pictures. One thing that is stunning me is the colors of the pro scans. From a negative I am not getting anywhere near those colors. The dynamic range of my scanner is so bad that, when converting back to a positive (inverting and expanding the colors to get a contrasted picture) will produce a flat result even with 48bit processing. 24bit is unusable. Even with 48bit pictures a lot of sensor noise shows up. Of course, I do not normally convert my negatives to positives myself - I let Silverfast do it.</p>

<p>A dedicated film scanner is right now too expensive for me. This has come to a point where I think I might drop the hole idea of digitizing family pictures and my artistic work as well. It will be way too expensive and take way too long time. It is very, very sad. I have hundreds of negative strips just fading right know - waiting to be scanned :-(</p>

<p>I really hoped that the V700 was the solution. With that I could scan up to 24 pictures in one batch with... not so good quality, I guess.</p>

<p>Maybe you think I'm crazy - what about using a high quality DSLR, a high quality macro lens and some kind of slide/flimstrip holder in front of it with built in diffuse light source? A "scan" will take the time to autofocous and press the button. What about the quality? I don't know. At least there will be virtually no sensor noise. Negatives could be "scanned" by taking several pictures at different shutter speed and then combined in post-processing to eliminate flat colors and sensor noise. If that's possible, I could save up for maybe a Canon 5D Mk II and a good macro lens.</p>

<p>What do you think about that? :-) I might be crazy... or maybe it is a good idea? :-) Tell me what you think.</p>

[kelly_flanigan1]

<p>About every week on photo.net somebody starts a thread about flatbeds; this goes back a decade.<br /><br />Next week another will "discover" that a consumer flatbed is not as sharp as a 10,000 buck high end drum scanner.<br /><br />In like manner; somebody on knife.net will "discover" a 99 cent made in China 99 cent store knife is less sharp than a 200 dollar knife.<br /><br />Such is life.<br />Since photo.net started there have been many many many hundreds of threads about flatbeds.<br /><br />There will be another next week where somebody wants to scan at 6400 dpi with a 100 buck flatbed; and "discover" the image is abit soft.<br /><br />In like manner some Kilroy will buy a 250 watt 25 dollar speaker and "discover" that there is some "fluff" in the marketing.<br /><br />Flatbeds are great tools; they cost a TINY fraction compared top what us service bureaus paid for a high end pro units 12 to 15 years ago. In that era a pro flatbed was 600, 800 or 1200 dpi; and cost say 3000 dollars. Today a consumer flatbed pulls in say 1800 to 2400 dpi of info; costs 1/10 as much; scans 10 times faster; has a way better dmax range; and folks whine about the performance!<br /><br />It has been common knowledge and mentioned on many many many hundreds of threads there is some fluff/bs in scanner specs; with scanners "hawked" to amateurs.<br /><br />Amateurs seem to have issues with not having any goals; thus it is way harder to use a lessor tool. Flatbeds have been used to scan originals for 12x48 foot billboards 12 to 15 years ago; in pro usage.<br /><br />In *sister* threads many folks also ask "how big can I enlarge XYZ image" too; again with no goals; not viewing distance; the the peanut galley in wait preaches all sorts of goofy bizarre answers; often off by factors of 10, 100 etc.<br /><br />************Unless one has some goals ie defined prints and viewing distance; there is no logical way to say if a flatbeds scan is ok; overkill; or not enough.<br /><br />The same goes with a knife; are we cutting butter; carving a model; or doing surgery? Lay folks get *all* wrapped up in gooble gook.<br /><br />***Just take a super sharp set of negatives and have them high end scanned.<br /><br />***Then scan these with your flatbed.<br /><br />***Then make some prints with both; view them target distance; ***see if it really matters.***<br /><br />It really is not rocket science; a few simple tests only are required<br /><br />Here we got our first 35mm scanner back in 1989; and bought flatbeds several years later. I have owned about as dozen plus flatbeds; high end Microteks for 3000 bucks; and many Epsons; even low end Musteks and a moderate Canon once too. I LOVE epson flatbeds!<br /><br />With any tool; you want to understand what it will do; and its limits are too. One needs a better saw for furniture than building a doghouse or fence.<br /><br />ACME could market a 10,000 dpi 200 buck flatbed tomorrow; a 5HP shop vac that has a 18 awg cord; a 2 Lb boombox with D-cells that says 300 watts on the decal; a TV with a 1:10,000 contrast; a developer that makes tri-x be rated at 100,000 iso. Many folks love BIG NUMBERS; thus the buy stuff soley based on the BIG NUMBERS. It is wise to understand hokem versus real world results.<br /><br />If the boom box hard clips at 10 watts RMS ; it probably is just a 10 watt amp .<br /><br />IF the ACME flatbed seems; smells; feels like a 2000 dpi device; it probably is. It really doesnt matter if peanut gallery here says it is a 10,000 dpi optical device; or a 300 watt amp. If one measures the performance; then knows the results; often lower than fluffy claims.</p>

