What are the film scanner options in 2017?

[duff photographer]

Howdo,

Looking through the Collaborative Large Format Scanner Comparison http://www.largeformatphotography.info/scan-comparison/ I see that many of the film scanners are no longer made, available, or supported although most of the high-end scanners (Imacon [now Hasselblad], ICG, Heidelberg, Creo, etc.) are either still produced or have a relatively strong refurbishment market.

As it's been several years since the last scans were posted, what is the general concensus on the quality of those scanners that are still available compared with what's available today, if indeed there has been any significant advance in the technology and consequent output quality of scanners?

Cheers,
Duff.

[Jim Andrada]

I'm not aware of anything new that would be an improvement (or even be equivalent!) to the good old stuff. I suspect that DSLR "scanning", either stitched or one-shot is where it's all heading. For now it seems that parts etc are relatively available for my IQsmart and I have no complaint about how well it scans - only about how SLOWLY it scans thanks to really old internal compute capability. On the other hand, no big deal to load it up in the evening and let it scan away until morning.

[duff photographer]

I'm not aware of anything new that would be an improvement (or even be equivalent!) to the good old stuff. I suspect that DSLR "scanning", either stitched or one-shot is where it's all heading. For now it seems that parts etc are relatively available for my IQsmart and I have no complaint about how well it scans - only about how SLOWLY it scans thanks to really old internal compute capability. On the other hand, no big deal to load it up in the evening and let it scan away until morning.

Thanks Jim,

Interesting. I know a couple of pro printers who now use DSLR's to 'scan' film and I can't fault the quality of the final image for it's intended purpose.

Duff

[Peter De Smidt]

See post 42 in this thread: http://www.largeformatphotography.info/forum/showthread.php?106507-Recommend-a-Good-Flatbed-Scanner/page5

Dslr scanners can be very good, but building a good system is non-trivial.

[pendennis]

I have a number of negatives and transparencies in 35mm, MF, and LF which needed scanning. I know that scanners like the flat bed models will be a compromise, but I also have to consider costs. Drum scanners are outrageously expensive unless you can justify sufficient volume, say for business.

I bought an Epson V850 earlier this year, and I'm very happy with the results. I've started shooting film again, so it will get ongoing use. I also just bought a couple of 4x5 cameras, and I'm gearing up for them. One of the additional items I've purchased is a film holder from www.BetterScanning.com. These are outstanding additions to the Epson and work much better than those supplied.

[Greg]

For several years I had access to an Imacon Flextight Precision II at an educational institution just a 30 minute drive way. Scanned many of my B&W negatives and color transparencies but not all of them. Hoping to try to approach or even replicate the quality of the Imacon's scans at home, so put together the following set up:

Image capture: Nikon D4 FX body. True a series D800 Nikon would give me a lot more pixels, but the image files of the D4 by far exceed the resolution of my Platinum/Palladium final prints.

Equipment: a Nikon Multiphot. Probably the ultimate apparatus for photomacrography. Alignment of stage, lens, and camera back probably the best one could ever wish for.

Lens: 65mm Macro-Nikkor for 35mm and the 120mm Macro-Nikkor for larger formats. Shooting at #3 aperture in both cases. For larger than 4x5 negatives, I use a LED (lightbox) panel. Mine is a 13x16 inch "LED COPY BOARD Model: A3" and I have found its illumination to be completely even.

Illumination: For 4x5 and smaller formats, 1/2 of the exposure made with the collimated Multiphot's condenser light source and the other 1/2 of the exposure with a 100% diffused substituted LED light box. The condenser illumination caused an over-sharpening effect and it produced a slightly distorted reproduction of a step wedge. Using the 100% diffused light source produced a great range of tonalities but apparent sharpness of the image suffered a bit. Fred Picker, up there, is probably saying "I told you so". For large than 4x5 formats, I use a LED light panel and do ever so little sharpening in Photoshop using an Action which I've tweaked over the years.

Results... I actually preferred the images captured with the Multiphot over the images captured with the Imacon. I am not enlarging the images all that much in making the digital negatives and then contact printing Platinum/Palladium prints. With greater enlargements of the scans, the ones made with the Imacon, I would think that they should be a bit better, but I haven't experienced this so far.

To be fair, depending on the subject matter, shoot some film negatives with only using the collimated Multiphot's condenser light source, some with 1/2 of the exposure made with the collimated Multiphot's condenser light source and 1/2 of the exposure with a 100% diffused substituted LED light box, and then finally some with using only the 100% diffused LED light box.

Comparing scanning the negatives with my Epson V750 Pro with the final output being an 8x10 print... with 4x5 or larger negatives I couldn't see any difference. With 120 was a toss up. With 35mm the Epson scans were (subjectively) inferior.

Some points to consider:
Films shot with the Multiphot were laid flat on glass, emulsion side down, and held in place using painters masking tape to tension the film.

In the end, for producing digital image files from 8x10 and 11x14 negatives (4 scans with photo merge for 11x14), the Epson was just a whole lot easier to use.

comments welcome

[jim10219]

I use the single shot DSLR scanning method for 35mm work. I tried stitching them together from multiple shots, but in practice there isn't a noticeable different in quality, and a huge noticeable difference in the time and energy it takes to go that route. Though every once in a long while I'll still go that route (and only for E6 emulsions with super fine grain where it actually makes a difference). I use my ancient Epson 4990 for LF stuff. If I have a LF shot that I particularly love and want to blow it up huge, I'll send it off for a drum scan, but generally speaking my flatbed works well enough for 99% of the stuff I do. I've found the flatbed isn't sharp enough for the small 35mm format. And the DSLR is too hard to stitch together for LF. I usually wind up getting errors somewhere if I let the software stitch the image, or it takes forever and a day for me to do it by myself. I'd rather just ship it off for a drum scan at that point, rather than spend that a day or more trying to get one decent super high resolution file. I've thought about building an automated DSLR scanning rig, but it always seemed like a ton of work, time, and money that was just too hard to justify the gamble on. Maybe if I had a friend who built one or some other way to get my hands on one first and see if it's worth the investment before heading down that road, I might change my mind. But for now, I can't say I have any issues using a single shot DSLR (for 35mm), a flatbed scanner for general LF work, and a postage stamp for the LF stuff I want to blow up big.

I don't do MF stuff very often, so I don't have a set procedure for them. But one thing I will say is that getting a good process down is more important than getting the best gear. And it's not just in the physical scanning portion that's important, but also in the PS manipulation realm. Color correction, dust removal, and sharpening are all more art than science.

[Peter De Smidt]

If you have a repeatable positioning system, setting up a stitching template in PTGui works beautifully.

[bob carnie]

I am considering the phase one repro system for my future scanning needs of old photographs, Unique prints that I make, and possible film reproduction but I would have to test the final one against my Eversmart Supreme and Imocan

[Peter De Smidt]

That would be an interesting test, Bob!

[bob carnie]

I am thinking if I have the right lens the setup could be quite good, really expensive but over 20 years I can do enough scans to pay for the setup, the phase would be tethered to a workstation , probably capture one then on to storage drives.

What I like about this is the ability to sit down at a workstation like my big scanner and work. In my past life I was a photo comp specialist which required many hours doing copy work and we used a Linhoff camera copy stand with polarized light and it was a dream to use.

I am looking for a device that can measure LAB numbers on prints where the readout is immediate, I think this would be killer for dead nuts repro.

[duff photographer]

For several years I had access... ...comments welcome

Interesting read Greg - thanks.

...and thanks everyone else (and for the link Peter).

I've been considering a DSLR set-up. I have a spare Leitz Diaplan that can be converted into a copy stand although I've started making plans for a set-up that will accomodate larger film size (up to 8x10) using a pro' light box, X and Y stage, and a few other odds and sods cobbled together with the aid of a lathe and milling machine.

Otherwise, seeing little advance in flatbed scanners the last 10 years, I have been considering a drum scanner such as the ICG but only if I'm able to also do some business with it, but with people adapting DSLR's this may be a non-starter. I'm sure a properly calibrated drum scanner will produce the ultimate in quality scans, not least due to the wet mounting but I wonder how far off a decent DSLR set-up with, say, a betterscanning wet mount station properly set up, would be.

Duff.

[Peter De Smidt]

I wet mount my serious dslr and Cezanne scans, but for high max density slides, nothing beats a Tango. I wish I could afford one!

[Rich14]

Peter,

My D4000 and Tango produce equivalent results concerning high density slides. Using my Nikon D800E with a Micro Nikkor 105mm f/2.8D (at f/8) to "scan" 35mm, MF or LF (4 scans, stitched) is significantly better in deep shadow detail (as well as all other image parameters).

And much faster

Rich

[jim10219]

Seeing as how you can resolve down to the grain and beyond with a good lens on a DSLR, and the dynamic range of the newer bodies, I don't think even a drum scanner could give you better results. But like I said before, there's a process that you have to get down. I've yet to come up with a reliable stitching method. I can do 4, no problem. But 9 is sometimes dicey, and 16 almost always gives me fits to the point of being useless to attempt. That's why I still occasionally use drum scanning services. I need an automated system. I also need new tires on my car and other life expenses, so that's on hold for now.

[duff photographer]

Peter,

My D4000 and Tango produce equivalent results concerning high density slides. Using my Nikon D800E with a Micro Nikkor 105mm f/2.8D (at f/8) to "scan" 35mm, MF or LF (4 scans, stitched) is significantly better in deep shadow detail (as well as all other image parameters).

And much faster

Rich

Using the DSLR, I assume one would be able to make an exposure for highlights and another for shadows and 'HDR' merge the two digital images produced to pull out all the available information from the slide.

[Pali K]

Peter,

My D4000 and Tango produce equivalent results concerning high density slides. Using my Nikon D800E with a Micro Nikkor 105mm f/2.8D (at f/8) to "scan" 35mm, MF or LF (4 scans, stitched) is significantly better in deep shadow detail (as well as all other image parameters).

And much faster

RichRich,

I am curious to see any samples you have to share. My tango can see through the velvia 50 film rebate with extended range profile with no contamination of color. I don't know how anything can be better but I would love to see what others are doing and seeing.

I might have to try this test myself too. I do have a 5D MK II and macro lenses.

Pali

[Peter De Smidt]

Using the DSLR, I assume one would be able to make an exposure for highlights and another for shadows and 'HDR' merge the two digital images produced to pull out all the available information from the slide.

You can do that, and it works, but it's not usually needed. A D600/800 class camera has about 1 stop more dynamic range than an Epson scanner, as tested with a Stouffer calibrated step wedge. Some newer cameras will be even better.

[Pere Casals]

You can do that, and it works, but it's not usually needed. A D600/800 class camera has about 1 stop more dynamic range than an Epson scanner, as tested with a Stouffer calibrated step wedge. Some newer cameras will be even better.

Is this comparison made with Epson using Multiexposure ?

[Pali K]

So I did a test and I am going to take a strong position and say that NO, DSLR cannot outdo a Tango for shadow detail. I took multiple photos of a severely underexposed slide using Canon 5D MKII and even exposed it so the highlight of the slide were completely blown away and still, shadow detail is no match for Tango. Tango scan was a no fuss setup and made at 2000 DPI.

Here is a cell phone photo showing how severely this slide is underexposed compared to a properly exposed photo.

http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Underexposed%20Slide%20Test.jpg

Here are the DSLR vs. Tango comparisons and I picked the best for the DSLR out of multiple tests. Tango is literally the 1st test scan I made with auto settings and not even attempting to pull shadow detail to it's max ability.

http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/DSLR%20vs%20TANGO%20B.jpg

http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/DSLR%20vs%20TANGO.jpg

Unless someone is willing to post apples to apples comparison scans and show how a higher end DSLR can do more, I have a hard time believing that something can beat the PMTs and the 12 BIT Log AD convertors combo on a Tango. Its not just about the resolution or the spots of detail, the purity in the color and quality of the scan is just no match. If your Tango doesn't do this, I highly suggest having Karl tune it.

Pali

[Peter De Smidt]

Pali, I agree, Although a 5d mk ii is not a great camera for dynamic range.

[Pali K]

... a 5d mk ii is not a great camera for dynamic range.

Also in full agreement here :)

Pali

[Pali K]

This test made me scan an image that I never thought would be useful. Tango proved me wrong :)


https://farm5.staticflickr.com/4290/35502787650_a568d31741_h.jpg (https://flic.kr/p/W6g4Ss)
Underexposed Slide Test (https://flic.kr/p/W6g4Ss) by Pali K (https://www.flickr.com/photos/palikalsi/), on Flickr

Sinar Norma 4x5 | Fuji Velvia 100
Heidelberg Tango PMT Drumscanner

Pali

[Pere Casals]

So I did a test and I am going to take a strong position and say that NO, DSLR cannot outdo a Tango for shadow detail...


Hello Pali,

I don't doubt that the Tango is a way better scanning system than a DSLR.

Anyway your test shows a very good result fron the DSLR source. It is true that the DSLR has way more stray light (Tango has near zero stray light...), but this signal level can be removed with PS.

IMHO the DSLR performed much better than I would guess:

With stray light:

167200



Just removing stray light level:


167201



With a quick color balance:

167202



IMHO a better match can be done with a proper LUT.


Regards

[Rich14]

Rich,

I am curious to see any samples you have to share. My tango can see through the velvia 50 film rebate with extended range profile with no contamination of color. I don't know how anything can be better but I would love to see what others are doing and seeing.

I might have to try this test myself too. I do have a 5D MK II and macro lenses.

Pali

Pali,

I posted messages about using a D800E for scanning a number of months ago and wondered when someone was going to ask for a comparison between such scans and a drum scanner.

I've though about this quite a bit. I am confident that I am seeing better scans with the D800E than with scans that were made over the years with a wide variety of scanners, including Tango, my current D4000 and other assorted machines. I have scanned the same Kodachromes, Velvia and Ektachromes with the D800E.

I no longer have access to the Tango. But I still fire up the D4000 from time to time.

While I know that my results are valid, I don't think that my posting my results will be proof to anybody. My drum scan technique could be faulted. There could be any number of criticisms that I'm not showing the drum scans at their best.

So here's what I suggest - send me a difficult chrome (or a number of chromes) that you have scanned on a Tango, or any other drum scanner. Post the image(s) here. I'll scan them with my D800E and post my results. (Of course I'll send your film back). We'll be able to compare the D800E to not only a Tango other than the one I used, but the capabilities of different operators.

It should be interesting, and will be done openly on the forum.

I'm not particularly interested in sending my film out as my collection now is all personal stuff, and I have no question about the ability of the D800E.

Rich

PM me if you want to get in touch.

[Pali K]

That's a great idea Rich!

I'll send you a PM and will plan to send you this very exact 4x5 Velvia sheet. I'll also make some proper scans from a Epson V700, Eversmart Pro, Scanmate 11000, and Tango that I have access to for comparative purposes.

Pali

[stefan dinu]

Hey guys, do you think 600USD is a reasonable price for a Fuji Lanovia C-550? I am thinking about an upgrade from V700. Also I was curious in terms of speed in Fuji will be an upgrade on this regard.

[Pere Casals]

While I know that my results are valid, I don't think that my posting my results will be proof to anybody...


Why not ?

Just scan a USAF 1951 resolution target slide and post the scan.

https://en.wikipedia.org/wiki/1951_USAF_resolution_test_chart

If group 6 element 2 bars are seen then your system is outresolving LF lenses, so enough for LF sheets. If you see Group 7 elements then you have a first class system...


Here you have the result from an X5 (Not a drum, but close, This is scanning 35mm):


http://www.filmscanner.info/Bilder/UsafHasselbladFlextightX5.gif


You can compare to that.

