Flat bed vs DSLR scanning

[Dismayed]

I still use my Nikon CS9000 for scanning because I'm unwilling to give up the infrared channel for cleaning up old color film.

 

[albireo]

Here are two examples I made with an APS-C sensor on a Fuji X-Pro3 to show the size difference for printing if you choose to make slices and use a panoramic feature such as I have done using Lightroom:


If you want to read the article I wrote about how I digitize:

Digitizing Film Part 2: Software, Contact Sheets, Bracketing, Stitching, Workflow | photoscapes

I perform several tasks while creating image files, including bracketing, blending, and stitching larger film sizes.

photoscapes.com

Digitizing Film Part 1: Copy Stands, Film Holders, Medium Format | photoscapes

Curious how I digitize film? In this post, I share the tools I use—without sales hype—plus links for specs. My digitizing process will follow in a future post.

photoscapes.com

I plan on writing a part three as soon as my home gets finished with construction after storm damage. I have been digitizing for a few years and also use a medium format digital back for really large printing.

Hi @darr thanks for sharing. These appear to be great results - I really like the tone transitions in your orchid close up, apart from highlights being a little hot, which might be just down to personal choice of course. Excellent write up too.

What I have to say though is that I'm not sure I would be able to consistently pick these up against say V750 scans of the same 4x5 negative in a double blind test so again I don't see the overwhelming advantage, unless of course one owns already an X-Pro3 camera and is willing to experiment.

 

[McDiesel]

@darr your work on https://photohaikus.co is superb, really enjoyed my time there!

 

[Adrian Bacon]

I still use my Nikon CS9000 for scanning because I'm unwilling to give up the infrared channel for cleaning up old color film.

That's a totally valid reason, and some of the reason why I still have and maintain a dedicated film scanner. Sometimes some old negatives will come in that haven't been kept well and the output isn't such high resolution that I need to be delivering huge files, AND the added scan time and IR cleaning is actually shorter than manually spotting out the dust after the scan. In those instances, absolutely, having a film scanner with a dedicated IR channel for dust removal is clutch.

 

[projectbluebird]

darr, thank you for sharing your posts.
I especially appreciate your explanation of the rational behind some of the steps in digitizing B+W negatives.

 

[Paul Verizzo]

A scanner is a lot bigger than a camera, so there's more material, which costs more money.

But I think it's really just the marketplace at work. They sell a gazillion digital cameras and only a few scanners. Guess which one will cost more?

That's a really strange....and wrong....observation. A scanner has a lot of air within, and the optics aren't even close to what a lens has to do.

 

[Paul Verizzo]

I just "scanned" this thread, pardon the pun, and here are my three cents:

First, what is the goal of why you are scanning? Archiving those family Kodachromes? You just want a decent 250 lines per mm? Or, are you making a billboard?

And the counterpoint is, what is your camera's capability?

If you have a bunch of 35mm slides of good and consistent image quality, a camera scan is fast. HOWEVER, if brightness range and density varies......not so fast, cowboy.

Behold, the versatile flatbed scanner. I have long used a long in the tooth Canon 8800F LED scanner. Currently selling for about $100 with shipping on eBay. It can scan 35mm negs and slides and 6x6 medium format. Sure, it's slow compared to camera scanning, but I just go do other things.

But where it really shines is when the negs or slides are less than ideal. The built in Canoscan software allows every possible image manipulation before scanning. Levels, tone, etc. A camera scan with a less than ideal image can only scan the image as is, except for exposure, and then too late to correct.

My software options include Autotone, Dust and Scratch Correction, Grain Correction (totally awesome with grainy B&W), Fading Correction, Backlight Correction, Color balance controls, Brightness/Contrast Controls, White and Black setpoints, Levels, and custom tone curves. Whew!

Can't do any of those with a camera.

 

[flavio81]

even high resolution full frame DSLRs are just starting to touch 2400-3200 DPI resolution, which is still well within what something like an Epson V850 can deliver in terms of raw resolution,

You'll need to be very very careful (i.e. play a lot of time with height adjustment and/or glass holders) to get real 2700dpi out of the Epson V850, and even then, it would be "smeary" resolution, with little clarity.

A 36mpx digital camera like the Sony A7R with a good macro lens can extract a good 7360 x 4912 image. This represents over 5100dpi.

 

[Paul Verizzo]

You'll need to be very very careful (i.e. play a lot of time with height adjustment and/or glass holders) to get real 2700dpi out of the Epson V850, and even then, it would be "smeary" resolution, with little clarity.

A 36mpx digital camera like the Sony A7R with a good macro lens can extract a good 7360 x 4912 image. This represents over 5100dpi.

And you need that because? Making billboards?

 

[grat]

You'll need to be very very careful (i.e. play a lot of time with height adjustment and/or glass holders) to get real 2700dpi out of the Epson V850, and even then, it would be "smeary" resolution, with little clarity.

Not really. The real problem is that 135 film doesn't really support more than about 3600 PPI anyway. I've got at least one 6400 PPI scan, which while it looks perfectly "OK", it doesn't look that much better than the 3200 PPI version.