<p>Here is a scan of a 35mm tri-x negative shot in 1969; with as super sharp Konica Auto S2 at F5.6 . It was flatbed scanned here with out great Epson 2450 about 8? years ago when the units first came out. The sequence is full frame; and two cropped sections.<br /><br />This is a typical scan we get at a 2400 dpi setting. A scan with our dedicated Nikon and Canon film scanners pulls out more info; in the Argus wind knob's knurling is defined better. The shutter speeds by "argus" are defined. The Argus A2's exposure meter matrix computer is defined better.<br /><br />A flatbed pulls out alot of info; but *not* the entire cigar. With many customers originals and slides; it really does not matter since the original is not critically sharp anyway.<br /><br />This SAME original here is MY gold standard; it is the same negative I scan with every new scanner for the last 21 years.</p>

<p>The scan below is with the stock Epson 2450 holder; stock software; it was scanned with win2000 on a 200Mhz PPro with 512 megs of ram; all old stuff.<br /><br />The Epson 2450's flatbed scan at 2400 dpi is BETTER than our 1989 35mms scanner that cost as much as a new car!<br /><br /><img src="http://i4.photobucket.com/albums/y148/ektar/teletach/tripods-244.jpg?t=1264346093" alt="" /><br /><img src="http://i4.photobucket.com/albums/y148/ektar/teletach/tripods-189.jpg?t=1264343472" alt="" /><br /><img src="http://i4.photobucket.com/albums/y148/ektar/teletach/tripods-190.jpg?t=1264343602" alt="" /></p>

[kelly_flanigan1]

<p>RE" Maybe you think I'm crazy - what about using a high quality DSLR, a high quality macro lens and some kind of slide/flimstrip holder in front of it with built in diffuse light source? A "scan" will take the time to autofocous and press the button. What about the quality? I don't know. At least there will be virtually no sensor noise. Negatives could be "scanned" by taking several pictures at different shutter speed and then combined in post-processing to eliminate flat colors and sensor noise. If that's possible, I could save up for maybe a Canon 5D Mk II and a good macro lens."</p>

<p>At least with Nikon and digital slrs this method/concept goes back to the first digtal slrs in the early mid 1990's. Folks used/use a Nikon PB4 bellows and slide/negative copy attachment; and often a relay/macro lens like the 50mm F2.8 El Nikkor; or even a 50mm F2 normal. On really is doing the basic "slide copy" settup of a 1940's Exakta VX 35mm slr; but now one has a digital body. Slide copy rigs have the diffused piece that backlites the slide or negative. Some folks use natural light; others flash. A shot with a camera is digital still has noise. The camera method uses a 2D sensor; flatbeds and dedicated 35mm scanners usually have a 1D sensor array.</p>

[swilson]

<p>Kelly,</p>

<p>In your scan sample from the scaling it look like that the last crop is at around 1830 ppi, not 2400 ppi</p>

