Plane of Best Focus of Flatbed Scanners

[sanking]

The folks who supply fluid mounting kits make the point that there is a plane of best focus of consumer flatbed scanners, and that finding this exact plane is very important fir best results. And to that end they supply testing kits and shims to adjust the scanner to the plane of best focus. If anyone has actually used and found one the testing kits to be useful I would appreciate your comments.

Let me say from the start that I am a bit skeptical. In my own tests of two different flatbed scanners, an Epson 4870 and a Microtek 9800XL, using a resolution target capable of much more definition in lppm than either scanner, I found no difference in results in terms of lppm with the target placed faced down on the scanner glass, elevated 1mm above the scanner glass, and elevated 2mm above the scanner glass. In all three cases, when scanning at maximum optical resolution, I got resolution of about 21 lppm with the 9800XL, and just under 40 lppm with the 4870.

If your experience is different from mine, please let me know. I am truly interested in getting the most possible out of my scanners and if there is something faulty in the logic of the testing I have done I would like to know it.

Sandy King

[Doug Fisher]

Hi Sandy -

Even if your scanner's focus is on target at 1 mm, I am really surprised you didn't see any difference in the sharpness of your scans when you tested with a 2 mm range. 2 mm is a relatively large range. On my 4870, I didn't find optimal sharpness until at a total of 2.8 mm above the glass (+1.8 mm over standard). 2.8 mm is on the high end of the range though. Most people find their's closer to the 1.7-2 mm range. For some people the difference can be small although I am really surprised you cannot see a difference at all between the glass and 2 mm. The advantage isn't so noticable with smaller prints and larger film formats, but with smaller film formats and prints larger than 5x7, the benefit becomes more apparent.

Here are a couple of sites other than mine showing examples for dry mounted film:

http://www.normankoren.com/Tutorials/Epson_flatbeds.html#Focus_spacing

http://www.larry-bolch.com/ephemeral/4870-shims.htm

Finding the optimal focus point won't turn your scanner into a Nikon 9000 but it can help you get a bit more resolution from your flatbed. Hardware stores sell shimming material or your can buy a sheet of consistently thick plastic at the art supply store and just layer it to create your own shims.

Doug
---
www.BetterScanning.com

[sanking]

Doug,

Thanks for the links. I had seen the first of the sites, but not the second.

Well, I tested with the resolution targets and enlarged the file on screen until the file pixelated, which would have made it a huge print. With the Epson 4870 I was able to read about 40 lppm with the target on the glass, at 1mm over the glass and at 2mm over the glass, with the 9800XL 21 lppm at all three distances. Is there something flawed about this type of testing, or is it fundamentally different from the way you test? My logic is that resolution testing with a target of this type is a much more objective method of testing than what was done in the two sites you mentioned?

Or perhaps it just happens that the plane of best of best focus with both of these scanners is about 2mm over the glass, and the 1mm differene on each side is just not great enough to make much difference in resolution?

Let me say that I am also surprised by these results. I used to own an Epson 836XL that allowed physical focusing of the scanner over a range of 5mm, 2.5mm on each side of the center. With that scanner I definitley found a difference in resolution of fine details at the extremes. And I have a Leafscan 45 that is highly critical to the plane of best focus, though the Leaf is quite unlike the flatbeds in design.

Sandy






Hi Sandy -

Even if your scanner's focus is on target at 1 mm, I am really surprised you didn't see any difference in the sharpness of your scans when you tested with a 2 mm range. 2 mm is a relatively large range. On my 4870, I didn't find optimal sharpness until at a total of 2.8 mm above the glass (+1.8 mm over standard). 2.8 mm is on the high end of the range though. Most people find their's closer to the 1.7-2 mm range. For some people the difference can be small although I am really surprised you cannot see a difference at all between the glass and 2 mm. The advantage isn't so noticable with smaller prints and larger film formats, but with smaller film formats and prints larger than 5x7, the benefit becomes more apparent.

Here are a couple of sites other than mine showing examples for dry mounted film:

http://www.normankoren.com/Tutorials/Epson_flatbeds.html#Focus_spacing

http://www.larry-bolch.com/ephemeral/4870-shims.htm

Finding the optimal focus point won't turn your scanner into a Nikon 9000 but it can help you get a bit more resolution from your flatbed. Hardware stores sell shimming material or your can buy a sheet of consistently thick plastic at the art supply store and just layer it to create your own shims.

Doug
---
www.BetterScanning.com

[paulr]

My 4870 was off by quite a bit ... at least 1-1/2 mm.

It's not a huge deal. The depth of field of the thing is enough that I don't see the difference except in big enlargements of high res scans. But I'm happy to take advantage of the shims when I scan negs. It's a painless enough way to eek out a bit more quality.

[sanking]

My 4870 was off by quite a bit ... at least 1-1/2 mm.

It's not a huge deal. The depth of field of the thing is enough that I don't see the difference except in big enlargements of high res scans. But I'm happy to take advantage of the shims when I scan negs. It's a painless enough way to eek out a bit more quality.

Just curious. Do you know the focal length of the lens, and at what f/stop is it operating? I don't remember reading anything about this.

Sandy

[Walter Foscari]

I found no difference in results in terms of lppm with the target placed faced down on the scanner glass, elevated 1mm above the scanner glass, and elevated 2mm above the scanner glass
Went through a similar exercise a while ago with an epson 3200 and reached a similar conclusion. I used several different negatives, and had the results inspected by other viewers but was never able to determine that one position was better than another.

[Stan. L-B]

I have been using the new Epsom V750 Flat bed scanner for film formats and prints to 10X8 or A4. I have not found any need to alter the focus platform for any scan - as for me the final rersults are well up to expectations.