While they vary wildly in resolution and pixel density, I'm don't think these qualify as "smeary" with "little clarity".

Sample Photos

Mostly test photos

flic.kr

A 36mpx digital camera like the Sony A7R with a good macro lens can extract a good 7360 x 4912 image. This represents over 5100dpi.

For 35mm, yes. For 4x5, in one shot, it's about 6140x4912, or about 1228 PPI. For a 6x6 image, you'll get about 2183 PPI, for a 4912x4912 image.

For a 4x5 image, at a "mere" 3200 PPI (easily done with an Epson v800), the final image is 14,966x11,776 (176.2MP).

You can do that with a DSLR, but you're probably looking at 6 frames stitched together, possibly more (It's late, and I don't feel like doing the math).

 

[flavio81]

Not really. The real problem is that 135 film doesn't really support more than about 3600 PPI anyway.

This is not true at all.

Anyways, this topic has already been beaten to death.

 

[flavio81]

I've got at least one 6400 PPI scan, which while it looks perfectly "OK", it doesn't look that much better than the 3200 PPI version.

Because, as I've said above and as you can find out in the internet if you do the research, your scanner doesn't have more than about 2700 real, true DPI. That's why your "6400dpi" scan won't look much better than "the 3200ppi" version.

Now, run the very same frame (assuming it's a sharp frame) through a really high quality dedicated scanner like a top Nikon scanner, and you'll see all the detail you've been missing.

Just a simple illustrative example -- although you'll find plenty on the web.

Epson V700 -- tested to reach about 2400dpi maximum. The V850 doesn't improve too much over this figure, even if using glass carriers and adjusting height.

 

[flavio81]

And you need that because? Making billboards?

You have a very valid point. Of course if one is mostly posting pictures online or printing to small sizes, in theory it's pointless to be pursuing very high resolutions.

The problem is, these Epson flatbeds also introduce chromatic artifacts and smearing of details. And, sadly, these add a fuzziness of details that is evident even in reduced-size images. Users counteract this loss of details by adding sharpening, and this adds artifacts as well. The end result is an unnatural look that has nothing to do with the very same image properly scanned or optically printed.

The other reason to require a high real (optical) resolution is to avoid grain aliasing problems, particularly with high ISO film.

To put an example of a good "low-res" scanner -- the Fuji Frontier scanner, the Fujifilm SP3000 if i remember correctly, is a mid res scanner, max output is about 13 MP if I recall correctly (about 4000dpi). But the optical system is very faithful, it doesn't add chromatic abberration or smearing. Even when you request a lower-res image from it, it gets a very clear result. However, due to the lowish optical resolution, put a high ISO (grainy) film through it and the grain will become way more pronounced than it is in reality, compared (for example) to an optical print of the same frame. This i've experienced firsthand.

 

[grat]

This is not true at all.

"most" 135 film seems to range from 50 to 100 lp/mm, with high resolution films like t-max 400 peaking at 200 lp/mm. I realize there's not a hard and fast rule for translating to PPI-- even those numbers are based on contrast ratios, which are going to vary from section to section, but 100 lp/mm is translated by SilverFast, for example, to about 3250 PPI

Anyways, this topic has already been beaten to death.

This is photrio. There are no dead horses that can't be beaten just a bit more.

 

[Alan Edward Klein]

Because, as I've said above and as you can find out in the internet if you do the research, your scanner doesn't have more than about 2700 real, true DPI. That's why your "6400dpi" scan won't look much better than "the 3200ppi" version.

Now, run the very same frame (assuming it's a sharp frame) through a really high quality dedicated scanner like a top Nikon scanner, and you'll see all the detail you've been missing.

Just a simple illustrative example -- although you'll find plenty on the web.

Epson V700 -- tested to reach about 2400dpi maximum

I've found that you need to sharpen scans when using an Epson flatbed. Without knowing what was applied, this is not a good sample. Maybe the operator applied no sharpening to the V700 scan.

Here's a comparison between an Epson V850 and Howtek 8000 drum. scans The results compare favorably after sharpening the Epson scan.

Howtek 8000 Drum vs. Epson V850 flatbed scanners

Here's a comparison between Howtek HiResolve 8000 drum scanner and Epson V850 flatbed. Negative is mine, Tmax 100, processed normal by North Coast Photo dip and dunk with Clayton F76+ developer (similar to Kodak D76). Taken on a Chamonix 45H-1, Schneider Symmar APO 150mm with B+W orange...

www.largeformatphotography.info

 

[grat]

Because, as I've said above and as you can find out in the internet if you do the research, your scanner doesn't have more than about 2700 real, true DPI. That's why your "6400dpi" scan won't look much better than "the 3200ppi" version.

Yeah, I've read those "reviews". And I have an Epson v800. Their results are, for lack of a better word, poor. They're using a glass target, and they're not focusing at the 3mm height that the Epson sensors are calibrated for for high resolution work (or if they are, they completely failed to mention it in their methodology). Garbage in, garbage out.