[peter_e]

<p>Mr. Haagerup: you are over-thinking this! You mention family pictures: if you follow my advice above (scan at high-rez, then down-rez and sharpen at least twice) you will get perfect results with a flat-bed to enlarge your family pictures to about 8"x12" size. But you also mention artistic uses: Flat bed will be fine to produce files for web posting. If you have fine-art large-format prints in mind you have to invest more. Alternatively, select your best 5-10 slides and have them professionally scanned and printed. I suppose you are not going to hang 50 large-format prints on your wall, right? But even then, depending on viewing distance and print size, film grain, lens used and many other factors, you will reach the limits of 35 mm slides sooner than later. Select slides for scanning based on their sharpness and contrast. Then start saving for a used DSLR. This is now the more economical and practical way of getting good digital images--I am not arguing against film and scanning here, but you'll have to spend more in time and money to get comparable results.</p>

[robert lee]

<blockquote>

<p>It is very, very sad. I have hundreds of negative strips just fading right know - waiting to be scanned</p>

</blockquote>

<p>Outsource. A <a href="http://www.scancafe.com/pricing/scanning">number of places </a> (that I've heard of, but haven't used) prices out to just about $0.30USD per frame.</p>

<p>Keep in mind also that the Nikon scanners don't depreciate much. Buy one used and resale when done. You can expect to get back the entire purchase price.</p>

[chrisnielsen]

<p>I recently upgraded from a Epson 4490 to a Coolscan IV that I got from KEH.com for $363. That's not horrendously expensive, at least in my book, and it's very sharp. Seems to resolve the grain as well, which is nice. Highly recommended!</p>

[peter_haagerup]

<p>Thanks on the info on the slide copy stuff. It is expensive as well, but at least it is a camera at the same time - that is a huge difference to me. Even if the results is not like a $$$$ film scanner, if it is like a good flatbed, it will be fine because it is very much faster. For family photos, it might be more than good enough. I have an idea that it could make very, very high quality "scans" if setup with care. I have some ideas why, that I want you "professional guys" to give your view on, if you like:</p>

<p>* The sensor is making a complete 2D picture at once. No moving parts, cheap optical lenses and glass platen to blur the image.</p>

<p>* The macro lens used can be stopped down to it's sharpest setting and focus can be done with precision - either manually or by using auto focus.</p>

<p>* High dynamic range can be achieved by taking several pictures, just like making HDR pictures. May give good results with negatives where the image information is narrow (low contrast) compared to a slide.</p>

<p>* Composing a picture of several exposures (HDR) will still give a sharp picture, while flatbed scanners have some trouble with this because of the not-so-precise motion of the sensor bar. (When using multi-exposure and similar in SilverFast, most of the pictures have slightly higher dynamic range and less noise but the result is always even more blurred than with a single scan)</p>

<p>The only thing I need to know is the price of a slide copy attachment and diffuse light source. And where to get it.</p>

<p>To the stuff about very expensive scanners and people buying stuff only reading numbers in the speclist, I can say that I do not buy stuff myself based on the specs. My greatest hobby besides photography is hifi. I own a $5000 turntable and a $1000 catridge - that's only part of my setup. Very, very expensive to me. But I use it almost every single day - and will continue to do so. My turntable has fairly good specs and that matters. What matters even more is the sound you actually hear.</p>

<p>To me it seems like false specs on most scanners. A scanner capable of scanning 3200 dpi should be sharp at 3200 dpi. Just like a 300w amplifier should play music with clarity and solid control when playing loud. The cheap amplifiers saying 300w in the specs are not even close to 300w. I guess that's the same with the scanners. It is a consumer thing I guess.</p>

<p>I am not going to make billboard size prints or anything like that. A4 is more realistic. A3 for really good photos maybe. I just like quality. It makes me take more photos. Just like a better hifi setup makes me listen to more music. That's just how it is :-)</p>