[Bob McCarthy]

I recently did a personal test. I found a 4 inch (10 cm) metal ruler that I propped up one edge 4 mm. The ruler had years of micro scratchs on the surface. The cm marks gave me a % scale to get the point of best focus. After scanning at high dpi (4000) I looked at the resulting scans. After much teeth nashing and some sharpening, I guessed the sharpest point was at the 3 cm reading on the ruler. So this implied 30% of 4mm or aprox 1.2 mm. But there was far more depth of field than I expected. With a little more sharpening, I'm not sure I could have found a definitive point.

Epson 4990 BTW

bob

[Kirk Gittings]

I have now owned 12 flatbeds over the years and currently have 4990(2), V750, 1800f and a Nikon 8000. I have done a ton of testing to try and optimize their performance.

Given how poor quality control has proven to be on some of the flatbeds, one cannot assume that they are in best focus right out of the box (some are). Since none of these entry level flatbeds have focusing and fairly limited DoF, finding the actual plane of focus is essential. Epson has acknowledged this by incorporating shims in their new scanners.Though I have not always done this with a test target, I have always found on all my flatbeds that there is a noticeable difference on small adjustments of the film plane, for instance, on the glass vs, in the film holder, emulsion up or down, small (gaffers tape thickness) incremental shiming of the film holder etc. There is no universal solution but must be figured out on each scanner. It is similar with scanning b&w in RGB to find the sharpest channel. The sharpest/noise free channel (blue or green) varies totally by brand but can vary within brands even.

[sanking]

Given how poor quality control has proven to be on some of the flatbeds, one cannot assume that they are in best focus right out of the box (some are). Since none of these entry level flatbeds have focusing and fairly limited DoF, finding the actual plane of focus is essential.

I have found a noticeable difference in resolution with targets placing them at different locations on the scanning surface, but very little with placement vertically from surface of the glass to + 2mm.

BTW, why do you say that flatbeds have limited DoF? It is my understanding that the lenses of flatbed scanners with CCD sensors such as the 2450, 3200, 4870 and 4990 have very great depth of field, and that the lack of sharpness observed is not due to variations in the plane of focus but due to, 1) the large area the lenses on these scanners have to cover and to, 2) the fact that the short focus lens is used at such a small aperture that it is diffraction limited.

Sandy King

[Kirk Gittings]

Maybe I am confusing terms. I say that because a 1-2mm difference in height of the negative can effect the sharpness of the scan. Other things being equal, what would it be but shallow depth of field? I can think of nothing else which would explain this. Though I have not played with this on the new V750, I have certainly found this to be true on earlier models. Epson acknowledges this on the new models by giving you spacers which can reset the distance above the glass from 2.5 to 3.5 mm. Also noted by: http://www.photo-i.co.uk/Reviews/interactive/Epson%20V700/page_13.htm

[sanking]

Maybe I am confusing terms. I say that because a 1-2mm difference in height of the negative can effect the sharpness of the scan. Other things being equal, what would it be but shallow depth of field? I can think of nothing else which would explain this. Though I have not played with this on the new V750, I have certainly found this to be true on earlier models. Epson acknowledges this on the new models by giving you spacers which can reset the distance above the glass from 2.5 to 3.5 mm. Also noted by: http://www.photo-i.co.uk/Reviews/interactive/Epson%20V700/page_13.htm

I don't know why Epson provides the spacers. It may me that the spacers help at certain *sweet* spots over the scanning surface. But spacing does not appear to improve resolution over the entire surface in my testing with the Epson 4870 and Microtek 9800XL. Spacers appear to improve resolution in some areas, but degrade it in others. I suspect that the improvement or degradation is due to the fact that the the lens is covering a wide area and by necessity light from some areas is reaching the lens aperture at very acute angles. I would speculate that this is potentially important because the net effect is to effectively increase or decrease the size or effective aperture of the lens opening, thus significantly varying the effects of diffraction.

In any event much of the literature on the subject suggests that the lenses in CCD sensor type scanners are short focus and are used at very small apertures which provide great depth of field, up to several cms. In fact, some people have even used them for scanning of 3-D objects. See some of the following sites.

http://www.normankoren.com/Tutorials/Epson_flatbeds.html

(See NLK's comments in response to email from Dennis Wilkins)

http://csdl2.computer.org/persagen/DLAbsToc.jsp?resourcePath=/dl/mags/cg/&toc=comp/mags/cg/2000/02/g2toc.xml&DOI=10.1109/38.824535

http://www.microscopy-uk.org.uk/mag/artapr01/dwscanner.html



Sandy King

[Kirk Gittings]

Sandy, Interesting information. I stand corrected on the depth of field reason for sure.


I would speculate that this is potentially important because the net effect is to effectively increase or decrease the size or effective aperture of the lens opening, thus significantly varying the effects of diffraction.

Fascinating.

One cras possibility is also that the spacers allow Epson to igore a level of QC by putting critical focusing back in the hands of the consumer.

[sanking]

Sandy, Interesting information. I stand corrected on the depth of field reason for sure.



Fascinating.

One cras possibility is also that the spacers allow Epson to igore a level of QC by putting critical focusing back in the hands of the consumer.


Kirk,

I am just speculating, not stating anything as a fact. However, there must be a reason why some people see an improvement in changing the position of the film plane, and others do not, and given the large DoF of these scanners, and the wide area covered by the lens, the diffraction concept seems a viable explanation. It would be nice if someone with a good engineering knowledge of how these scanners work would comment on the DoF question. I just don't know enough about the optics involved to make the calcuations to determine how much difference in resolution would result from changing the plane of focus 1-4mm.

Sandy King