Please lift me out of my ignorance - scanner resolution

<p>Say I buy a dedicated film scanner that says "xxxx optical dpi"and scan a film frame.</p>

<p>If I buy a flatbed that says also "xxxx optical dpi" and I scan that same film frame, with its film adapter, do I get the same resolution?</p>

<p>I suppose so, an inch is an inch, but maybe I am neglecting a technical detail?</p>

<p>Yes you should get the same size and resolution in the image file. Some users feel that the claimed resolution for flatbed scanners is somewhat optimistic, they may effectively only deliver around half of the maximum claimed in terms of actual image detail. They scan through glass and the focusing may be less accurate than a dedicated film scanner.</p>

<p>Well, not all flatbed scanners scan through glass. Those with glassless film holders inserted into the scanner body do not (like my cranky old Microtek 1800f.) That said, flatbed scanners are still a compromise, and only work really well with MF and LF film. Scans of 35mm film are disappointing and do not come up to the quality of dedicated film scanners. Those issues may or may not be related to max optical resolution of the machines...</p>

<p>Scanners only have a true optical resolution based on the pixel density of the sensor. Its an array that then moves up or down the scanner bed via a stepper motor. So when you see 6000x8000, the first (smaller) value is the optical resolution. The 2nd is the steeper motor. Since pixels are square, you can only get 6000x6000 (optical scan) OR 8000x8000 (interpolated data from the 6000 actual pixels). </p>

<p>The manufacture's rated resolution and the actual measured resolution are two different things. This is what causes the confusion. Here are links to 3 tests of various scanners. You can see that 2 are exaggerated and only one is close to its stated value.</p>

<p>Epson V600 - Stated Resolution 6400ppi - Measured Resolution 1560ppi.</p>

<p>http://www.filmscanner.info/en/EpsonPerfectionV600Photo.html</p>

<p>Nikon 5000 ED - Stated Resolution 4000ppi - Measured Resolution 3900ppi.</p>

<p>http://www.filmscanner.info/en/NikonSuperCoolscan5000ED.html</p>

<p>Plustik 7600i - Stated Resolution 7200ppi - Measured Resolution 3250ppi.</p>

<p>http://www.filmscanner.info/en/PlustekOpticFilm7600i.html</p>

<p> </p>

<p>I see...very interesting</p>

<blockquote>

<p>You can see that 2 are exaggerated and only one is close to its stated value.</p>

</blockquote>

<p>I'm sure the fact that the linked site sells scanning services has no impact on the objectivity of their results.</p>

<p>I've seen a number of supposed "experts" claiming that flatbed scanner resolution is much lower than the manufacturers' claims. So they're accusing the manufacturers of lying and false advertising. I'd love to see one of these experts on a witness stand repeat that nonsense under oath.</p>

<p>Scanner resolution is a function of sensor design. There are a certain number of sensor elements per inch. You can count them if you have a microscope. This is not some kind of smoke and mirrors (well, mirrors yes, but no smoke).</p>

<p>There are many factors that will influence the resolution of your results, but scanner dpi is not one of them.</p>

<p>Let's take a simple example, using a scanner that has 1000 sensor elements per inch.<br>

Suppose you put a target with exactly 1000 lines (500 line pairs) per inch on the bed, and align it such that each sensor element sees half of one line and half of its neighbor. The resulting scan will be a uniform gray, with no discernible detail. This does not mean that the 1000 dpi spec is wrong, it means the test methodology is flawed.</p>

<p>- Leigh (using an Epson V750)</p>

Leigh,

 

I agree. There are more than a few inaccuracies on the German's website.

 

If anyone has access to the test pattern negative he uses, I'll buy one and test, and post results. Ironically of all the

scanner stuff he sells, I could not find him selling that. Personally, I don't think he uses a negative at all, but does his

tests with a printed card of some type and tests using reflective scanning. Which I think is a false test.

 

His suggestion to scan at lower settings because there is no difference in the output, is absolutely wrong when I've

tested my new Epson V500. I will state this, the resolution of the V500(at highest settings) is greater than using my

Beseler 67 with EL Nikkor 80mm on Ilford paper. And the response curve/dynamic range is spot on to the negative,

which it is not with paper.

<p>Unfortunately even if the Epsons sensors have enough resolution the lenses in the v500 type scanners are not that great.<br>

Epson V600 - Stated Resolution 6400ppi - Measured Resolution 1560ppi.<br>

That seems to be about fair maybe a good example may resolve a little more but these are not 6400ppi scanners. They may produce files that are 6400ppi but the is just not 6400ppi of real detail in those scans.<br>

If you make scans with a v500 at 2400ppi you will get pretty good scans for smaller sized prints, but scans at 6400ppi just won't be that much better if at all.</p>

<p>Hi Richard,</p>

<p>That's a standard USAF 1951 resolution target, available from everybody.</p>

<p>But to adequately assess performance at the 4800+ dpi level requires a very high quality target indeed. I would certainly expect to use a deposited chrome on optical glass target for this test. They're available from Edmund Optics and other companies.</p>