Other practical tests can be done with an IT8 slide target, for density, etc, Also you can measure stray light.

https://www.flickr.com/photos/timparkin/26276532831/in/faves-125592977@N05/



IMHO numeric tests do not explain all but a lot is explained...

[bob carnie]

That's a great idea Rich!

I'll send you a PM and will plan to send you this very exact 4x5 Velvia sheet. I'll also make some proper scans from a Epson V700, Eversmart Pro, Scanmate 11000, and Tango that I have access to for comparative purposes.

Pali

I would contribute by scanning on a Eversmart Supreme and a Imocan if someone handles the dynamics of sending the original around.

[Pali K]

I would contribute by scanning on a Eversmart Supreme and a Imocan if someone handles the dynamics of sending the original around.

Bob, I can certainly send you this particular 4x5 Velvia image but does it make sense for us to better organize this effort? Does anyone have access to the original images that were used for the scanner comparison here for a more meaningful 2017 update to these tests that were done many years ago?

Pali

[Rich14]

Why not ?

Just scan a USAF 1951 resolution target slide and post the scan.

https://en.wikipedia.org/wiki/1951_USAF_resolution_test_chart

If group 6 element 2 bars are seen then your system is outresolving LF lenses, so enough for LF sheets. If you see Group 7 elements then you have a first class system...


Here you have the result from an X5 (Not a drum, but close, This is scanning 35mm):


http://www.filmscanner.info/Bilder/UsafHasselbladFlextightX5.gif


You can compare to that.

Other practical tests can be done with an IT8 slide target, for density, etc, Also you can measure stray light.

https://www.flickr.com/photos/timparkin/26276532831/in/faves-125592977@N05/



IMHO numeric tests do not explain all but a lot is explained...


Fair enough. The last AF test target scans I made on several D4000s and a Tango resulted in a resolution figure a little better than about 4300 dpi. I forget the formula for translating between the Air Force target and DPI. Anyone have that?

I'll do a D800E scan of my AF target this weekend and post the results.

I'll also scan an IT8 target. I'm not sure how much that would satisfy the question. It's a reference target and using its calibration file will certainly produce a color reference frame for a particular machine, but I don't think individual machines can be compared to each other by looking at their images of different IT8 targets. Someone set me straight if my logic is off here.

FWIW, producing scanner ICC profiles from an IT8 target never really helped me with any scanner I've used. It always resulted in crushed blacks, despite the Don Hutchinson "trick" of covering the darkest patches on the target with metal foil tape to make them absolutely opaque. Scanning without a scanner profile and subsequent color correcting on a calibrated monitor, using the original transparency transilluminated by daylight/calibrated light source, or by comparing to known reference images on screen was always much better.

[Pere Casals]

Fair enough. the formula for translating between the Air Force target and DPI. Anyone have that?

Here http://www.filmscanner.info/en/Aufloesung.html you have these tables, no conversion needed, bottom table says dpi from Group and Element:

167247

This is the common (I'd say controversial) way to tell "scanner optical dpi" an scanner resolves. Here we can have some controversy because we can say we need 2 points to resolve a line or we can say we need way less than 2 points, depending on the modulation transfer % we consider.





I'll do a D800E scan of my AF target this weekend and post the results.

I'll also scan an IT8 target. I'm not sure how much that would satisfy the question. It's a reference target and using its calibration file will certainly produce a color reference frame for a particular machine, but I don't think individual machines can be compared to each other by looking at their images of different IT8 targets. Someone set me straight if my logic is off here.

FWIW, producing scanner ICC profiles from an IT8 target never really helped me with any scanner I've used. It always resulted in crushed blacks, despite the Don Hutchinson "trick" of covering the darkest patches on the target with metal foil tape to make them absolutely opaque. Scanning without a scanner profile and subsequent color correcting on a calibrated monitor, using the original transparency transilluminated by daylight/calibrated light source, or by comparing to known reference images on screen was always much better.

One important thing is measuring the amount of stray light, this is a transparent slide with a little black patch, compared with a totally black slide.

I about color... scanner and DSLR sensors may have different spectral responses, but difference can be narrow after using a conversion LUT.

I'd like to see your tests, thanks in advance.

[Rich14]

Here is the detail of the USAF resolution chart (sorry for the dirt in the scan - I didn't clean the film):

167307

The full size, original file is here for viewing or downloading.

https://www.dropbox.com/s/wu9354fcu1k3lnp/USAF.jpg?dl=0

This is a jpeg processed from the RAW NEF file in ACR and Photoshop. No adjustments or sharpening of any kind. The D800E's white balance was probably set to that of the LED tablet on which the target was scanned.

I can read to Group 5, element 6, maybe a partial read of element 7. That puts the resolution in the neighborhood of 5000 ppi. Since the camera is focused at 1:1 and the pixel density is 4900 across the 1 inch dimension, resolution can, of course be no higher than that. Suffice to say the camera and lens are are delivering well above 4000 ppi, and in the neighborhood of a 5000 ppi "scan." I have never seen any usable image data from commercially available film stocks above about 3500 - 4000 ppi no matter what users of machines with so-called scan resolutions of 8000 ppi or 11,000 ppi claim. So the D800E/Micro Nikkor f/2.8 AF D (shooting aperture 3 stops down from wide open) is pulling everything that can be seen on film at this reproduction ratio as far as resolution in concerned. The shooting aperture was set 3 stops down from wide open. That's usually f/8. At the 1:1 reproduction ratio, that is actually f/13, which the camera reported. I call it f/8 for simplicity as that's where I set the lens for all reproduction ratios as that's where it performs best, whatever the f/ratio truly is mathematically.

Here's a Kodak IT8 target scan.

167308

And here's the link to the full-sized jpeg for viewing or downloading.

https://www.dropbox.com/s/rbsuh2w0krkjyth/IT8.jpg?dl=0

This is a scan of the full 4x5 transparency as one shot with the D800E. I usually scan 4x5s in four slightly overlapping quadrants and stitch them together via ACR and Photoshop. I did scan the IT6 target that way, but the resulting files gave Photoshop's stitching algorithm fits in putting the four pieces together. This is the first time I have run into this kind of problem. It really surprised me, but the problem became obvious after a few retries at large and smaller scan sizes, trying to change the overlap which is what I first thought was the issue. It has nothing at all to do with the D800E's capability for scanning.

First, Photoshop completely removed the gray border around the central area of color patches, cropping the image to just the color patch region.

Second it assembled the four images out of order.

My take is that the algorithm saw the uniform border as being superfluous overscan outside the desired "image." Then it couldn't make heads nor tails out of the four quadrants as there is no unifying "image theme" like there is in images of the "natural world." I assume some other kind of test chart might have fared better, but the semi-random color patch data is not something the algorithm is made for. I have never experienced difficulty stitching the quadrants of a 4x5 image of the natural world. I doubt I ever will.

If anyone wants to see the individual quadrants, I'll give links to them, but they give no more information than the image I've posted as far as the use of the IT8.

The posted image is about 1250 ppi. If I could have correctly assembled the quadrants, the resulting image would have been 2500 ppi. I've never had occasion to scan a 4x5 at higher resolution. There is no additional resolution in a 4x5 image above about 2000 ppi. If super large prints are needed (above about 50 inches in the short dimension) upscaling a 2000 ppi scan gives identical results to scanning a 4x5 at 4000 ppi (which takes a long time on any drum scanner).

The D800E's white balance was set to that of the LED tablet. I adjusted camera exposure until I got flashing "blinkies" in the white patches. No adjustments of any kind have been made in ACR or Photoshop. No sharpening. The NEF file was simply sampled from 16 bit to 8 bit and saved as a JPEG.

When I get the test image in the mail I'll post my scan.

Please let me know if there are any problems with the links. I will leave the files up for the foreseeable future.

Response to this will be interesting.

Rich

[Pere Casals]

I can read to Group 5, element 6,


You have to read the table in the other direction... 5/6 is the column that has the "5" header in the top, so it is around 2900 dpi (instead 5000), not bad anyway !!!!

Thanks for posting this interesting test.

The IT8 test shows an amount of stray light, as lighter colors (pastel) have more white. This can be solved in part in post process, calibration wont solve it because different scenes will deliver diferent amounts of stray light.

My interpretation of the test is that it is a viable system for LF, in special for BW, the single issue, IMHO, is that one has to remove the stray light level in PS, and perhaps it won't work very vell for the case of high densities present in some slides, and in few negatives.

[Rich14]

Sorry for misreading the chart. The Lasersoft target is labeled differently. I should have realized there is no way the lens and sensor working together could give a result of close to 5000 ppi since the sensor's pixel density is itself 4900 photo sites per inch. Theoretically the system should produce about half of that figure. But resolution is not an issue since it could be increased simply by using a higher magnification ratio. I can't do that with the Micro Nikkor 105 without adding an extension tube. Easy enough. Then there would be a stitching step for 35mm scans as well as for larger formats. Also trivial.

As it is, the 1:1 set up is sufficient for all but gargantuan prints from the format. Using a Pentax 645Z, Hasselblad X1D or Fujifilm GFX all of which contain the 50mp Sony sensor of equal or better performance than their 36mp D800 sensor would remedy that in a single step.

I'm not sure what you're referring to as "stray light" in the shot of the IT8 target. The only light entering the camera is that passing through the transparency which is completely masked all around on the LED light table. The scan was done in a relatively dim room. The image does look lighter in the highlights (less contrast) than the target looks visually. But that's the point of using a calibrated target to generate an (ICC) profile of an image system. We expect nonlinearity. That's what the profile corrects (mostly).

Do you mean that the image is being "self-contaminated" by light because the camera sees the entire area of the transparency at once, while a drum scanner sees only an individual scan spot at a time?

The real test is going to be comparing a (corrected) scan from my D800E against that done on a Tango by someone else.

Rich

[Pere Casals]

Do you mean that the image is being "self-contaminated" by light because the camera sees the entire area of the transparency at once, while a drum scanner sees only an individual scan spot at a time?

Rich


Yes, this is. A lens self-generates parasite (stray) light. This depends a lot on the number of optical groups and on the coating performance.

The amount of parasite light also depends on the particular negative (or slide) you scan. A very "transparent" slide will trow a lot of light to the lens, and a little share of it will be dispersed. If you have also deep shadows in the slide this contamination will make a difference, and you'll need to manage that in the digital processing.


A flatbed only illuminates a row at a time, so there is less chance that a lot of stray light arrives to the sensor, also the Micro 105 is a good lens generating limited amounts of stray light, but as you see a big area this worsens the thing in front of a flatbed.

...and a drum is near free from stray light, as a single point is illuminated and read.


You can measure the stray light amount by placing an small opaque patch in the middle of the negative, that reading in the opaque patch is the amount of the stray light, you can compare with a reading made with illumination closed.

Of course to measure that you have to take completely raw images, without any in-camera correction.

[Rich14]

Understood. Yes that's an issue. Both theoretical and practical.

But in my completely unscientific testing, the path through the camera's exposure compensation, inherent in-camera RAW settings and ACR/Photoshop "post" processing produce results equal to or better than my own D4000 or Tango scans on screen or printed on an Epson 7890.

I originally tried using some old 150mm and 300mm repro process lenses. As well as my 50mm f/2.8 EL Nikkors (both original and "newer" designs). I expected the process lenses to give excellent 1:1 performance. But nothing works as well as the Micro Nikkor. I believe a big part of the problem with the others was flare in the extension tube and bellows setups. I may be able to improve on that but the Micro Nikkor is just so easy to use.

I imagine there are better lenses for this purpose, but this one is hard to beat.

[Peter De Smidt]

A 75mm Rodagon D, the 1x version, works very well for this.

[Pere Casals]

But in my completely unscientific testing, the path through the camera's exposure compensation, inherent in-camera RAW settings and ACR/Photoshop "post" processing produce results equal to or better than my own D4000 or Tango scans on screen or printed on an Epson 7890.


I agree a lot that accurate post processing may be very important, IMHO sometimes it is more important than the pure scanner performance.

Sometimes drums deliver better results simply beacuse a good operator is on-board.

Real shooting conditions are field conditions, not lab conditions, this means that the "on film" actually resolved lp/mm are far from lens maker specifications: those are ultimate performance in ideal conditions, with perfect alignment, perfect focus, optimal aperture, no vibration and a contrasty subject.

TMX, a sharp film, resolves 200 lp/mm for 1:1000 (micro)contrast, but for common contrast you can find in textures it may resolve just 50 lp/mm at extintion. What I'm suggesting is that at the end most times a very good scanner will make not a great difference, simply because the inferior machine is able to extract most of on film information yet.

IMHO there are some shots that deserve a good scanning machine, but a number of factors are required: A technically perfect shot, with a very sharp lens, at optimal aperture, with a sharp film, and lots of microcontrast, and a big print.

If it is not the case the most important part is edition, using the r¡ght sharpening and the right PS downsizing algorithms. A perfectly sharp image can look bad simply because web browser resizing.

This is what I personally concluded... but I'm still a learner, and sharpness is a really complex concept, more complex than it looks, IMHO.

For MF and 35mm a very good scanner may also help to depict film grain more naturally... but this is also a complex thing :)

Pere

[Peter De Smidt]

Rodagon D 75mm f/4 at 1x magnification with a D600. Raw developed with Capture One with no sharpening. Image cropped at 100% view in Photoshop.

https://www.dropbox.com/s/ysiy8zg6enz7blw/Rodagon_D_1x_D600_no_sharpening.jpg?raw=1

[Pali K]

That's very clean Peter - thanks for posting. Out of curiosity, is your DSLR able to see through the lines? On my tests, the lines are light grey as the scanner sees right through them.

Does anyone wish to see similar tests from Drum Scanners? I ask because the resolution is far greater and the test is silly to compare in my opinion.

Pali

[Pere Casals]

That's very clean Peter - thanks for posting. Out of curiosity, is your DSLR able to see through the lines? On my tests, the lines are light grey as the scanner sees right through them.

Does anyone wish to see similar tests from Drum Scanners? I ask because the resolution is far greater and the test is silly to compare in my opinion.

Pali

Yes... it would be interesting. X5 flex can see Group 7 if doing 35mm, but a lot less with larger formats. It would be interesting to see where drums arrive...

[Peter De Smidt]

That's very clean Peter - thanks for posting. Out of curiosity, is your DSLR able to see through the lines? On my tests, the lines are light grey as the scanner sees right through them.

Pali

Hi Pali, I used Capture One's 'auto' tone curve and no exposure adjustment. Looking at the raw file, when the whites are at 250 the dark bars are at 33 with the linear tone curve. This is with Edmund's chrome on glass high resolution target.

[Rich14]

I've rescanned the USAF target after realizing my camera was not focusing accurately. I have been using Autofocus and the D800E's autofocus fine tune is on -20, the end of its range. Even at this setting, the image was not in focus. If there were a -25, that would probably be spot on.

I manually focused the following image which is a slight improvement over my previous. In my version of the USAF target, group 5 corresponds to group 7 in Pere's and Peter's. Here I can read all of that group and perhaps the first element of the next group.

Peter's Rodagon image certainly looks good.

It looks like I will need to send my camera and lens together to Nikon for them to adjust the autofocus fine tune for this lens. It can be done for newer lenses, don't know about this one. Which leaves me to using manual focus for scanning. Not really a problem.

Here's the full image:

167353

And the detail showing group 5:

167354

Rich

[Peter De Smidt]

Hey Rich,

At 1x magnification, using the lens to 'focus' changes the magnification instead. It's better to set the lens for 1x magnification, and then moving the whole camera + Lens system forward and back until the target is in perfect focus. I use a quality mm ruler to set the lens focus. Match the ruler reading to the size of your sensor, which'll give you 1x, and then don't change it.