<p>If I was going to use a film scanner very, very often, I think it is OK for me to buy a expensive Nikon state-of-the-art film scanner. I will probably use a DLSR more often than a film scanner, so if the slide copy thing is worth trying out - I think that's the way to go.</p>

 

[kivivuori]

<p>Try different film holders and make also some scans without holders. I´ve found that holders that come with the flatbedscanners are sometimes too thick. 0,2 mm can make a huge difference in results.</p>

<p>BR<br>

Esa Kivivuori<br>

Finland<br>

http://esakivivuoriphotography.ning.com/</p>

[peter_haagerup]

<p>Thanks - I will try that :-)</p>

[leo_papandreou1]

<blockquote>

<p>I really hoped that the V700 was the solution. With that I could scan up to 24 pictures in one batch with... not so good quality, I guess.</p>

</blockquote>

<p>The V700 is a good scanner. Its output sharpens nicely in PS and it can sometimes pull more detail from slow film properly exposed in broad daylight than my DSLR is capable of recording on its sensor. You can usually make decent prints with V700 scans, if you don't print large. If most of your scans are destined for screen viewing then the V700 is A+. If you can't wait to afford a CS 5000, which really is the bomb, then a V700 plus Vuescan is adequate. Better than Costco or a slide attachment anyway. The V700 can also scan 120 and sheet film. With Doug Fisher's holder, which I have, 120 film + an Epson V700/750 is usually better than a DSLR. My DSLR anyway.</p>

<p>If you shoot 35mm rarely then you will rarely have pictures worth enlarging. It's just statistics. You can have the best scanned professionally and feed the rest to Canoscan. If you love and shoot 35mm intensively then it's just silly IMO to buy a flatbed rather than to save up for a Coolscan. If you're somewhere in between those two extremes, the V700 is ... adequate. While adequate is not great, most things aren't even adequate. The impression I get from reading your comments in this post, and it's just an impression, is that you should stay with the Canoscan. Buy a V700 only if the money's burning a hole in your pocket, which is the other impression I'm getting.</p>

 

[peter_haagerup]

<p>I did a test scan of a slide - without film holder (very hard to get the film flat against the glass), with film holder and a multi exposure with film holder.<br>

Looks like I did not have trouble with multi exposure this time. Looks as sharp (or unsharp) as the non-multiexposure. However, the scan without the film holder looks less sharp. Not by much, though. The optics in this scanner are so bad that it is almost impossible to see the difference ;-) hehe</p>

<p>This daylight picture is Fuji Sensia 200 - definitely shorter shutter speed and higher f-number than my first posted picture.</p><div></div>

[StuartMoxham]

<p>Here is you image with a a bit of sharpening some curve adjustments and the saturation lifted a bit.</p><div></div>

[StuartMoxham]

<p>If you were to print your 100% crops at 300 ppi they would appear in the print to be around 1.5 cm high. If you sharpen correctly you could get a reasonable 8x12 inch print from that scan without too much problem. When you view them a 100% it is rather like looking a a rather large print say between 70-100 cm.</p>

[mauro_franic]

<p>If you have sharp shots on film, a flatbed will throw away a lot of resolution, color, sharpness and add noise. If you only have a flatbed it is better than nothing. </p>

<p>Check out this comparison (the first crop is at 100% 4000dpi from medium format - obviously 3200 should look even better).<br>

http://www.photo.net/film-and-processing-forum/00SlJT</p>

<p> </p>

[sebastianmoran]

<p>Peter -- I was quite interested in exactly your question. Here's what I have found:</p>

 

<ol>

<li>My Epson V500, while an excellent bargain for scanning, delivers only 1500-2000 ppi resolution and the results with 35mm are not good enough in my opinion. I get decidedly better image quality from a DSLR, so I would use that instead of film-to-V500. Even with 120 film, I have trouble getting image quality from the V500 that is better than my DSLRs.</li>