<p>I've considered buying Edmund's scanner resolution target, which is made with that technology and goes down to 2-micrometer (12,700 dpi) target sizes IIRC. But it's $450, and I'm not that interested in proving to myself what I already know.</p>

<p>- Leigh</p>

<blockquote>

<p>... these are not 6400ppi scanners.</p>

</blockquote>

<p>Do they claim to be? Is that optical or interpolated resolution?</p>

<p>My V750 only claims 4800ppi optical. I find it hard to believe that Epson would bring out new products (the V7xx) with lower performance than previous (V600) models.</p>

<p>- Leigh</p>

<p>Yep the v500 claims to be 6400 ppi optical resolution. I threw my box away some time ago otherwise I would snap a shot of it.</p>

<p>One 6400pp scan from a v500 full image plus 100% crop. I don't what others think, for me it is pretty good for a flat bed but it likely is not as good as a Nikon coolscan.</p><div></div>

<p>Interesting.</p>

<p>What's the maximum size of an original at 6400 ppi?</p>

<p>- Leigh</p>

 

<p>8298 x 5141 was the size of that particular scan at 6400ppi. It varies from mounted slides to unmounted slides/negatives as more image is visible on an unmounted slide but there is not much in it. I think to that some brands of mount may mask more than others but 8298x5141 was the size of that scan.</p>

<p>Hi Stuart,</p>

<p>I meant the dimensions of the original. The V750 will do up to 4"x5" at high resolution. I was wondering if the V500 would do that size or if it was limited to smaller originals.</p>

<p>Thanks.</p>

<p>- Leigh</p>

<p> </p>

<p>Sorry just regular 35mm Fuji Chrome. The V500 only scans film up to 6 cm wide so 120 is the limit with a v500.</p>

As Andrew pointed out there is the true electronic or optical resolution based on the pixel pitch of the scanning bar.

So by that standard if the manufacturers are being honest, any 4000 ppi scanner is a 4000 ppi scanner.

 

However there is often a difference between resolution of the real detail in the film and what you get in the film. Just

as with a camera you get more resolved detail with any scanner when it is actually focused on the plane of the film's

emulsion. If you cannot focus the scanner by either adjusting the height of the film plane or of the scanning bar unless

you are very lucky you won't get the full amount of detail the scanner is capable of resolving. Focusing the scanner is

also complicated by the fact that most film has a curve to it, so ideally you want to either hold the film flat or focus

somewhere in the curve so that (hopefully) there is enough depth of field that the entirety of the frame you are

scanning is in adequate focus.

 

Without getting into the virtues of drum scanners or even the Hasselblad/Imacon Flextight virtual drum scanners there

are third party fluid scanning options made for various desktop flatbed scanners and Aztek made one for the now

discontinued Nikon Coolscan 8000 and 9000 scanners. These options let you set the optimal height and hold the film

flat. Epson also includes a fluid scanning tray with the V750 scanner.

<p>Hi<br>

Here you can find an interesting part of a Epson V750 flatbed scanner review, and one of the conclusions was that the best 35mm film scan seems to come out at 3200 dpi, well below the maximum indicated optical resolution:<br>

http://www.photo-i.co.uk/Reviews/interactive/Epson%20V750/page_4.htm<br>

(you can find this conclusion together with the opinion on definition. at the bottom §s of page 4)<br>

This review is quite complete and recognizes the advantage of dedicated scanners, but it also puts things in perspective, namely concerning prices and the intended uses for the scans.</p>

<p><strong> </strong></p>

<p>OK. That explains the difference. You need higher resolution for smaller originals.</p>

<p>Thanks.</p>

<p>- Leigh</p>

 

<blockquote>I've seen a number of supposed "experts" claiming that flatbed scanner resolution is much lower than the manufacturers' claims. So they're accusing the manufacturers of lying and false advertising. I'd love to see one of these experts on a witness stand repeat that nonsense under oath.</blockquote>

<p>I would like to see one objective measured testing result showing a flatbed scanner meeting the manufacturer's rated resolution. Please post a link to the testing results for all to see.</p>

 

<blockquote>

<p>Please post a link to the testing results for all to see.</p>

</blockquote>

<p>Ask the manufacturer. It's their specs you're challenging, not mine.</p>

<p>- Leigh</p>

 

<p>As you can see from the responses, there are other variables to a scanner's output quality than the resolution. A scanner is sort of like a camera; film size matters. Scanning a smaller negative requires much more resolution (since the image will be effectively 'enlarged' more on printing), and things like lens quality and film flatness become much more important. </p>

<p>The other thing to keep in mind is that since the image sensor produces many 500x500 dpi images, and then combines them (random numbers there folks), larger images will benefit more from this as well. If a negative is very small, the scanner's sensor may not be able to move itself into enough places to take a full XXXX dpi image without overlapping, and thus wasting pixels. This is why you could scan a print or 4x5 negative and get almost stated resolution, while 35mm negatives may look the same at stated resolutions as they do at half of stated resolution. For instance, I doubt you'll get much more detail out of a 110 negative at 4000 dpi than 1500 dpi for this reason, along with many others.</p>