Same thing as before, except this is with a reversed 50mm Componon-S at 2x magnification:
https://www.dropbox.com/s/7qmow9nrlhcmgxs/Componon_s_2x.jpg?raw=1

[Rich14]

Hey Rich,

At 1x magnification, using the lens to 'focus' changes the magnification instead. It's better to set the lens for 1x magnification, and then moving the whole camera + Lens system forward and back until the target is in perfect focus. I use a quality mm ruler to set the lens focus. Match the ruler reading to the size of your sensor, which'll give you 1x, and then don't change it.

Peter, the exact magnification ratio is seldom important (at least to me). Focusing the lens itself is the only smooth focusing mechanism I have with my current set up. I don't have the camera on a focusing rail.

I know magnification changes slightly with the way I'm doing it. I had thought of adapting my view camera to hold the D800E body and making a lens board for the Micro Nikkor. But even my Sinar P back can't focus as smoothly as the helical focusing collar of the Nikkor.

[Peter De Smidt]

My Z-axis: https://www.dropbox.com/s/t3jwny820r3szvd/Z_axis.jpg?raw=1

I'd be surprised if any auto focus has the precision and accuracy to focus ideally at these magnifications. But I'd love to be wrong.

[Rich14]

Wow I think I have to have one of those!

That and a usable live image. (D800E leaves a bit to be desired in the live focus department).

Does the Radagon have a focusing collar? I do have a helicoid adapter for F mount which I got for testing the various lenses I mentioned previously.

[Peter De Smidt]

The camera and extension tubes are bolted to an Arca slide. Thus there is no helical in the system. The lens extension is set to focus exactly at 1x, which is ideal for my lens. With my D600, live view is not accurate enough to find perfect focus. I have no idea why. During setup, I do a series of exposures at different distances from the target, find the best one in my raw processor, and then fix everything at that point. The vertical stage is a Velmex 4000 series Unislide. A D800e should be able to get about 800 dpi more with the same lens setup than with my D600. (Daniel had a D800e and ran tests with it in a similar setup.) How much test slide resolution is practically helpful is a difficult question to answer. My guess is that many people overestimate how much detail they are actually getting one film. I scanned a purposely shot and highly detailed 35mm tech pan negative. Shot on a very heavy tripod with mirror lockup at my lens's best aperture. Such a setup should resolve significantly more detail than any LF negative. Doing test scans, I found that going up i, magnification (with lenses optimized for that) lead to better test slide results, but I couldn't see any advantage with the 35mm Technical Pan negative, which means that going after better results than I get with 1x with the Rodagon doesn't appear worth the effort, especially with larger film. My point is that since you're getting such great results with your current system, don't let test slide scans push you to spend more time and money chasing a dubiously useful improvement.

[Pere Casals]

Same thing as before, except this is with a reversed 50mm Componon-S at 2x magnification...


This is some 5000 or 6000 dpi, impressive



I scanned a purposely shot and highly detailed 35mm tech pan negative. Shot on a very heavy tripod with mirror lockup at my lens's best aperture. Such a setup should resolve significantly more detail than any LF negative. Doing test scans, I found that going up i, magnification (with lenses optimized for that) lead to better test slide results, but I couldn't see any advantage with the 35mm Technical Pan negative, which means that going after better results than I get with 1x with the Rodagon doesn't appear worth the effort, especially with larger film. My point is that since you're getting such great results with your current system, don't let test slide scans push you to spend more time and money chasing a dubiously useful improvement.


I also think this is a very good way to know what performance is needed or not, a good TP shot with a good 35mm glass is a safe reference...

[Rich14]

EL Nikkor 50mm f/2.8 (2nd generation) at f/8 unreversed:

Full frame

167366

Detail

167367

I had previously thought the Micro Nikkor 105 was doing a better job than the El Nikkor. But the EL Nikkor now looks a tad sharper and more contrasty at 1:1.

I'm still focusing by moving the lens (on a helicoid mount between the lens and the extension tube). So the size "breathes" a little as I focus in and out. I will eventually need to make a proper "copy stand" or several which I'll lock down at correct focus for 1:1 and for scanning 4x5 in four pieces.

That is, if I can be sure the camera truly locks down accurately in the same place with the tripod mount. Don't know how reliable that might be. Its tolerance may be too great. Cant leave the D800E locked down as a dedicated scanner!

Rich

[Rich14]

I can't remember if I've posted the following article on this forum before. I know I posted it on another forum. I came across it some months ago. It's attribution and URL is in the first lines. The site is no longer active and seems to go off line frequently.

It was originally posted in 2011 I believe, before the D800E and its Sony sensor, or any of the more recent sensors and DSLRs, mirrorless and MF DSLRs were available.

It was amazingly prescient then about this entire topic of Drum scanner vs modern digital sensor scanning of film and the dilemma for drum scanner users re the inevitable loss of all avenues of repair-replacement of equipment in the (now long) absence of any manufacturing. It articulates the situation far better than I have. It is spot on about the superiority of ACR/Photoshop over any scanning software ever made.

Another issue is this: There still exists large quantities of film archives that need scanning. For that purpose, of course we need the equipment. It is a great sense of relief to me that, rather than fear the day when my scanner ceases to work, and I have no way to remedy that, there exits a technology that is not only more than up to the task now, but is still accelerating in its development and capabilities.

BUT - what about future film? Some of us see hope that the rapid extinction phase is over and that not only will there be some stable small niche of film users and manufacturers but that there appears to be a (tiny) renaissance of sorts with older photographers coming back "to the fold" and younger members joining the ranks of film users. Kodak (or "new Kodak") is offering the best color negative material they have ever made and TMax and Tri-X are alive and pretty well. Ilford is still around and other small European manufacturers continue.

On the other hand, Fuji appears to be going the other way. Velvia 50 (4x5 and 8x10) cannot be obtained other than knowing someone in Japan who can buy it there and ship it. And soon LF Acros, a marvelous film will go the same way. Is the market really too small and Fuji's withdrawal an indication of further manufacturer collapse?

See the article in the next message, as the total is too long for a single post. Remember, this was about 6 years ago!

Rich

[Rich14]

(From previous message - an article posted about 6 years ago)

http://theagnosticprint.net/future-of-scanning/

Camera Scanning and Raw Software
Walker Blackwell

What’s happening?

Only a few years back, most companies stopped making film scanners. The game was up; scanning equipment no longer pulled a profit. The future was in digital capture and the world marched on without looking back. It left educational institutions in a pickle. They were already reeling from high silver costs and a change in photo curriculum. Suddenly companies stopped updating software for Intel chips and repair service on older scanners dropped like a stone. It put pressure on the education world to go all digital and those ripples were felt everywhere. For us high-end scanning labs that were outside of the educational “prosumer” world, we fared ok although young clients were coming to us with 8 megapixel files instead of 40 megapixel (equivalent) 6×7 film. We already went through this in the early 2000s when companies stopped making drum scanners. Over the years, most of us learned enough about our various drum scanning machines to fix the beasts ourselves. Third party service vendors, mostly past employees of the very corporations that build the crazy things in the first place, took care of the rest. But recently the support has slipped and it gets harder every month to maintain high-end equipment. Today, prosumer scanners are taking the same track but at a more accelerated pace; everyone feels the heat including professional photographers who prefer 35mm film.

A little history

In the mid 90s to early 2000s a boom in the drum scanning business helped propel the magazine and advertising world into the digital age. Companies like Esko-ScanView and Howtek began building machines small enough to sit on a desk. These futuristic contraptions could scan every piece of silver grain from Kodak Tech-Pan 25 and TMX100 film at a resolving power as small as 3 microns. They sold for sixty thousand dollars and up and only large service bureaus could handle that overhead. Small independent printmakers couldn’t. Mostly, we weren’t in the digital game yet so we didn’t care or notice.

But in short order the competition got stiff for the drum scanner manufacturers. Then the economy turned bad, and digital capture in the advertising world went through the roof with new medium format backs hitting 16 megapixels. In just a few years the drum scanning business silently imploded leaving only a few survivors and no capacity for new scanner sales. The used market boomed with the advent of eBay and drum scanners could be bought for one tenth their original cost.

I got into the game in late 2005 when I realized that all the artists I printed for needed high quality scans. They couldn’t afford a drum scanner on their own and didn’t want to spend the time operating one if they had the money; but they weren’t getting good service at their labs either. None of the large repro shops with drum scanners new how to do ultra-high resolution work for artists. They set the variables wrong and my own output quality suffered because I was printing crap files. I didn’t know a lick about drum scanning but my goals were high and I knew I could follow through. Eventually Nathan Baker (my business partner) and I settled on the Scanmate 11000 drum scanner. I learned so much about scanning in such a short amount of time. Nathan went even further and now fixes the things. At a resolution of 11,000dpi it remains one of the top scanners of all time. It uses a mechanical interpolation by multi-sampling individual over-lapping parts of the negative. It actually scans with a physical resolution of something around 5000 or 6000 dpi but with this mechanical stepping-motor interpolation, the resolution it can really pull is much higher. I describe this because the very same thing will apply in the near future of digital scanning when drum scanners become obsolete.

The machine we use at Black Point Editions was born in 1996. At 15+ years of age, it still does its job remarkably well. It pumps out nine-hundred megabyte (and up) scan files day-in and day-out. The files it produces are still too large for the most contemporary computers to edit quickly today! Talk about ahead of its time. When the Nikon 9000s and all the rest of the prosumer scanning equipment stopped being built or serviced in 2008, us printmakers new the feeling of loss felt by our fellow photographers. The scanners we depend on are important to us; and they are for individual photographers as well. We digital printmakers have lived in limbo trying to keep our scanners operational for as long as possible in a world less and less equipped to repair them. Now everyone will be in the same boat. For the individual photographers, this environment pushes them to buy dSLRs and many aren’t happy with that mainly due to a relative decrease in dynamic range in even 5D Mark 2s compared to color neg film.

But there is one world were the lack of prosumer scanners and increasing lack of drum scanners could have really hit hard: the institutional archive. What will we do with all of our millions and billions of culturally important photographs if we can’t scan the film? Much of that film was printed to dye paper that will degrade under light. Do they just languish in some dank office somewhere? We haven’t scratched the surface of our film heritage in this country. A large part of the reason for not getting our digitized house in order is that every film scanner made to date runs as slow as a toad! (That is true of scan backs fyi.) It takes too damn long to sit and wait ten minutes for a high resolution scan. The labor involved with the flimsy, bendy, all-around-lame, film holders “invented” by the various scanning manufacturers just stops massive digitization projects in their tracks. Institutions and individuals combined have slid back and bought extremely low quality “fast” ccd flatbed scanners resulting in millions of perfectly blurry thumbnails scattering the playing-field of image history.*Now we run the risk of not even having the “good” hardware (clunky as it was) to get the job done.

What to do?

There is an old workflow that we are forgetting about: remember inter-negs and slide duping? That’s right, the dSLR has finally taken the stage of the prosumer (and in the last year even highend) scanning world. The “camera scanner” has arrived! Most people don’t know it yet, but the capabilities of a dSLR go (or will go) way beyond a CCD scanner and someday soon even a PMT (photo multiplier tube) drum scanner. The secret sauce is the interaction between raw cameras and raw editing software.

When Apple came out with the iPhone everyone went stupid over how good the hardware looked. But the killer feature was how well the operating system worked with that hardware. The fact that we see multiple YouTube videos of babies working iPhones and just somehow knowing what to do, is proof. Well, the scanning world might be due for a similar treatment all-be-it in a slightly more professional manner. With the soon demise of Nikon 8000 and 9000 scanners and the high sticker price of Imacon/Hasselblad CCD scanners (the static in them just sucks doesn’t it?), a good Canon 5D Mark II holds its own. Couple that with Lightroom/Bibble 5/CaptureOne and we have a temporary winning combo for 35mm scanning.

Let’s face it. Anyone who has done a color negative scan using normal scan applications has stumbled across the fact that our scan input software isolates us behind a vast mote, far away from the workflow we use to visualize and print those files. It forces us to quickly interpret the preview image in some way we think we want, scan the file, and then fix (or rescan) the image later if we don’t like it. VueScan, Silverfast and Flexcolor have attempted to fix the problem by giving us invert/edit ability in their most recent applications. But those edits are not saved to XMP and are proprietary code. Also, they don’t work with RAW files from Cameras. Well, if we scan directly into Lightroom/Bibble5/CaptureOne as RAW, we don’t have to bother with that separation. Our scanning software is our visualization software (and what a visualization software it is); we can always revert to the raw negative any time we want and rebuild from scratch. Not only that, but our development settings aren’t linked to some “negative profile” saved in a proprietary hunk of software from 2004. Those settings are embedded inside each file in readable XMP format. They are portable and reversible and can stand unique to every scan. Sounds logical right?

As a professional photographer, you most likely already have a 5D or equivalent twenty megapixel CMOS camera. That is good. You can use that. All you need is a prime 100mm macro lens; put the camera on an old-school strobe interneg base like the Bowens Illumitran, and you can do a 3500dpi scan in less than ten seconds. Not only that, but the scan is RAW meaning future raw editors will get out more detail from the scan in future releases. All the things you can do with a raw file apply with this workflow including mutli-frame real HDR scanning giving you the ability to scan massively contrasty and dense silver film.

I’ve recently built such a system for the University of Vermont Slide Library. Where MIT spent many “cutting edge” years hand-scanning every slide in their archive, we hope to digitize our 150,000 slides in under three months (pre metadata inclusion) with only $3400 in equipment. That is all made possible by the full frame pro dSLR.

That idea applies to the personal photographer as well. You can scan negatives (both color and silver) with the 5D and interpret them directly in your raw editor of choice. These applications let you invert and “develop” negatives (with a few nudges I might add). The develop settings can be saved and applied at the time of capture: meaning you can place a negative in front of the 5D and if the scene contrast, exposure, developing, and film type are all the same from one negative to another, the live-tethered image will show up in seconds ready to interpret in Photoshop without any extra work! Couple this with the fact that you are using a “non-destructive” editing system, XMP metadata support, not to mention a fantastic image database (in the case of Lightroom), and I would say this combo takes the blue ribbon. For the photographer looking to scan their backlog of 35mm film but unable to find a decent CCD scanner at budget and not wanting to waste time waiting for a CCD scanner to “rrrrrr” its way through every frame, I think this workflow will catch on. One plus for Bibble 5 and CaptureOne: you can do full Linear interpretations of the raw file with inverse curves and levels. This lets you get the right gamma characteristic of the color neg film and keeps your color luminance and saturation in check. Over-all, I would say Bibble/CaptureOne have a better color/tonal rendition. If Lightroom adds Curves, Levels, and Linear support in Version 4 (these are top feature requests btw) it may take the hat.

The Future

The logical ones among you are probably saying, “3500dpi? Really? Professional scanning hardware pulls 8000dpi for 35mm. Why would I backtrack?” For us high-end scanning companies, a new player arrived this year that might change the game: the Hasselblad H4D-200MS. It sports a two-hundred megapixel 645 camera back. To get this resolution, it captures six photographs in sequence all at fifty megapixels. For each photograph in the sequence, minuscule piezoelectric motors shift the entire sensor a small amount. This results in a mechanical interpolation of the primary fifty megapixel image to a two-hundred megapixel whole while retaining full color, low noise, and zero moire. This method resembles the old drum scanners we use today that are interpolating the physical resolution of 5,000dpi up to 11,000dpi by just moving the drum motor slower. The result from the new Hasselblad is a pixel dimension almost that of a drum scan of an 8×10 film sheet (an 8×10 at 2000dpi is 320 megapixels but suffers from some physical defects of the scanning hardware.).*The question is whether a 645 sensor could actually pack the necessary sensor element density required to capture grain detail from an 8×10 film sheet and if the software used to build the 200 megapixel image uses that resolving power in the right way. We have the glass but the jury is out on the sensor. (I would need the $47,000.00 camera system to test with.) It looks like the technology may be hitting drum-scanner quality at 4×5 film sizes. And it couldn’t be at a better time as the machines we depend on are starting to decline and degrade. If we can simply place a 4×5 sheet of film under a camera, push a button, and see that film digitized in seconds at a print quality of 50×60 inches, well good riddance to waiting. The magic will live in how we backlight the film and interpret the raw data when it arrives on screen.