<li>Side by side comparisons of 35mm film scanned with Nikon Coolscan V vs. Nikon D200 DSLR, it's about a toss up in my judgment. The results are slightly different, but no clear winner/loser. The film scan requires noise reduction and sharpening, and there is sharpening, of course, in the digital capture. I have excellent 8x12 prints from both. Color smoothness and dynamic range are better in the film. The D200 image looks "better;" it's probably a little crisper.</li>

<li>I did try using a DSLR as a scanner. The key question is to have a way of focusing accurately enough; looking through the viewfinder is not good enough for precision macro focusing. With live-view 10x for focusing, and a good macro lens, I got "scans" that are much better than from the V500. With 35mm slides or negs, non-flatness of the target will cause trouble. I have not compared this to the Coolscan V.</li>

</ol>

<p>I hope this is helpful. </p>

 

[janne_moren]

Peter, you mention that your focusing screen in the camera is off. It shouldn't be. In fact, it should normally be very precise; certainly good enough to nail the focus at f/2 or even wider open. If it's off, then there's something wrong. It could be that the mirror needs adjustment, but it could also be misaligned optics in the lens or any number of other issues that will end up giving you overall soft pictures as a result. I'd really have a camera place look over the camera before trying to figure out the scanning bit.

 

BTW, the V700 is no dedicated film scanner, but it does give me 35mm scans that are good enough for the web or for printing small. For medium format it's really good. But if you aren't going to use medium format, then a good dedicated 35mm film scanner doesn't cost any more than the V700 does.

[robert lee]

<blockquote>

<p>To me it seems like false specs on most scanners. A scanner capable of scanning 3200 dpi should be sharp at 3200 dpi. Just like a 300w amplifier should play music with clarity and solid control when playing loud. ... It is a consumer thing I guess.</p>

</blockquote>

<p>Pretty much.</p>

<p>I used to do some speaker building, where I experimented with open baffle woofer systems. Accuracy for designs of this sort is unmatched, but reasonable extension at adequate volume called for lots of amplifier muscle. It turned out that the truly pro use, sound reinforcement PA's from the likes of QSC and Crown delivered. No fancy metal work or enclosures, just products that actually reflect the spec.</p>

<p>It's like this too for film scanners. The Nikon pair, the 5000 and 9000, are the last of hobbyist accessible high quality scanners. It's worth considering if film is likely to be of enduring interest to you. After these are gone, the only thing left at the "low end" are the $20K Hasselblad Flextights. It's gonna be a sad day indeed, however inevitable, when Nikon finally exits the scanner market.</p>

[peter_haagerup]

<p>To me it really seems like 35mm is simply too small to get a good result with a normal flatbed. It is made for scanning paper, prints and maybe medium format and large format where the large magnification is not needed. On 35mm scans, the problems of the scanners really shows up. Even on the V700 I think, but it is clearly better than the Canon 8400F.</p>

<p>I am considering getting a slide copy system or making one myself. I need a good macro lens though. I think it is the best solution. Of course the limits are the sensor of the camera and the lens. With a high quality macro lens and even my Canon 400D I expect better results than the Canon scanner. With a very cheap lens, the results might be comparable I think.</p>

<p>I have seen a few photos scanned with this method - it looks very good. And it is fast - that is important as well. If the lens allows and care is taken to setup the system - I cannot see why it should not be sharp at 100% - even in the corners because the lens is stopped down.</p>

<p>I might get a 35mm IT8 calibration slide as well - to get everything just right.</p>

<p>I think it is worth a try because the lens can be used for normal photography as well. I really don't like to spend insane amounts of money on a slow, comparable quality dedicated film scanner.</p>

[roger_smith4]

<p>Peter, I'm not sure you can get an IT8 target to work well with a digital camera. It will be important to keep the light source, exposure and other settings identical from the time you make a profile to the time you make your "scans" or it will be invalid.</p>