<p>Lastly, flatbed scanners almost all scan at a single preset exposure, and digitally alter the exposure to provide 'brighter' or 'darker' scans. This can lead to shallower dynamic range, and more grain in shadow areas. Some dedicated scanners, and most drum scanners, can actually alter exposure time like an enlarger, giving you a MUCH higher-quality scan of even slightly under or over-exposed images.</p>

 

<blockquote>

<p ><a name="00Z06e"></a><a href="../photodb/user?user_id=6310086">Richard Sperry</a> <a href="../member-status-icons"><img title="Subscriber" src="../v3graphics/member-status-icons/sub1.gif" alt="" /><img title="Frequent poster" src="../v3graphics/member-status-icons/2rolls.gif" alt="" /></a>, Jul 06, 2011; 02:16 p.m.</p>

</blockquote>

 

<blockquote>

<p>His suggestion to scan at lower settings because there is no difference in the output, is absolutely wrong when I've tested my new Epson V500. I will state this, the resolution of the V500(at highest settings) is greater than using my Beseler 67 with EL Nikkor 80mm on Ilford paper. And the response curve/dynamic range is spot on to the negative, which it is not with paper.</p>

</blockquote>

<p>While this is true, I think you're leaving out some of the story. For starters, I find Ilford paper to be much more contrasty than other brands (I've switched to Kentmere, which is actually almost the same brand, but does expose differently), and that will influence your results. I also find - and my negatives tend to be very dark, so your milage may vary - that being able to have "true" exposure control, rather than a digitally edited single exposure, usually results in much better dynamic range from my negatives. Again, Kentmere paper, low contrast, and often different filters in different sections. I also use Moersch paper developers and I'm printing 6x6, so I really want to stress the 'your milage may vary' part :) Personally, I find that if I really work hard on it and waste a lot of paper, I can get much better tonal range with an enlarger than even with the Coolscans.</p>

<p>Another thing to remember is that most inkjet printers are extremely contrasty in the shadows, and like to clip details. Even if your <em>scan</em> has great dynamic range, that doesn't always mean that your <em>print</em> will have great dynamic range.</p>

<p>I'm in no way arguing with you Richard - just pointing out that depending on how your own personal negatives are shot and developed, you may get drastically different results. My guess is that your average shooter can probably expect results somewhere in between what you and I are getting with our negatives.</p>

 

Zack,

 

When I view a bare negative I can visually see detail in both the light and dark areas. When I wet print to get that

detail in both I need to use filters(usually split filtering, with dodging and burning) with Ilford papers. If I straight print for one, I

lose the other.

 

"Personally, I find that if I really work hard on it and waste a lot of paper..."

 

I concur. But I have to do neither with the V500. And it really does give me a good idea of what I need to do with a

negative to achieve the results that I intend.

 

Ive purchased Adorama paper(which some say is Kentmere) and will be testing that soon.

 

I have no intention or desire to ever inkjet or process print any scans. Film is only a hobby for me,

and when I do print negatives it will be to gelatin silver paper. I also have no intention of ever buying a $4000 used Nikon scanner, it's just something that will not ever happen.

 

 

In any regard, film flatness appears to be the limiting factor. Not the optics or resolution.

<p>Richard, have you experimented with other papers or developers? Your eyes and the negative are both capable of seeing a much broader dynamic range than either form of paper print can show without a lot of coercion. It's possible that if your paper is too contrasty, you'll lose shadow or highlight detail. Also your average variable contrast paper will have a little shorter tonal range than a graded paper, particularly in the shadow areas. To counteract this, I use a graded paper (usually #2) and make my prints a little bit lighter than I normally would want them. I use Moersch neutral developer, which develops a little more evenly across the tonal range than your average Dektol or the like. Lastly, I use a selenium toner to darken up shadow areas. It's sort of like the analog equivalent of shooting a low-contrast RAW file, and then bringing up the black levels.</p>

<p>Also, just to play devil's advocate ... if you don't print any scans, and you have no intention of printing them, how do you know if the range really is better than your darkroom prints?</p>

<p>Since film is just a hobby for you, I'd advise screwing around with some different stuff next time you trot out the camera. I don't know your system and I don't presume to, but I'd try some PanF+ 50 if you havent yet. The film can be really contrasty using standard developing methods, but the tonal range is absolutely phenominal, and it's silly sharp. I find that if you expose it at 25 ISO and develop it at 50, but for lower contrast, you get an even broader tonal range.</p>

<p>The only downside to this method is that it produces a very dense negative. With an enlarger that's no problem: you just print longer. But with a scanner, the denser negative may result in clipped highlights or shadows, unless it's one of those scanners that can vary exposure times. For what it's worth, the widest tonal range I've ever had was on PanF+, overexposed and developed as mentioned, and then printed with a Rodenstock 150mm at f/11. My printing time was a whopping 5 minutes, but that probably wouldn't have happened if I wasn't printing from a few feet away. It actually got up to about 20 minutes when I decided to vignette the edges and burn in a few areas :P</p>