While the drum scanner isn’t dead yet (they are selling for a rock bottom price between $300 and $4200 on eBay right now) we may see more and more labs going to camera scanners and not looking back. Gone are the costly consumables in fluid and clear Mylar. Gone is the static electricity and dust of the Imacon days. Gone is the clunky software used to invert (and generally mess up) color negatives. Gone are the hundreds of variables to keep track of: white balance strips, gear lubrication, PMT calibration, dark and white noise, bulb life, micro aperture/focus, crazing. Now we can find a winning combo of settings, save that, and let the digitizing begin. The future of film scanning may have just got a little brighter if not temporarily more expensive. For 35mm and 120mm film, the Hasselblad camera scanner is most likely already better than drum scanners. For 4×5 and 8×10? We’ll see.

[Peter De Smidt]

Gigamacro, Phase One Cultural Heritage, and others have scanners aimed at these institutional customers.

[Rich14]

Peter,

- Hiding in plain sight -

One of my process lenses is a Rodenstock Magnagon APO 105mm f/5.6 (marked 103mm). Otherwise known as Rodagon. I had poor results with it before, but that's because I didn't have an adequate extension tube or fine focusing capability. I'll post results later.

I also have an Eskofot Ultragon (love the names) 150mm f/9 which I also think is an APO. I don't have the means to focus this lens on the D800E (yet). I think this is a job for my view camera.

Rich

[Peter De Smidt]

I've got a Magnagon, too. Just a second....

https://www.dropbox.com/s/gjwjxoyz2mj4212/magnagon.JPG?raw=1

It looks closest to the Rodagon D f/4.5, which is optimized for 2:1 or 1:2.

[Rich14]

Digging around in the lens adapter, flange, etc collection . . .

I've mounted the 105mm and 150mm lenses on Linhof 96x99mm boards. They now snap onto my Sinar P 4x5 front standard lens board. (Gosh, with those lenses usable on the 4x5, I'll just have to get a Sinar shutter!)

I've got a Nikon F mount male flange which I'll epoxy onto a suitable aluminum plate to fit the rear standard film holder/film-digital-back clamp and I'll be in business with the D800E on the Sinar.

I'm getting dangerous now. The sky's the limit . . .

[Peter De Smidt]

Alignment and freedom from vibration are critical.

[Rich14]

I've tried testing the 105 Magnagon using my extension tubes. It requires two full sets of tubes. And the whole setup is too ungainly. I can't focus accurately enough with this rig. Things will have to wait until I can mount the D800E on the view camera. I will do that in the next few days, but my day job rears its ugly head for the next several, and that's that.

A good part of the focusing difficulty is trying to focus at f/5.6. It's significantly more difficult than at f/2.8. I don't want to even think about testing the f/9 150.

Remembering operating a process camera, when focusing was necessary - the camera's movements were set up to handle that pretty "automatically - we used the "parallax" method, focusing through a clear central area of the film back. Which, of course was very bright and didn't involve trying to judge the quality of the blurred/unblurred image at all.

In any case, in multiple tests that wound up just in front or just in back of the true focal plane, it's apparent that even when I'm able to focus accurately, the result of the APO lens is not going to be much better than what I've already got. We're at the point of rapidly diminishing returns. It looks like I might get one more element out of the USAF chart.

The 50mm EL Nikkor seems to be giving the best performance so far by a whisker. And that's good for anyone who wants to set up such a system. These great lenses can be had for about $30.

As a "proof of concept," I am more than happy with the results I've been able to show.

Still waiting for Pali's transparency. I'll scan it as soon as I get it.

Rich

[Pere Casals]

The 50mm EL Nikkor seems to be giving the best performance so far by a whisker.

A reversed 50 EL (f/2.8 version) is a top performer for macro photographers, in special for flat jobs, in the 1x to 2x range.

Then there is the reversed Componon 28mm f/4, an amazing lens for up to 5:1, 5x. (The 35mm is easier to find, for up to 4x)

As said previously, perhaps there is no need for such high dpi from 5x. ...but it would be amazing making one x5 scan, just to know what +10000 dpi is, with lots of stitching :)

[Peter De Smidt]

Well, if I get really bored next winter, I'll do a scan with a Nikon 5x Measuring Microscope lens.

[Pere Casals]

Well, if I get really bored next winter, I'll do a scan with a Nikon 5x Measuring Microscope lens.

The "measuring" Nikon lens can be amazing !!

[Rich14]

Hello everyone,

This is my D800E scan of a Velvia transparency that Pali sent me. A lot of explanation follows. Don’t judge the scan too much from the display jpeg. I have no idea how much compression the forum software is applying to the file nor how much it’s been mangled by the deviantart.com site where I’ve parked it simply for display here.

http://img11.deviantart.net/9169/i/2017/217/f/5/panorama_by_rich14-dbiz1rw.jpg

All my original files are available for download. The Dropbox link is at the end of this (multipart) message.

The purpose of this post is to compare my D800E’s abilities to drum scanners, and specifically to a Tango scan which Pali will submit. I have said that my D800E produces scans with equal or better dynamic range compared to my Howtek D4000 drum scanner and to scans that I have done on a Tango and on other drum scanners. I don’t have access to those scanners now, only the scan files from my own transparencies which I did several years ago and extending back more than 15 years. Re-scanning those transparencies with my D800E, I can improve on every file, no matter which drum scanner I used at the time.

But showing those differences shouldn’t convince anyone that the D800E is the equal of a drum scanner. It just shows my own capabilities or shortcomings in operating a number of different machines. Presenting this scan vs attempts by others using their equipment and the same test transparency is a better comparison.

I am very eager to see how other people and equipment can handle Pali’s Velvia 4x5.

First some comments about the 4x5.

It is a very dark image. We agreed on that as an ultimate test of the DMax of whatever machine is doing the scanning. This Velvia 4x5 had deep blacks as dense as they get. Did I mention it’s a very dark image.

It’s a great test image for DMax capability, but understand, this image is a very unrealistic one to use for quality reproduction. It will challenge any scanner and operator to set up the scan parameters, black point, white point, other software settings. And at best, if this transparency were presented to a commercial scanning service, they would be well within their responsibility to tell the photographer that their efforts would have to be recognized as an effort to “rescue” the image as best they could and that all bets were off as to how well the image would reproduce when printed, either by a digital printer or by commercial printing methods.

Over the years, I have had the pleasure to work with many pro photographers who understood the needs of high end magazine color reproduction. It is a fact of life that the dynamic range of the original subject matter captured on transparency material cannot exceed 5 stops when the highest quality of magazine reproduction is the goal. 5 Stops original subject contrast.

Think Vogue magazine covers, or Architectural Digest. If the subject contrast is not controlled to fall within that 5 stop range, the deepest shadows or the brightest highlights can not be held. Period. The real pros, using color reversal material in advertising and editorial photography know that well, and bring back the bacon every time.

We always advised that examining a transparency by reflected light off a white piece of paper was the only way to anticipate how the image would reproduce on the printed page (any kind of printing method). Looking at the slide by transmitted light (on a light table) resulted in a gorgeous image, dripping with dynamic range and contrast, but was a completely unrealistic expectation of what could be pulled out of the image by scanning and printing.

I had many (uninformed) photographers insist that because they could see brilliant detail in their (underexposed) slides by holding them up to a hot light, that the detail should be reproducible. Nope. Look at the slide by reflected light. That’s what it will look like in reproduction.

And looking at an underexposed slide by reflected light, usually results in a useless, dark, murky image. Sorry. Welcome to the laws of optics and physics.

Pali’s 4x5 doesn’t just look impossibly dark if examined by light reflected off a white surface (black cat in coal bin dark!) - it looks impossibly dark examined on my light table, with a 4x5 mask blocking all extraneous light!

Examined with a hot light and a loupe I can see great image detail down in the very darkest parts of the image. But can I dig that stuff out?

More caution: The image itself represents a scene that should not be completely reproducible in a one-shot effort. Essentially the scene contains a light source itself (open sky) and objects illuminated by that light source (trees, water, ground) and deep, deep shadows. It is not normally possible to hold detail all the way through the dynamic range of a scene which includes the light source which illuminates the scene. Such scenes can have a range of luminance exceeding 20 stops. Did I mention, it has deep, deep shadows. Inky black. Deep.

(continues)

[Rich14]

(continued)

Excuses:

I had some technical problems at my end when I first tried scanning this image. First, you’ll notice flare at both sides of this scan, right side greater than left. That problem led me down a rat hole of tearing my scanning setup apart to find the source. I tried scanning with a number of lenses, leading me to discover that one APO process lens I have (f/5.6 105mm Magnogon-R, which I believe is a Rodenstock) is toast because of strange bubbling and pitting of the front and rear surfaces, resulting in terrible lens flare. I don’t know what environment that lens was used in. Nasty.

I have several other APOs including a 210mm Rodenstock Ultragon made for Eskofot process cameras. This lens is giving me superb results, but I still don’t have a reliable way to use it. I will have all that solved when I get my Nikon F mount camera adapter finished for my Sinar 4x5. I also have a 305mm Ultragon which also should be superb but is way too big and heavy to use on anything smaller than a 20x24 process camera. I haven’t had one of them for 20 years.

I never was able to solve the problem of the two areas of “flare” on this image. I do not get such flare from any other 4x5 scan that I do with my equipment. Is the flare in the image, caused by Pali’s camera? I don’t see it in the image. Maybe it’s there and only shows up when scanning? I don’t know. This has me stumped at the moment.

I have mentioned that my Micro Nikkor 105 mm f/2.8D lens does not correctly focus using the camera’s autofocus. I became aware of this earlier in this thread, and Pali’s 4x5 really made me notice it. I can’t correct the problem with the in-camera fine tune settings of the autofocus. I need to send the camera and lens to Nikon for pairing and calibration.

So I’ve been learning to use the camera’s “live image” video capability to focus. Nikon’s implementation of live focus is not very good on the D800 series. I understand it’s better on the 810 and expectation is that it will be much better in future cameras. The image at highest magnification is very grainy. That combined with the fact that the slightest vibration of my set up is also greatly magnified (the image dances around with the slightest touch of the focusing collar) and the fact that the focusing collar in manual focusing mode is sloppy and loose (a common affliction of AF lenses) made learning to use this technique hard.

But the image is coming right off the sensor. No proxy here, no reflex mirror or prisms in the path. Live view offers the best theoretical way to focus the image. It took a while, but I learned how to do it and can now focus extremely accurately. I have to say that using live view is the only way to really nail focus in macro mode.

So here’s my image as well as I can scan and process it. I had anticipated scanning both as a one-shot and using multiple shot HDR technique to increase the scan dynamic range (just for fun). But I didn’t need to. As you can see if you look at the Photoshop file, the image histogram completely fits in the camera’s dynamic range with room to spare at both ends. Extending the dynamic range really would not have improved anything. The image is captured in 14 bits (16 bit file), so there’s plenty of “head room” to lift the shadows in software.

The photoshop file has two adjustment layers. One to lift shadows a little. The other is a high pass filter to “output sharpen” for the monitor. You can turn either layer off to see the difference. To print this file needs a different output sharpening layer. This is a pretty big file, 1.5G.

The four quadrant Nikon NEF raw files are available to anyone who wants to see the original scan files and stitch them into one image. If you download them, make sure to also download their associated xmp sidecar files. These will allow you to see my ACR settings. I applied no exposure adjustment, just brought the files in as shot. Custom white balance was set in the camera off the LED light table. I applied “capture sharpening” by a combination of the Clarity setting to 15 and in the sharpening tab, set Amount from its default 25 to 75, Radius from default 1.0 to 1.5 and Detail from default 25 to 50.

I consider all the software settings, including the curve applied in Photoshop and the capture sharpening settings as the equivalent of analogous hardware and software settings that would be applied with any particular drum scanner.

The jpeg in the folder is simply sampled down from the very large Photoshop file for display. No attempt was made to optimize it in any way, so don’t rely on its image for any critical purpose.

I apologize for dust in the scan. I didn't want to clean the 4x5 too aggressively.

The color balance on my calibrated monitor is as close as possible to the visual appearance of the 4x5 viewed by daylight.

(Dropbox link in next post)

[Peter De Smidt]

Good stuff, Rich. Thanks for doing the test! Daniel might also be willing to scan the slide with his D810 at 1x magnification. One huge advantage of this method over, say, a consumer flat bed is that it's easy to adjust how much light is hitting the sensor.

[Rich14]

Here's the Dropbox link to the folder, "Velvia"

BTW, the vignetting is not from the scan. It's in the 4x5 image.

https://www.dropbox.com/sh/flir6e4j7oxnlca/AABpFAztTRvfaZQEw9q9ljyPa?dl=0

Rich

[Sasquatchian]

I don't know. I downloaded the psd file and to my eye, which includes nearly twenty years of drum scanning, it does not look very good. The shadow lift curve does make the shadows lighter, but doesn't really produce any more usable shadow detail and, as you can clearly see, mostly washes out the deepest blacks. The blue channel is clipped or so close to being so that it might as well be. There seems to be some weird grain texture very visible at the edge of the waterline that looks like bumps in the emulsion but are probably the effect of not fluid mounting, which, btw, would have drastically reduced the overwhelming amount of dust and dirt in this photo of a piece of film. Of course, it's also hard to know when the Bayer filtration will come into play and give you less than desired color by simply not having all the color information there to begin with. What I'd like to see is the same piece of film scanned by a good operator on a Howtek or ICG at 4000 ppi. And finally, the resolution on this "scan", as it were, when image size is adjusted down to about 4.75 inches across, is a bit under 2400 ppi. Okay, but not nearly what the film with a good lens and good technique will support. While this is undoubtedly much better than an Epson scan, it's not up to what I'm used to seeing from a well adjusted Howtek on a daily basis.

[Pali K]

Thank you for doing this test Rich. I am not sure what is causing the flare on your end. I don't see it on my scan and haven't had any issues with my camera that was used for this image.

Here is a link to my unadjusted tiff file from NewColor V2. As I mentioned to you via email, this scan I made was a quick n dirty dry-mount scan and I'll do a proper wet mount scan soon.
Tango Quick Shadow Test Scan (http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/ShadowDetailTest3-Rotated.tif)

BTW, I am happy to send this to others who want to test it using different scanners. I know that Bob offered to test it using ES Supreme and Imacon and I'll be happy to send this to him next. I also have an ES Pro with me so I'll make a scan using it for comparison purposes as well.

Pali

[Rich14]

I don't know. I downloaded the psd file and to my eye, which includes nearly twenty years of drum scanning, it does not look very good. The shadow lift curve does make the shadows lighter, but doesn't really produce any more usable shadow detail and, as you can clearly see, mostly washes out the deepest blacks. The blue channel is clipped or so close to being so that it might as well be. There seems to be some weird grain texture very visible at the edge of the waterline that looks like bumps in the emulsion but are probably the effect of not fluid mounting, which, btw, would have drastically reduced the overwhelming amount of dust and dirt in this photo of a piece of film. Of course, it's also hard to know when the Bayer filtration will come into play and give you less than desired color by simply not having all the color information there to begin with. What I'd like to see is the same piece of film scanned by a good operator on a Howtek or ICG at 4000 ppi. And finally, the resolution on this "scan", as it were, when image size is adjusted down to about 4.75 inches across, is a bit under 2400 ppi. Okay, but not nearly what the film with a good lens and good technique will support. While this is undoubtedly much better than an Epson scan, it's not up to what I'm used to seeing from a well adjusted Howtek on a daily basis.


Sasquatchian,

You raised some good points -

Re scan resolution. I was switching between a number of lenses. When I made the final scan, the camera was a bit further from the light table than I usually capture a 4x5 because I just left it at the position for the previous lens, a 105mm Magnogon and scanned with the Micro Nikkor 105. That resulted in much more overlap of the four quadrants than usual.

Even on the drum scanner, I seldom scan a 4x5 at 4000 dpi. There simply isn't much additional resolution on even a very good 4x5 transparency much above 2500 dpi and precious little usable resolution above 3000 dpi. There may be unusual images that are remarkable for resolution above 3000, but I have very seldom encountered one. Upsampling from 2000 dpi is often equal to scanning at 4000 dpi for 4x5. And four times as fast. Of course, if a client orders 4000 dpi, that's what they truly get.

You didn't see any additional image detail with the shadows lifted, because there isn't any. The entire image histogram already fits within the D800E sensor's dynamic range. I only added the shadow lifting curve for visual reasons, and because the image would probably print easier that way. If you turn off the shadow lifting curve, you will see the image about the way it looks on the light table. The shadows are very dense.

The scan shows detail that is simply not readily visible on direct viewing without using a very bright hot light - which is a completely unreasonable way to predict useful shadow detail that can be pulled off a piece of film. It has always been frustrating to be able to see it, but not to reproduce it.

I'm not sure what grain texture in the water you might be seeing. Not that it might not be there. There certainly may be some interaction between the demosaic algorithms in ACR and certain frequencies of image detail. All such algorithms do this to some degree and create false resolution or microscopic banding and drum scanners are free of that (although they have other warts). I can see several in this scan in the trees that I'll leave "undiscovered" for now. They would only be visible if a 40" x 50" print were viewed from close up.

I'll try to get a 4000 dpi, oil-mounted drum scan of this image done tomorrow with my trusty, dusty ol' D4000. I really should have done that anyway. Let's see what turns up!

What I really would like to see are scans from some high-end medium format camera sensors. Especially the IQ3-100 or new Sony 50 MP sensor in the Fuji GFX/Hasselblad X1D/Pentax 645Z. Rumor also is that Nikon's new DSLR will have a 50MP sensor.

Rich

[Peter De Smidt]

Magnagons were often machine vision lenses, and so Rich's lens might've been used in a pretty hostile industrial environment.

I'm not sure what the advantage of the larger format chips would be, except maybe covering a bigger area of a negative at once. According to DxO, they don't have any dynamic range advantages over full-frame, and I wonder how lenses do over a bigger area. Daniel uses his D810 in 5:4 mode, as he saw some quality drop off in the corners of full frame.

In my experience, the best lenses are those that are optimal at f/4-f/8, as there will be diffraction losses at smaller apertures, which are exacerbated by shooting close up.

[Sasquatchian]

I'm pretty sure you're going to see less flare (at the micro level, not the huge overall flare veiling a large part of the frame) from the drum scan, as long as your lens is clean and you're autofocus is working correctly. It should also be sharper overall given the way the optical system on the Howtek works and the fact that the lens on the Howtek is only recording a very small area of the film as it passes by the lens and does not have to worry a bit about whether it's sharp at the corners. I've never used the older 4000 scanners from Howtek, only the 4500, 7500 and 8000 and I don't know if the 4000 has the lens on end of the fiber optics cable that focuses the light into the actual scanner lens that the 4500 has but was removed from the 8000. The 8000 definitely benefits from opaque taping outside the film right up to the edges of the film, which will remove any tendency to see flare coming in from the clear drum area. The 4500 only does this to a very minor degree. It might be a good idea to tape off the film on your scanner just in case. I used to use that expensive metallic slide masking mylar tape but when that got harder to get and more expensive, I just started using simple, cheap, black electrical tape, taped to the outside of the C-42 overlay material. Simple, cheap, and very effective. Phil Lippincott once gave me a long convoluted explanation for the flare and claimed that it was caused by Trident using "log" mode to scan with. He was wrong. It wasn't and his software was actually worse in that regard when I did a side by side comparisons between DPL and Trident, much to his chagrin. Having Evan do a custom alignment of the optical path of the scanner did make a small visible improvement, so that was worth it. If we can keep these scanners going another fifteen or twenty years, I'll be happy, or perhaps, dead. Ha!

[Rich14]

I've added two files to the Dropbox folder, https://www.dropbox.com/sh/flir6e4j7oxnlca/AABpFAztTRvfaZQEw9q9ljyPa?dl=0

They are Panorama2.tif and Panorama3.tif

Here's a low resolution jpeg of these images:

http://pre06.deviantart.net/ce91/th/pre/f/2017/218/8/7/panorama3_by_rich14-dbj39rf.jpg

Panorama2.tif is a 4000 dpi Howtek D4000, oil-mounted drum scan of Pali's Velvia image. It's a ridiculously large file. So large that I couldn't save it with any layers as either a tif or Photoshop file.

So I downsampled to 2000 dpi (Panorama3.tif) and added a monitor (output) sharpening layer and a layer to slightly lift the shadows. You can drag these layers to the larger file (recommended).

These scans have less than 1% of the dust seen in my D800E scan, but I didn't do any spotting, so there may still be some.

The color in the D800E scan is much closer to the original transparency. Although the D800 scan is a little under-saturated. Reds are exaggerated everywhere in the drum scan. Honestly it is very hard to see the actual color in the deep shadows with a loupe on the light table or with a hot light. The shadows are just too dense. The drum scanner is digging detail out there and a combination of the Velvia color curves going haywire in the underexposure and the PMTs operating at full bias are resulting in color distortion.

Detail is significantly better reproduced in the D800E scan. Look at the little white sign on the water's edge below the trees. I may have slightly blown it out in the camera scan, but it is sharper and the tree trunks and branches on the shoreline are far more three-dimensional and detailed. They are flat in the drum scan by comparison. Their color and the color of leaves and other debris on the shoreline in the camera scan are more faithful to the transparency and significantly more life-like.

I will do one last effort - an oil mounted D800E scan. I still haven't figured out what is giving me the flare on the sides of the image. I've got to get that under control. It's only happening with this image.

I'm surprised and delighted that my camera/lens combination is optically outperforming the optical bench in the scanner. Going into this effort, i doubted that would be the case. When I first saw that months ago, it was the biggest surprise of this entire exercise. And I think I can better the performance of the Micro Nikkor 105mm AF 2.8D with one of the APO lenses I have. That will be some time away, though.

This whole effort was only going to compare the "DMax" capability of the D800E vs a Tango scan. I've put a lot of time into this. After I post the oil-mounted D800 scan, I'm going to give this topic a rest for a while. I'll leave all the files in place for a while. Feel free to download them for analysis or comparison.

I'm very eager to see more scans of this image which I'll be mailing back to Pali shortly. Please feel free to comment on any aspect of this whole exercise.

Rich

[Pali K]

Rich, thank you for continuing your efforts with this image. Download and take a look at the Tango image I posted in my previous reply for your comparison as well. What I like about this exercise is that everyone gets to see the output from various methods and can make their own conclusions.


I would contribute by scanning on a Eversmart Supreme and a Imocan if someone handles the dynamics of sending the original around.

Bob, please PM me your address if you are still interested to give this image a go. I'll mail it to you once I have had the change to scan it on my ES Pro and Tango (Wet Mount) again.

Pali

[Sasquatchian]

I downloaded the 4000 ppi drum scan. Quick question: Did you slightly rotate that either in Trident (I'm assuming that's what you're using) or later in Photoshop? I'm asking because your scan is showing telltale signs of a bug in Photoshop's resampling math when it comes to image rotation. I discovered this and traced it to the method of interpolation used in the rotation. You need to use Bicubic Smoother and nothing else and preferably do it with Free Transform where you can conveniently choose your type of interpolation in the options bar. What I'm seeing are the large regular soft cross hatching that is the result of rotating with the wrong interpolation.

Secondly, how has your scanner been profiled and why are the blacks so red? How did you get from Scanner RGB to ProPhoto? Was that in your scanning software or in Ps? What is your current scanner input profile?

Well at least the skies aren't clipped anymore.

[Rich14]

I downloaded the 4000 ppi drum scan. Quick question: Did you slightly rotate that either in Trident (I'm assuming that's what you're using) or later in Photoshop? I'm asking because your scan is showing telltale signs of a bug in Photoshop's resampling math when it comes to image rotation. I discovered this and traced it to the method of interpolation used in the rotation. You need to use Bicubic Smoother and nothing else and preferably do it with Free Transform where you can conveniently choose your type of interpolation in the options bar. What I'm seeing are the large regular soft cross hatching that is the result of rotating with the wrong interpolation.

Secondly, how has your scanner been profiled and why are the blacks so red? How did you get from Scanner RGB to ProPhoto? Was that in your scanning software or in Ps? What is your current scanner input profile?

Well at least the skies aren't clipped anymore.

The scan was done with Silverfast AI 6.

I was very surprised at the excessive red also. I commented on that earlier. The shadows are indeed quite red, but not this exaggerated.

I can't remember which software I used for the ICC profiling. I used both a Kodak IT8 4x5 and a 35mm Velvia from Silverfast. I haven't profiled in a while, but no other scans show any kind of color shift. I don't rely solely on a profile to produce scans. The final scan is specified by the customer to meet one of a number of criteria. Either to look essentially the same as the transparency when viewed by daylight (very common request) or to match a piece of artwork that has been captured on the transparency, or balanced for skin colors or a particular product. The image space is specified by the profile in the scanning, not in PS.

I would have color balanced this one to appearance by daylight, but it's so dense, that I really can't see the color well. As I said, the D800 scan is much closer to the visual appearance of the film. Pali did a dry mount scan which appears on page 3 of this thread. It also has a lot of red.

I rotated the image 90 degrees in Photoshop CS6. No other rotation applied anywhere. I've not been aware of a rotation bug in PS. The only interpolation was downsampling the 2000 dpi image, Panorama3.tif.

I realize that every aspect of the scan is up for scrutiny. But I was interested only in the ability to see into the shadows in these scans, As I mentioned in my first long post after I had finished the D800E scan, I consider this image useless for reproduction, and any scanning attempt is a "rescue effort" that is intended to dig data out of the deep murky shadows at the expense of other image qualities. Without question, if requested to print this image, I would work on the color balance. But a drum scanner is not a magical device that can turn any kind of image into a thing of beauty.

[Rich14]

Correction:

I rotated the image in Photoshop CS2. The scanner software is Power PC software. I run it under Mac OS X 10.6 Snow Leopard. The scanner writes a .tif file which I then usually bring into PS6 running under Mac OS X 10.12.6 Sierra. I have PS CS2 on the Snow Leopard partition and I opened the file there and rotated it and saved it before bringing it over to the Sierra partition and PS CS6. One single 90 degree rotation.

[Rich14]

I've put one more file in the Drop Box folder. A 2000 dpi scan that I've brought into Photoshop CS6 only and rotated 90 degrees. Let me know if anyone sees artifacts of the rotation.

And how come no one noticed the "scan lines" in the previous scans? They're more noticeable in this one.

Fixed now. A recalcitrant power supply lug that apparently vibrated loose from its tab and just needed to be reseated and tightened.

Part of the care and feeding of these beasts is to periodically open them up and reseat every connector and chip in the machine. After all, it's a 27 year old computer. And although it's built like a tank electrical connectors that are not soldered in place don't stay perfectly in contact forever. Even soldered joints can eventually cause problems.

[Sasquatchian]

I believe there were also some electrical improvements to the later 4000 scanners, so if you have an early 4000, that could be an issue as well. Never any kind of fan of Silverfast, but depending on what you're using as a scanner input profile, that could indeed be causing a loss of shadow detail, particularly if you didn't use an opaque patch for your 0,0,0 in the target. If you don't do that, the input profile can clip the very bottom end of the shadows. So the target matter, the software used to generate the target and the scanning software too - all matter in getting the most out of the scanner. Even if you think you're not using an input profile, you have to be somewhere even if it's not apparent. The other thing to consider, and I have no way of knowing the history of your scanner, is the condition of the pmt's and whether they need replacing or not.

If you "just" rotated ninety degrees exactly, that will not cause an issue with the furrowing but if you're rotating even a fraction of a degree, either in Ps or SF, you could be bitten. After I get back from Monterey in the middle of the month I'll put together some samples for Chris Cox at Adobe to give him something meaningful to re-code, but you have to be able to show him what to look for and how to reliably repeat the problem. We used to see that cross hatching back in the mid 90's on certain scans we used to get from Howtek's back then, and then, when I started working with Trident and giving John at Colorbyte feedback, we pretty much attributed the effect to a bug in the bicubic interpolation routines within Trident which was avoided by scanning only at hardware optical resolutions.

[Rich14]

I have used every version of Polaris and Trident software Howtek made available. I never want to see them again. Why would you "assume" that's what I'm using, never having communicated before? And why do you assume I don't know anything about scanning?

You're describing problems I left behind twenty years ago. How strange to have that software brought up now. Ancient history.

Yes, I am familiar with the mangling that Polaris/Trident did to image data. Specifically the rotation bug. And that software's endearing trait of throwing away data when downsampling, either to an odd value or an exact divisor of the original.

The only use I had for Trident was as control software to physically run the scanner to get a file into the computer, keeping that software's image editing controls as far away from the file as possible, until I could get it into Photoshop. Howtek made great hardware. But the software they made available was truly painful through the early and mid 90s and not much better after that.

I thought you were referring to a problem in Photoshop. Do you see any artifacts in the latest 2000 dpi image that I rotated in PS CS6?

Silverfast is the best software I have ever used with Howtek machines. I have no experience with it otherwise.

I am quite familiar with Don Hutchinson's advice to apply opaque tape to the IT8 black target to avoid crushing blacks with the resulting ICC profile. In fact, my scans of my targets have his term "stretched" in their title to indicate that they were made with the opaque tape in place.

Silverfast references a ProPhoto ICM file, I installed for that purpose, via Apple ColorSync during the scanning process to output the scanfile in ProPhoto color space.

My intent here was not to get involved with critique of my scanners or their calibration. I am very familiar with their operation and construction, their mechanics, electronics and optics. Drum scanners are becoming extinct. The effort to keep them alive becomes more difficult with each passing month. And each of us who owns these machines knows that it is not a matter of "if," but "when" we will have powered them up for the last time because some major or minor part has finally failed and there are simply no replacements anywhere, anymore. Has anyone tried to get a replacement drum in recent memory? Aztek got out of the drum scanner business for many reasons, but being unable to source drums, anywhere, was right up there.

I agreed to post scans from my D800E for an honest comparison to other drum scanner operators' results with Pali's dark 4x5 so we all would have a basis for comparing a DSLR scan to a drum scan. It will be very interesting to see those scans as they become available.

Thanks for your observations. If you have some experiences with "Film scanner options in 2017" (to me options means alternatives) please chime in. But please try to spare us the name-dropping and references to irrelevant and obsolete technology.

[Rich14]

To all,

I'll have my D800 oil mounted scan up here tomorrow. Just want to tell you it is hard to oil mount to a flat surface (my light table). Real hard.

I've been a drum scanner operator for a long time, and can practically oil mount in my sleep on the curved surface of a drum. A flat surface fights you all the way. I just don't feel that the film is down tight on the surface the way securing it around the drum circumference feels. On a drum you know that film is down.

I taped one end, used a squeegee to smooth it down and taped the other end. Couldn't really tell if the oil flowed out smoothly between the film and glass. Couldn't see the leading edge of the oil film ahead of the blade like I can on a drum.

With the transparency right down on the light source, it's impossible to see any bubbles. On a drum and mounting station up above the light table, bubbles and other imperfections light up like little light bulbs. Not so right down on the deck.

Not fun.

[Sasquatchian]

"I thought you were referring to a problem in Photoshop. Do you see any artifacts in the latest 2000 dpi image that I rotated in PS CS6?"

Yes I WAS referring to a bug in Adobe's interpolation schemes and I also said that if you rotated in even 90 or 180 degree increments, there's no problem because there's no interpolation. Well, I didn't exactly say that but I implied it. Then, I added that there was a resizing bug in Trident, but apparently, it's in a lot of different applications that use variants of bicubic interpolation. Sorry if that confused you.

I just downloaded your 2000 ppi scan and it's actually worse than the 4000 ppi original. Your scanner is not working correctly. You have a problem with both the green and blue channels, like the pmt's for them. Look at your file, either of them or both, at 100 percent, but just look at the green and blue channels by themselves. See, what you're doing here is comparing a broken drum scanner against a dslr. You might think that this technology is ready for the trash heap, but it's not really. Your scanner might be, and all three of the other Howtek models are significantly better than the 4000, so that might be coloring your findings as well.

But please, don't get pissed off at me for pointing out the areas of obvious issues. You seem more than a bit touchy about any sort of constructive criticism, taking it all too personally. I've used Nikon Scan, Linocolor, Epson Scan, Silverfast, DPL, and Trident on various scanners over the last twenty years, and, in MY experience, nothing touches Trident 4.0 for overall quality of scans, whether it's color or black and white negs, reflected or color transparencies. There are a few minor bugs that never got fixed, but it's still fantastic. Maybe you never used the latest version which was vastly improved over the earlier ones, but since you claim to have used every version I can only assume that maybe you never learned to get the most out of it.

Not sure why you seem so tweaked about mentioning Evan or Chris. I mentioned Chris only because he'd probably be the dude to actually fix the Ps issue, and you can thank my communications with him many years ago that moved us beyond the 30,000 pixel limit in earlier versions of Ps.

I wouldn't waste any more time scanning on your old 4000. It needs repair and will only give you the kind of results you're posting to your DB.

[Sasquatchian]

"It's a ridiculously large file. So large that I couldn't save it with any layers as either a tif or Photoshop file."

But you could have easily saved with or without compression in .psb format (large document format) with no practical limit to file size at all.

[Pere Casals]

"It's a ridiculously large file. So large that I couldn't save it with any layers as either a tif or Photoshop file."

But you could have easily saved with or without compression in .psb format (large document format) with no practical limit to file size at all.


You can also show the quality of a large file by also posting selected (small area) crops of the image, shown with all detail possible, like this Epson V750 sample:

https://www.flickr.com/photos/125592977@N05/32535835184/in/dateposted-public/

[Rich14]

I've uploaded my final scan, an oil-mounted D800E scan, Panorama5.psd.

I have new-found respect for anyone who oil mounts on a flat surface. My next attempt to do that will not be directly to the surface of the light table. That's certainly not the way to do this. I'll fashion a glass stage to sit just above the light table's surface on which to mount the film. That way, I can examine the quality of the mounting just like is possible with mounts on a drum. Has anyone else had experience oil mounting on a flat surface?

Let's see if anyone with eagle eyes can spot the one bubble I left untouched! ;-)

I'm not yet sure the oil mounting improved the image. It's slightly different from Panorama1 due to slightly adjusted exposure. Let me know.

With regard to file saving, and file size, I uploaded full scans, not cropped details so everyone could see everything in the image. But when I ran into gargantuan files, trying to keep all layers intact, I also considered I have finite storage in this Dropbox account, and I'm very close to the limit right now. So I reduced file size as I saw fit.

As far as our name-dropping "contributor" is concerned, I find his comments offensive and unhelpful and will ignore him from this point lest I really tell him what I think. His conclusions of my scanner's functioning are incorrect.

Pali, I'll be mailing your 4x5 later today.

Rich

[Rich14]

FWIW,

I am still getting the flare at the sides of the image. Much less so with this oil mounted one. Stumped.

[Rich14]

I have removed my drum scans from the Dropbox folder.

As I explained way back in this thread, this is not about my drum scans vs D800 scans. It's about my D800E scans and drum scans done by others on other equipment, including Tango scanners.

I explained that my scanner was showing symptoms of a loose power supply connection that actually got worse thru the course of the two scans I presented. That's been easily remedied. I ask that anyone who has downloaded those two scans or the downsampled file of the first scan kindly discard them.

I've pretty much supplied all the D800E scan data that I think can be squeezed out of my equipment.

Thanks,

Rich.

[Jim Andrada]

What are you using for oil mounting? I use Kami on flat surfaces all the time with nearly zero issues. What helped me was to use a small plastic squeegee to chase out the bubbles. You can get them at art supply stores. I've tried the little rollers/brayers, but find the squeegee works better for me.

[Rich14]

Jim,

Thanks. I also use a squeegee. It's probably a "muscle memory" issue with me. When one tapes one edge of a film to a drum and pulls the other end tight and tapes it, the tension also pulls the film toward the center of the radius of the drum. That tension in the plane of the film pulls it down against the drum surface.

That "feeling" was absent to me when I pulled the film and taped it to my light table. It felt like there was no assurance it was down onto the surface. Of course, it was, the squeegee pushed it down and the surface tension of the mounting fluid held it there. But it actually was just parallel to the surface, held by the fluid. No amount of tension by the tape made it any tighter to the surface.

I wonder if a piece of film that has a built in upward curl could eventually "pop free" and escape the fluid's surface tension?

No big deal. Probably just me. For practical purposes, the problem probably doesn't happen.

BTW, my favorite squeegee for a long time was a 3.5" "floppy" disk hard plastic case.

Rich

[Pali K]

Got the image back today and made quick scans using Scitex Eversmart Pro scanner. The scanner is old but wow!

Scan Details: 2000 DPI in AdobeRGB, 8 Bits, Custom End Points. No other adjustments.

JPEGs of scan.

1- Rotated and Resized Scans
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/EversmartPro/ES_ShadowTest-1600px.jpg

2 - Above Scan with levels adjustments only
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/EversmartPro/ES_ShadowTest-Levels-1600px.jpg

3 - Crop of above.
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/EversmartPro/ES_ShadowTest-Levels-Crop-1600px.jpg

Download Original ES Pro Scan (http://www.netsoft2k.com/Docs/Media/Pictures/Scans/EversmartPro/ES_ShadowTest-2.tif)

Pali

[Pali K]

Heidelberg Tango Scan

Scan Details: Wet Mount 2000 DPI in ProPhoto RGB, 16 Bits, +10 Exposure Compensation (Max setting). No other adjustments.

JPEGs of scan.

1- Rotated and Resized Scans
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2-1600px.jpg

2 - Above Scan with levels adjustments only
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2-Levels-1600px.jpg

3 - Crop of above.
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2-Levels-Crop-1600px.jpg

Download Original Tango Scan [470 MB File] (http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2.tif)

Pali

[Two23]

Understood. Yes that's an issue. Both theoretical and practical.

But in my completely unscientific testing, the path through the camera's exposure compensation, inherent in-camera RAW settings and ACR/Photoshop "post" processing produce results equal to or better than my own D4000 or Tango scans on screen or printed on an Epson 7890.

I originally tried using some old 150mm and 300mm repro process lenses. As well as my 50mm f/2.8 EL Nikkors (both original and "newer" designs). I expected the process lenses to give excellent 1:1 performance. But nothing works as well as the Micro Nikkor. I believe a big part of the problem with the others was flare in the extension tube and bellows setups. I may be able to improve on that but the Micro Nikkor is just so easy to use.

I imagine there are better lenses for this purpose, but this one is hard to beat.


I own a D800E and a Nikon 105mm Micro f2.8 AFS (current model.) Do you have a photo of your set up? I've been using an Epson v750 to scan 4x5 & 5x7, but it is slow. Sounds like taking four shots of a negative and stitching in CC would be faster and actually better.


Kent in SD

[sheel]

Heidelberg Tango Scan

Scan Details: Wet Mount 2000 DPI in ProPhoto RGB, 16 Bits, +10 Exposure Compensation (Max setting). No other adjustments.

JPEGs of scan.

1- Rotated and Resized Scans
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2-1600px.jpg

2 - Above Scan with levels adjustments only
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2-Levels-1600px.jpg

3 - Crop of above.
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2-Levels-Crop-1600px.jpg

Download Original Tango Scan [470 MB File] (http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2.tif)

Pali

Wow. This is clear evidence if ever anyone doubted PMT technology against DSLR chips. Dont waste your time or money on dslr "scans", you'd be better off with a used ES Pro or a new epson v series if you can't find a drum scanner.

Can't believe the garbage I've read in this thread.

[Pere Casals]

I've added two files to the Dropbox folder, https://www.dropbox.com/sh/flir6e4j7oxnlca/AABpFAztTRvfaZQEw9q9ljyPa?dl=0

http://pre06.deviantart.net/ce91/th/pre/f/2017/218/8/7/panorama3_by_rich14-dbj39rf.jpg



Rich, in this scan you have 2 problems (IMHO) that are easy to correct with few clicks in Photoshop: Stray light and color temperature of your illumination source.

You also need to make a calibration profile with an IT8 slide target.


First, correcting stray light:

168113


Balance:

168114


Saturation:

168115


Result:

168116


I took me just 1min, of course it can be matched better... A good match may require you make a color calibration of your system, with that IT8 target slide and proper software.

Every scanning device may have a slightly different spectral footprint from illumination SPD (CRI... to also be considered) and particular spectral absortion of the on pixels dyes of D800 (made by Sony) sensor.

An spectral footprint is not better than another, after calibration results will look very similar, but not completely the same.


IMHO your project has next challenges: Stray light control, deep shadow (you can use a multiexposure like strategy), selecting the best illumination and color calibration.

There are a lot of advanced techniques to control stray light you may explore (polarization).

[Jim Andrada]

The original question was what one's options for scanning were now and maybe a few years in the future.

And there aren't all that many - an Epson, an old drum scanner, an old industrial strength flatbed ... or a digital camera. And more and more the pendulum seems to be swinging toward the digital camera as the old stuff (and old folks) get scarcer.

I went with an IQmart as a step up from the Epson. I'm sure it will outlast me. Others (particularly younger folks) may make different choices.

[Peter De Smidt]

As far as I can see, there's been nothing surprising here: A Tango in excellent condition will do very well with high density film. Not all drum scanners give the same quality. A good dslr using a good setup can give decent results. Talk about scanners long enough, and someone will behave poorly.......

[SergeyT]

DSLRs are relatively good at "scanning" 35mm negative film(both color and B&W) at around 4000 dpi max. Beyond that one would need to look for a scanner.

[Peter De Smidt]

DSLRs are relatively good at "scanning" 35mm negative film(both color and B&W) at around 4000 dpi max. Beyond that one would need to look for a scanner.

Why? I showed a dslr scan of a resolution target that went higher than that earlier in the thread. Need more resolution? Increase the magnification.

[Pali K]

Peter, I agree 100%. Your resolution target matches the performance of my Tango - maybe even exceeds it. I have heard some Tango units do close to 6000 DPI in tests but that's about the max they go. My Scanmate 11000 does around 8000 DPI but won't get the shadow detail that Tango does. For me personally, it is not about the resolution but the clarity and purity of the scan.

At the end of the day, the most exciting part is that film is so much alive that we are now trying to make the most from it using the latest technology :)

Pali

[Peter De Smidt]

Yeah, the importance of super high scanning resolution is of dubious value, especially with large format.

[Jim Andrada]

Amen!!!

[Rich14]

Rich, in this scan you have 2 problems (IMHO) that are easy to correct with few clicks in Photoshop: Stray light and color temperature of your illumination source.

You also need to make a calibration profile with an IT8 slide target.


First, correcting stray light:

168113


Balance:

168114


Saturation:

168115


Result:

168116


I took me just 1min, of course it can be matched better... A good match may require you make a color calibration of your system, with that IT8 target slide and proper software.

Every scanning device may have a slightly different spectral footprint from illumination SPD (CRI... to also be considered) and particular spectral absortion of the on pixels dyes of D800 (made by Sony) sensor.

An spectral footprint is not better than another, after calibration results will look very similar, but not completely the same.


IMHO your project has next challenges: Stray light control, deep shadow (you can use a multiexposure like strategy), selecting the best illumination and color calibration.

There are a lot of advanced techniques to control stray light you may explore (polarization).

Pere,

As I've previously explained, this was a defective scan, please discard it.

Rich

[Rich14]

Ok,

Now we can compare the cropped areas of the Tango and D800E scan, as well as the Eversmart. (Eversmart image in a previous message)


Tango (oil mounted)
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2-Levels-Crop-1600px.jpg



D800E (crop of Panorama 5, oil mounted)
https://orig00.deviantart.net/8466/f/2017/223/d/f/panorama5_crop_by_rich14-dbjp3mb.jpg

My scan was done without an ICC profile. I intend to make one this weekend. But the D800 scan is still very close to the visual appearance of the Velvia.

Pali, it appears that the Tango is emphasizing reds a little, due to the nature of the Velvia density. Not anywhere as much as I was seeing in my D4000 scan, but a little more than was visually apparent to me on the 4x5. What do you think?

[Rich14]

I own a D800E and a Nikon 105mm Micro f2.8 AFS (current model.) Do you have a photo of your set up? I've been using an Epson v750 to scan 4x5 & 5x7, but it is slow. Sounds like taking four shots of a negative and stitching in CC would be faster and actually better.


Kent in SD


Kent,

I'll photograph the setup this weekend.

Rich

[Pali K]

Ok,

Now we can compare the cropped areas of the Tango and D800E scan, as well as the Eversmart. (Eversmart image in a previous message)

Tango (oil mounted)


D800E (crop of Panorama 5, oil mounted)


My scan was done without an ICC profile. I intend to make one this weekend. But the D800 scan is still very close to the visual appearance of the Velvia.

Pali, it appears that the Tango is emphasizing reds a little, due to the nature of the Velvia density. Not anywhere as much as I was seeing in my D4000 scan, but a little more than was visually apparent to me on the 4x5. What do you think?

Rich, You are correct that Tango is emphasizing reds. I didn't have a matching target to calibrate the scanner so slight variation is to be expected. I usually make minor corrections after the scan so that colors feel true to the scanned image.

Have you done any color negative tests with your DSLR setup? I am hoping to have something to proof my scans before drum scanning and my Epson just doesn't do the originals much justice.

Pali

[SergeyT]

Why? I showed a dslr scan of a resolution target that went higher than that earlier in the thread. Need more resolution? Increase the magnification.
It would require a special equipment to move DSLR precisely in parallel with film plane as well as spending time on stitching and dealing with vignetting \ uneven illumination on edges of each frame, overlap errors, etc.
The Velvia reproduction examples from DSLR so far don't look smooth (somewhat lose the appeal of film usage if you will).

[faberryman]

Now we can compare the cropped areas of the Tango and D800E scan, as well as the Eversmart. (Eversmart image in a previous message)

Tango (oil mounted)
http://www.netsoft2k.com/Docs/Media/Pictures/Scans/Tango/Tango-ShadowTest-2-Levels-Crop-1600px.jpg

D800E (crop of Panorama 5, oil mounted)
https://orig00.deviantart.net/8466/f/2017/223/d/f/panorama5_crop_by_rich14-dbjp3mb.jpg
Just looking at those two images, I think the Tango scan wins hands down.

[Two23]

Just looking at those two images, I think the Tango scan wins hands down.

I'm not so sure. The bottom one looks a little "cleaner."


Kent in SD

[Peter De Smidt]

It would require a special equipment to move DSLR precisely in parallel with film plane as well as spending time on stitching and dealing with vignetting \ uneven illumination on edges of each frame, overlap errors, etc.
The Velvia reproduction examples from DSLR so far don't look smooth (somewhat lose the appeal of film usage if you will).

Yes. It requires a good system. Here's an early version of mine: https://www.youtube.com/watch?v=gmRHTausFls

Here's Daniel's: http://www.youtube.com/watch?v=PXy7RJwIBAo

The stitching can be automated using, for example, a template in PTGui.

[Peter De Smidt]

Just looking at those two images, I think the Tango scan wins hands down.

I agree. The Tango has much more shadow detail, more natural detail and contrast, although the highlights are reddish, as has been mentioned. Notice the white sign. In the dlsr version, the raw file reads clipped even with -100 highlights in Capture One. It looks natural in the Tango scan. But the Dslr scan is pretty good, and this is a slide that plays to the Tango's strengths. It'd be interesting to see this slide scanned on an Epson.

[EH21]

Wow, the Tango looks far better to my eye than the D800E
Just curious - have you tried any kind of multi exposure with the DSLR?

[Pere Casals]

...have you tried any kind of multi exposure with the DSLR?


I'm also thinking that some kind of multiexposure would be useful with the D800.

It looks that Velvia (DMax) slide dynamic range does not fit in the around 14 bits per channel the D800 can deliver. White the D800 top of the trees is near clipping while the shadows have slight less shadow detail than with drum.

It would be interesting taking a D800 shot in with exposure for the shadows under trees an another one ideal for the top of the trees, this would determine if ME can help.


Well, that D800 scan is not any bad, still not as good as the drum but it shows an surprisingly interesting potential, at least to me.

[Rich14]

Rich, You are correct that Tango is emphasizing reds. I didn't have a matching target to calibrate the scanner so slight variation is to be expected. I usually make minor corrections after the scan so that colors feel true to the scanned image.

Have you done any color negative tests with your DSLR setup? I am hoping to have something to proof my scans before drum scanning and my Epson just doesn't do the originals much justice.

Pali

Yes. Scanning negatives is easier with the camera and ACR software than any other method I have ever used. The values of B&W or color negs can be inverted in ACR and controlled completely. Far better than scanning software.

I bring the resulting positive image into Photoshop, sample the mask which is now blue, then create a fill layer above the image, fill it with the mask color and set that layer to Divide. That properly removes the mask where it exists in the image. The mask is just that. It's not an overlay and exists from deep mid-tones through the shadows/blacks. Then I do final color correction.

Rich

[Peter De Smidt]

I'm also thinking that some kind of multiexposure would be useful with the D800.

It looks that Velvia (DMax) slide dynamic range does not fit in the around 14 bits per channel the D800 can deliver. White the D800 top of the trees is near clipping while the shadows have slight less shadow detail than with drum.



This is from a long time ago now, but it shows using HDR techniques with a dslr film scan: http://peterdesmidt.com/blog/?p=657

The question is whether that would help. It looks like the shadows are completely captured, i.e. the histogram extremes, at least on the shadow side, seem captured by the D800e scan, what's gone is the variation in the dark areas that the Tango captures. Look at the shadowed forest floor, for comparison. This could be due to processing, or a loss of shadow detail due to flare, or, as Pere suggests, simple lack of dynamic range, i.e. the shadows fall on the toe of the camera's response curve. Regarding processing, I would avoid all use of clarity, structure,.....,in raw processing. The negative needs to be very well masked for stray light, the room should be dark, and no bright surfaces should be able to affect the image. The lens needs to be very clean, and the interior of extension tubes, bellows.....should be checked for any shiny surfaces. (This is often a problem with macro photography. The serious macro people often put light absorbing baffles inside extension tubes.) My guess is that flare is causing the most significant loss of shadow detail in this case.

DSLR scans are, in my experience, better suited for scanning negatives than for scanning underexposed Velvia.

[Rich14]

It would require a special equipment to move DSLR precisely in parallel with film plane as well as spending time on stitching and dealing with vignetting \ uneven illumination on edges of each frame, overlap errors, etc.
The Velvia reproduction examples from DSLR so far don't look smooth (somewhat lose the appeal of film usage if you will).

It doesn't require special equipment. If it did, no enlargement ever made could be called critically sharp. I carefully set the camera and film plane parallel with a digital bubble level. It's quite easy to do. You have both scans to compare, as well as the Eversmart. Where in the D800 scan is there evidence that the image was not optically aligned?

Taking four images takes a few seconds and stitching them in Photoshop takes about a minute, depending on your computer. The whole scanning process is just minutes compared to a comparable-sized scan on a drum scanner of 15 minutes - 1 hr.. (About 15 minutes at 2000 ppi. My process gives about 2600 ppi. To get that on a drum should be done at 4000 ppi and subsampled down. That would take 60 minutes.).

There is no unevenness of illumination in obtaining the four quadrant scans. The LED light table is absolutely even in brightness across its area. The film ins not moved on the light table. The whole tablet is repositioned. There are no overlap errors.

You are making up problems that don't exist.

My scan may not look as "smooth" as the Tango scan. This could lead us to a very long discussion of the optical path of Tango scanners and that scanner's ability to set it's aperture to the optimum for the highest resolution scans. Tangos scan at a higher aperture than Howtek machines. At 2000 ppi the Howtek automatically selects a 8u aperture size i believe (at 4000 it's smaller - I may have these values slightly wrong) The Tango uses an aperture no smaller than 10-12u, smoothing detail somewhat. This topic has been beaten to death over the years. It's a design decision Heidelberg made and stayed with. At typical magazine page reproduction size (the target market for which the Tango was intended) the smoothed (not less sharp) appearance was the goal. There is no equivalent aperture/scan spot size situation with the camera scan. But it's obtaining it's image at the equivalent of setting a scanner's aperture to minimum.

My scan may look a little more grainy or gritty. I may have overdone my capture sharpening by a small amount (10-15%). Maybe not. I'm used to bringing the image in sufficiently capture sharpened to go on to a large final print size. I print a lot at 20"x30". The image may look a little over sharpened on the screen, but it is correct at this point before output sharpening for the large print. I'll reprocess the RAW file with less capture sharpening and post that later.

We are looking at the finest detail that's in this Velvia shot. (Even though we are "only" looking at about 2000 ppi here (higher scan resolution would have revealed no more). Understand, the area that we're looking at in these crops was so dense that on the light table it visually is just featureless black. It takes a bright light and a loupe to see this image detail. We're almost seeing film "grain" and some film noise in the D800 scan, which is what you get when digging detail out of dark, underexposed film. There is no false image data. there are no halos. Compare the tiniest detail that you can see in the Tango scan to the D800. The latter shows that detail more distinctly everywhere, without any "extra"cruft.

I have done many D800 rescans of images that I have done or acquired over the last 30 years from my drum scanners and from others. I have been in the commercial printing business for a very long time and I have drum scanner files going back into the 1980s. Some from machines that never were available outside the industry) I have the magazine reproductions that we printed then. And I started printing digitally in the early 90s on IRIS machines, then Epson equipment from the late 90s to today. Every single rescan that I have done on the D800e has been an improvement of the original drum scan, whatever its origin. In many cases astoundingly so. The D800 shows micro-contrast, brilliance, tonal transitions and image detail that the older scans just didn't capture. And they were done by very good equipment operators.

I've replaced some of the pictures that were hanging in my home and my wife has remarked how "bright and clear" the new images look compared to the old ones and that she "can easily see things in the pictures she had never noticed before."

I wish you all could personally see the Velvia transparency Pali and I have imaged here (I hope others volunteer to scan it). Even though it is very underexposed, like most transparencies, the image has sparkle and "life" that you can see visually with transmitted light that is an enormous challenge to bring through in a scan. In my opinion, the D800 scan captures this significantly better than the drum scan.

Rich

[faberryman]

I'm not so sure. The bottom one looks a little "cleaner."
I think the cleanness you are referring to is caused by significantly higher contrast. Compare the shadow detail of the forest floor in the lower left quadrant of the images for example.

[Rich14]

This is from a long time ago now, but it shows using HDR techniques with a dslr film scan: http://peterdesmidt.com/blog/?p=657

The question is whether that would help. It looks like the shadows are completely captured, i.e. the histogram extremes, at least on the shadow side, seem captured by the D800e scan, what's gone is the variation in the dark areas that the Tango captures. Look at the shadowed forest floor, for comparison. This could be due to processing, or a loss of shadow detail due to flare, or, as Pere suggests, simple lack of dynamic range, i.e. the shadows fall on the toe of the camera's response curve. Regarding processing, I would avoid all use of clarity, structure,.....,in raw processing. The negative needs to be very well masked for stray light, the room should be dark, and no bright surfaces should be able to affect the image. The lens needs to be very clean, and the interior of extension tubes, bellows.....should be checked for any shiny surfaces. (This is often a problem with macro photography. The serious macro people often put light absorbing baffles inside extension tubes.) My guess is that flare is causing the most significant loss of shadow detail in this case.

DSLR scans are, in my experience, better suited for scanning negatives than for scanning underexposed Velvia.


I believe the D800 scan is aquitting itself very, very well.

I'm getting the sense that there is a lot of negativism just based on the feeling that, "this just can't be.' This method can't equal a drum scanner."

Luddism, anyone?

Seriously, I think all the suggestions about improving the conditions of the D800 scan are well taken. And I think we need to discuss all this. But I also think the remarkable qualities of this D800 scan are being ignored.

I didn't do an additional shadow exposure and run an "HDR" process to capture more deep shadow detail, but I could have. Sorry, but I thought I was obtaining all the shadow detail there was. It would have been easy enough to do. Two exposures for each quadrant. That would have doubled the time to acquire the image, but still, far less time than for a drum scan.

Pali, want to send the film back to me? :-)

I have done everything I can currently do to reduce flare. Oil mounting the film helped (of course). I am not using any extension tubes. I can't reduce flare in my Micro Nikkor 105 any more than its internal structure allows. I believe my lens was as clean as I can get it. My room was quite dim. The only light was from the light table and the 4x5 was completely masked by a cardboard aperture. There was no stray light other than that from the image itself.

Drum scanners are incredibly robust and sophisticated, complicated pieces of equipment, developed by teams and engineers and built in factory conditions with machining capabilities of incredible accuracy. They are refined and refined and refined before they are released.

I'm using a cobbled-together, jury-rigged set up, mounting my camera on a tripod, set up on my dining room table and shooting from a cheap LED tablet. The whole rickety setup vibrates and oscillates. I can see every micro vibration on the live view screen dance the image all over the place. I use mirror lockup, hold my breath for 10-15 seconds and gently trip the shutter with an electronic cable release which I carefully make sure is not itself touching anything. - And I'm still getting sharper, more detailed images than come through on the drum scan. It's as simple as that, folks.

I don't agree that capture sharpening shouldn't be done in ACR during the RAW processing stage, but I will post another crop with capture sharpening reduced to default levels.

Rich

[Peter De Smidt]

I didn't say anything about capture sharpening. I have over 5 years experience doing dslr scans. I wouldn't have put the effort I did into if I didn't think it was a viable option, which isn't the same thing as saying that it's the absolute best in all (or even any!) circumstances. Why the defensiveness?

[Rich14]

I didn't say anything about capture sharpening. I have over 5 years experience doing dslr scans. I wouldn't have put the effort I did into if I didn't think it was a viable option, which isn't the same thing as saying that it's the absolute best in all (or even any!) circumstances. Why the defensiveness?

Peter,

Sorry. I was replying to a combination of messages. If I came off as defensive, I did not mean to. I hope to be able to present data here in an objective way.

Rich

[Rich14]

Reprocessed D800E oil mount, Panorama 5 RAW image quadrants. No capture sharpening. No Levels or Curves adjustment to lift shadows. No post processing sharpening.

https://orig07.deviantart.net/955b/f/2017/224/2/6/shapening_off_by_rich14-dbjspla.jpg

[faberryman]

Reprocessed D800E oil mount, Panorama 5 RAW image quadrants. No capture sharpening. No Levels or Curves adjustment to lift shadows. No post processing sharpening.

https://orig07.deviantart.net/955b/f/2017/224/2/6/shapening_off_by_rich14-dbjspla.jpg

IMHO, the color, contrast, and shadow detail looks better in the Tango scan.

[Rich14]

Interesting.

Since I had the transparency, I can say that the D800 scan colors are more faithful. And the relationship between colors and their relative saturation and contrast are the way they looked visually.

I see very little difference in shadow detail between the scans.

I see far better contrast in the D800 image in all areas. It has sparkle that the drum scan is lacking.

The contrast of the dead tree trunks standing at the water's edge against the darker values in the D800 scan is the way the transparency looked visually. They were quite whitish with distinct three dimensional modeling. That doesn't come through as well in the Tango image where they are almost uniformly gray. There are many subtle colors muted in the Tango in the fine detail of the grasses and foliage at the water's edge correctly reproduced in the D800.

The Tango is emphasizing reds a bit and showing contrast of reddish detail against green that isn't in the transparency. The D800 is emphasizing greens a bit and is doing a better job differentiating between different green hues and shades. Those differences exist in the transparency, but a little less distinctly than the D800 is showing them. These kinds of subtle differences are always seen comparing one imaging system to another.

Again, we are talking about extracting data from an exposure that would be trashed and never used for a quality reproduction. These machines are optimized to extract data from normally exposed film.

And the differences I see in the two scans is way less than I have seen in multiple "shoot-outs" over the years between a number of drum scanners all of supposedly excellent reputation.

I do admit to being biased, but I pick the D800 image.

Again, I really hope we get other scans of this 4x5 posted.

Rich

[Pere Casals]

This is from a long time ago now, but it shows using HDR techniques with a dslr film scan: http://peterdesmidt.com/blog/?p=657

The question is whether that would help. It looks like the shadows are completely captured, i.e. the histogram extremes, at least on the shadow side, seem captured by the D800e scan, what's gone is the variation in the dark areas that the Tango captures. Look at the shadowed forest floor, for comparison. This could be due to processing, or a loss of shadow detail due to flare, or, as Pere suggests, simple lack of dynamic range, i.e. the shadows fall on the toe of the camera's response curve. Regarding processing, I would avoid all use of clarity, structure,.....,in raw processing. The negative needs to be very well masked for stray light, the room should be dark, and no bright surfaces should be able to affect the image. The lens needs to be very clean, and the interior of extension tubes, bellows.....should be checked for any shiny surfaces. (This is often a problem with macro photography. The serious macro people often put light absorbing baffles inside extension tubes.) My guess is that flare is causing the most significant loss of shadow detail in this case.

DSLR scans are, in my experience, better suited for scanning negatives than for scanning underexposed Velvia.

Well, the lack of dynamic range it is a concern. Velvia may deliver +3.6D, so contrast can be 1:4000, this is x4000 more light in the transparent areas than in the shadows.

The 14 bits a DSLR may deliver is some 16000 levels of gray, so this will leave 2 bits to describe detail in the shadows, if the adjusted exposure also has to describe higlights...

What I mean is that (in theory, if I'm not mistaken) 14 bits are not enough linear range to catch 3.6D well. Also we have to consider that the least significant bits (describing shadows) may have lots of noise...

So, IMHO, these simple calculations are suggesting that ME can be useful if it is a contrasty slice...


About flare, if image is dark there is less chance that it has a role. I'd suggest a way to test flare importance, just masking all film but a 1cm hole (or 0.5cm) in a very dense area, in this way most of the flare can be cancelled.

[faberryman]

Interesting.

Since I had the transparency, I can say that the D800 scan colors are more faithful. And the relationship between colors and their relative saturation and contrast are the way they looked visually.

I see very little difference in shadow detail between the scans.

I see far better contrast in the D800 image in all areas. It has sparkle that the drum scan is lacking.

The contrast of the dead tree trunks standing at the water's edge against the darker values in the D800 scan is the way the transparency looked visually. They were quite whitish with distinct three dimensional modeling. That doesn't come through as well in the Tango image where they are almost uniformly gray. There are many subtle colors muted in the Tango in the fine detail of the grasses and foliage at the water's edge correctly reproduced in the D800.

The Tango is emphasizing reds a bit and showing contrast of reddish detail against green that isn't in the transparency. The D800 is emphasizing greens a bit and is doing a better job differentiating between different green hues and shades. Those differences exist in the transparency, but a little less distinctly than the D800 is showing them. These kinds of subtle differences are always seen comparing one imaging system to another.

Again, we are talking about extracting data from an exposure that would be trashed and never used for a quality reproduction. These machines are optimized to extract data from normally exposed film.

And the differences I see in the two scans is way less than I have seen in multiple "shoot-outs" over the years between a number of drum scanners all of supposedly excellent reputation.

I do admit to being biased, but I pick the D800 image.

Again, I really hope we get other scans of this 4x5 posted.

Rich

I guess I can't trust my eyes.

[Pali K]

Just so it's clear for everyone, you are looking at an unadjusted file from Tango in my samples. My Tango is calibrated for Velvia 50 so the slight variation are expected and it will take me 5 seconds to fix the red. I am also sure I can make this raw scan shine but that would just show my Photoshop skills not scanner performance.

Rich, correct me if I am wrong but I do believe you needed to perform multiple tests to get the final results. If that is the case, it's important to note that Tango (and drum scanners) in this are are no fuss scans. Meaning that it's painful to setup but once you scan, you know you are getting the best possible file. Regarding your point about drum scans from the 90s, please keep in mind that software evolved well past that at least for main scanners. Tango software that I use is from 2004 and there were many enhancement to USM, shadow, and main scanner firmware at those latest iterations of the software. This is just another FYI statement since you mentioned you were comparings your scans from 90s and such.

I also saw Peter's YouTube on DSLR setup and also the linked one from Daniel. If you can trust me, I am telling you that I can mount negatives on a drum scanner in equal or faster time than what I saw in Daniel's video which is also one frame at a time. The scan time for 2000 DPI 4x5 is quite fast on a Tango. So unlike popular belief, all drum scanners are not really that slow.

I think there is enough in this thread already to at least show a controlled sample between Tango and DSLR. At this point, everyone is free to make their own conclusions.

I know that I will be taking gentle care of my drum scanners for many many years in the future :)

Pali

[Rich14]

Pali,

Many people will interpret this thread in their own way and with their personal biases. I look on it as a comparison between DSLR scanning, and in this case D800 scanning and Drum scanning. I offered originally to do the D800 scan so that we could have an example of the kind of performance I have been getting. I never wanted this to be a competition. You have been very helpful in all of this. And as I've said, I hope others will join in to establish a data base here of what's possible.

I can practically oil mount and do a 4x5 drum scan in my sleep, the muscle memory is that automatic anymore. But despite the "mickey-mouse" nature of my DSLR scanning setup, I can produce a 4x5 scan at all the useful resolution that can be wrung from it in just a small fraction of the time the drum takes. Less than 5 minutes total after oil mounting. How long does the Tango take to do a 3000 dpi scan? My memory is well more than 30 minutes. (Howtek is 60 min. The scan has to occur at 4000 ppi, no matter what resolution is desired higher than 2000.) My time includes settings in ACR ("scanner settings") which will bring the file into Photoshop essentially finished. Further corrections, if needed are mostly artistic.

I think DSLR scanning is not just a curiosity. If the need to scan film were greater, If that market were still viable for high-end commercial reproduction reasons, I do believe CMOS sensor scanning would completely eclipse PMT technology. And we would look at drum scanner technology with much the same bemusement that we now remember the relative crudeness of process camera pre-press work compared to a digital work flow and imagesetter output or direct to plate technology. It's only the lack of need for large scale film scanning anymore that this hasn't happened.

I can remember very well "old timers" laughing their butts off in the mid 80s about "toy" Postscript desktop work stations and how they never would make the slightest dent in serious production work flow and the huge, refrigerator-sized "mini" computers that ran the industry. I also remember printing house after printing house who couldn't make the transition go belly up. I operated my business with two Amiga computers and a 13" wide imagesetter and it was like driving a motorcycle through a disorganized herd of buffalo as I outmaneuvered those companies with "toy" equipment. Macs then were monochrome SE30s. It wasn't long before we had Macs with color capability and our toy computers suddenly were mainstream and the buffalos were all extinct.

Now those of us who are still nursing drum scanners along and those of us interested in the curiosity of DSLR scanning are simply niche artisans. For those few who are actually still making a living in the film niche, I think the knowledge that there is a viable technology available to do that job, a technology that is not extinct, but instead is expanding and thriving should be a sense of great encouragement and relief that there is a direction to take when their ancient drum machines finally can't be coaxed to do one more scan.

[Rich14]

I guess I can't trust my eyes.

Okay.

[Rich14]

Requested by Kent - my low tech D800E scanning setup:

https://orig11.deviantart.net/bf1b/f/2017/224/d/f/scan_set_up2_by_rich14-dbju601.jpg

Gitzo G2220 tripod, ancient Bogen 3262 ball head, quick release, D800E, Micro Nikkor AF 105mm f/2.8D lens, electronic shutter release, 9x12 LED panel with 4x5 mask, dining room table.

For 2600 dpi scans, (four quadrants) the camera sensor plane is 15.75 inches above the light table surface.

[Greg]

Requested by Kent - my low tech D800E scanning setup:

https://orig11.deviantart.net/bf1b/f/2017/224/d/f/scan_set_up2_by_rich14-dbju601.jpg

Gitzo G2220 tripod, ancient Bogen 3262 ball head, quick release, D800E, Micro Nikkor AF 105mm f/2.8D lens, electronic shutter release, 9x12 LED panel with 4x5 mask, dining room table.

For 2600 dpi scans, (four quadrants) the camera sensor plane is 15.75 inches above the light table surface.

f/stop used?

[Two23]

Couldn't this also be done by fastening the box to the wall? I'm guessing you shoot four overlapping quadrants here.


Kent in SD

[williaty]

I have a very, very strong recommendation for the guys kind enough to do testing and post the results for the rest of us to pick at:


Stop telling us which file is which. Also, make sure to strip the metadata out so that doesn't betray the origin either.


Post the two images, label them A and B, 1 and 2, or whatever you like. Let everyone look at them for a week or so and post which one they think is superior and why. Only after everyone has gone on record as saying which one they think is technically superior do you reveal which file came from which origin. People here are being crippled by their expectancy bias. Whatever they believe about a DSLR scan or a drum scan, they are convincing themselves they see in the examples posted. It's no different than the guys who claim there's an audible difference between a cable that cost $100/m and one that costs $10,000/m. Stop letting us know which file is which before we all have to go on record with an opinion.

[faberryman]

It's no different than the guys who claim there's an audible difference between a cable that cost $100/m and one that costs $10,000/m. Stop letting us know which file is which before we all have to go on record with an opinion.
It's a little different. Unlike high end cables, where you have to squint your ears to tell a difference, the D800 and Tango images are obviously different.

[axs810]

Has anyone tried the Epson Expression 12000XL Photo scanner?

I'm actually pretty interested in it now after hearing it can autofocus or manual focus on negatives. I think it would be really great for 8x10 neg scans vs the v700/800 series

[Rich14]

f/stop used?

I close the lens 3 stops. That would normally read f/8. But in the macro range it's f/13. still 3 stops from wide open.

[Rich14]

Couldn't this also be done by fastening the box to the wall? I'm guessing you shoot four overlapping quadrants here.


Yes, four overlapping quadrants. With the sensor plane a little farther than 15.75 inches, Photoshop's stitching algorithm can't put the quadrants together, although they do overlap. At 15.75 inches, they overlap just enough and the resulting scan is about 2600-2800 ppi.

For 35 mm, the shot is 1:1, 5000 ppi, and the whole rig is much closer to the subject.

I suppose the LED panel could be put on the wall, but this setup is much easier than doing that. Why fight gravity any more than necessary?

[Rich14]

I have a very, very strong recommendation for the guys kind enough to do testing and post the results for the rest of us to pick at:


Stop telling us which file is which. Also, make sure to strip the metadata out so that doesn't betray the origin either.


Post the two images, label them A and B, 1 and 2, or whatever you like. Let everyone look at them for a week or so and post which one they think is superior and why. Only after everyone has gone on record as saying which one they think is technically superior do you reveal which file came from which origin. People here are being crippled by their expectancy bias. Whatever they believe about a DSLR scan or a drum scan, they are convincing themselves they see in the examples posted. It's no different than the guys who claim there's an audible difference between a cable that cost $100/m and one that costs $10,000/m. Stop letting us know which file is which before we all have to go on record with an opinion.

I completely agree. I was going to suggest this.

I was also going to suggest shooting a roll of Velvia, same subject, same exposure for the whole roll. Send the samples out to whoever volunteers as "standard observers." Then we could do the scans and everyone who has samples can evaluate them in terms of comparison with the "original" examples they have. Right now I don't have a film camera, but that's solvable.

Pali, are you up for some more comparisons? I could scan some difficult, but otherwise "real world" Kodachromes that have given my colleagues and me some challenges over the years and send them on to you to scan.

Rich

[Rich14]

. . . you have to squint your ears to tell a difference . . .

Not easy to do! ;-)

[Peter De Smidt]

There are some pretty detailed threads in the diy section on dslr scanners.

[EH21]

Rich,
If that is an LED light box, then quite possibly you will be missing some color. I've been working with different digital camera capture scanning methods myself, but the color of the LED and other light sources can cause problems. I am using the Sekonic C-700 spectrum analyizer to measure the light source.
CRI is one thing but a light source can still report relatively high CRI and have some colors attenuated - often in the Reds. Quite honestly I am skeptical that your D800e with that setup is giving you more accurate colors unless you have somehow made an input profile for that light source for your camera.
Certainly the Tango scan looked much better to me in the images shared.


Requested by Kent - my low tech D800E scanning setup:

https://orig11.deviantart.net/bf1b/f/2017/224/d/f/scan_set_up2_by_rich14-dbju601.jpg

Gitzo G2220 tripod, ancient Bogen 3262 ball head, quick release, D800E, Micro Nikkor AF 105mm f/2.8D lens, electronic shutter release, 9x12 LED panel with 4x5 mask, dining room table.

For 2600 dpi scans, (four quadrants) the camera sensor plane is 15.75 inches above the light table surface.

[Pere Casals]

Has anyone tried the Epson Expression 12000XL Photo scanner?

I'm actually pretty interested in it now after hearing it can autofocus or manual focus on negatives. I think it would be really great for 8x10 neg scans vs the v700/800 series

12000XL it is very interesting for ULF until 31 cm x 43.7 cm., or to scan A3 prints.

IMHO 12000XL has a comparable or lower performance than V700-850 for 8x10", depending on how you scan.

The 12000XL it only has 2400 hardware DPI Horizontal, while the V750 has 4800 (for 8x10) delivering around 2000 with resolution USAF 1951 targets. In the V750 the sensor outresolves the internal lens system. If the V750 scans until 5.9" wide then it uses the HiRes lens and has a pixel density of 6400 instead 4800, resolving some true 2300 to 2800 dpi. IMHO the 12000XL won't reach these figures.


Scanner Focus has some importance for 35mm scans, but way less for larger formats. But for 135 use a dedicated roll film scanner to get most in case of challenging jobs.

[Pere Casals]

Rich,

Certainly the Tango scan looked much better to me in the images shared.


This has been (IMHO) an interesting thread, because it makes think how things work.


Beyond stray light, IMHO the better Tango result is a lot from dynamic range, PMT (Photo multiplier RGB sensors) from drum has more DR than the D800, so the D800 should need Multiexposure to match.


And the end the scanner reduces spectral color information to only 3 rgb values, this is after film has reduced scene spectral information to 3 dyes. As we are talking about Spectral to tristimulus information CRI may have a reduced impact that can be solved by color calibration. I'm guessing, more than CRI a significative criterion may be that RGB illumination peaks should match the transmission peaks of film dyes (each film can have different peaks) .

I'm not completelly sure, but I guess that if illumination and (particular) color film have the spectral peaks in the same points then color calibration will work more accuratelly.

[Rich14]

Rich,
If that is an LED light box, then quite possibly you will be missing some color. I've been working with different digital camera capture scanning methods myself, but the color of the LED and other light sources can cause problems. I am using the Sekonic C-700 spectrum analyizer to measure the light source.
CRI is one thing but a light source can still report relatively high CRI and have some colors attenuated - often in the Reds. Quite honestly I am skeptical that your D800e with that setup is giving you more accurate colors unless you have somehow made an input profile for that light source for your camera.
Certainly the Tango scan looked much better to me in the images shared.

I haven't made a camera input profile yet. I'm going to try to get that done later today. I have only set the camera's white balance to the LED panel as a Custom setting.

A histogram of the light source shows three RGB peaks of about equal height and width, for whatever spectral information that conveys.

I'll make the input profile using my IT8 transmissive Q60 4x5 and generate the ICC profile with Argyll CMS. I wish there we're a way to use that profile in ACR. But ACR only accepts reflective input profiles made from a Color Checker chart, saved in DNG format (as far as I know). I would be able to apply the ICC profile in PS after processing the RAW file, but I want to apply it in the RAW processing stage. ACR doesn't expect cameras to be profiled as scanning devices.

I believe some of the other RAW processing programs, RAW Therapee, for instance accept ICC input profiles. Anyone know?

[Peter De Smidt]

The light source matters for dslr scanning. I've tested a bunch of them. An input profile is not a panacea.

[Two23]

The light source matters for dslr scanning. I've tested a bunch of them. An input profile is not a panacea.


Since I only shoot b&w LF, this would all be academic for me.:)


Kent in SD

[Peter De Smidt]

Perhaps a discussion of light sources for DSLR scanning would best be undertaken in the appropriate thread in the DIY section: http://www.largeformatphotography.info/forum/showthread.php?87536-DSLR-Scanner-Light-Sources/page21

[EH21]

The light source matters for dslr scanning. I've tested a bunch of them. An input profile is not a panacea.

Very true - if you want to scan color you need a source that supplies all the colors to start with. Xenon flash works usually pretty well, tungsten lights also pretty well, LED and fluorescent lights not so well.

[SURF]

For me personally, it is not about the resolution but the clarity and purity of the scan.

At the end of the day, the most exciting part is that film is so much alive that we are now trying to make the most from it using the latest technology :)

Pali
Good point. Any ideas what makes that clarity and purity? I feel my eyes relaxed when I open Tango scans. Why is that? On a PC monitor...