 

[flavio81]

Here's a comparison between an Epson V850 and Howtek 8000 drum. scans The results compare favorably after sharpening the Epson scan.

Howtek 8000 Drum vs. Epson V850 flatbed scanners

Here's a comparison between Howtek HiResolve 8000 drum scanner and Epson V850 flatbed. Negative is mine, Tmax 100, processed normal by North Coast Photo dip and dunk with Clayton F76+ developer (similar to Kodak D76). Taken on a Chamonix 45H-1, Schneider Symmar APO 150mm with B+W orange...

www.largeformatphotography.info

Thanks, you're supporting and demostrating my previous claim about the smearing of details by the Epson scanner -- it's completely evident on the link you're supplying. It demonstrates the clarity and the fidelity of the drum scan, which represents the real detail on the negative.

So, thanks, here's the comparison from the link you've provided. It's evident how the Epson output is smeared. Yes you can add sharpening, here it works because the image, being large format, is devoid of grain problems. Try the same scan, on a 35mm frame, add sharpening to the Epson output and the difference would be dramatic.

But since you want to compare to drum scanners... Here's a medium format (6x7) frame, V850 vs Heidelberg Tango drum scan. Epson at the right.

And this isn't a 100% crop! Thus substantianting my point that the differences are evident even if the image is reduced in size. Click to enlarge.

 

[flavio81]

Yeah, I've read those "reviews". And I have an Epson v800. Their results are, for lack of a better word, poor. They're using a glass target, and they're not focusing at the 3mm height that the Epson sensors are calibrated for for high resolution work (or if they are, they completely failed to mention it in their methodology). Garbage in, garbage out.

No, you're just choosing to ignore the evidence.

Check out this carefully done test of Epson V850 with film height adjustment, still no more than 2600dpi possible.

There are also other tests of glass carrier vs no-glass carrier (all with height adjustment). The difference is minimal. Which is something people who have been using enlargers for decades already knew.

Well, to each its own, i already provided enough info.

 

[Alan Edward Klein]

Thanks, you're supporting and demostrating my previous claim about the smearing of details by the Epson scanner -- it's completely evident on the link you're supplying. It demonstrates the clarity and the fidelity of the drum scan, which represents the real detail on the negative.

So, thanks, here's the comparison from the link you've provided. It's evident how the Epson output is smeared. Yes you can add sharpening, here it works because the image, being large format, is devoid of grain problems. Try the same scan, on a 35mm frame, add sharpening to the Epson output and the difference would be dramatic.

View attachment 309430

But since you want to compare to drum scanners... Here's a medium format (6x7) frame, V850 vs Heidelberg Tango drum scan. Epson at the right.

And this isn't a 100% crop! Thus substantianting my point that the differences are evident even if the image is reduced in size. Click to enlarge.

View attachment 309428

You;re comparing to the V850 without sharpening. Here's after sharpening. (Epson on the right) Who can tell the difference to the Howtek drum?
Screen shot compares scanners at 100 percent after sharpening by Alan Klein, on Flickr

 

[Alan Edward Klein]

No, you're just choosing to ignore the evidence.

Check out this carefully done test of Epson V850 with film height adjustment, still no more than 2600dpi possible.

There are also other tests of glass carrier vs no-glass carrier (all with height adjustment). The difference is minimal. Which is something people who have been using enlargers for decades already knew.

Well, to each its own, i already provided enough info.

When I got my V850, I tested the sharpness at each of the five different heights you can select on the film holder. I was surprised that there is a noticeable difference in sharpness between the best position and the ones immediately adjacent to it. Everyone should check this out on their machine to verify the best position and mark it for future use.

 

[Alan Edward Klein]

Regarding 35mm, I suppose the v850 cannot provide sharpening as good as a drum scanner. But it does a pretty good job certainly acceptable for the web. Here's 35mm with Tmax 400.

Removing ground brush with fire - 35mm Tmax 400

New Jersey Forest Fire Department clears bush by burning. Near Craig House, Monmouth Battlefield.

www.flickr.com

 

[jtk]

Who scan's "for the web" primarily/only/at-all ?

 

[McDiesel]

@Alan Edward Klein Effectiveness of sharpening depends on subject matter. The wood texture apparently is a better subject than skin imperfections in Flavio's sample. For this reasons scanning devices must be compared with sharpening turned completely off.

There's another variable not being discussed: appearance of grain. It is a much higher bar to reach. Because even at smallish sizes the grain quality impacts the appearance of an image. Scanning is tough on grain. It either gets smeared or exaggerated, and there's an inverse correlation between grain quality and sharpening. For this reason I am in a "minimal sharpening while scanning" camp.

 

[flavio81]

When I got my V850, I tested the sharpness at each of the five different heights you can select on the film holder. I was surprised that there is a noticeable difference in sharpness between the best position and the ones immediately adjacent to it. Everyone should check this out on their machine to verify the best position and mark it for future use.

Sorry, you misunderstood me. I'm not saying there's no difference when adjusting height. Of course there is! What i mean is that it will not enable you to go beyond the optical limitation of the machine (around 2700dpi).

And what makes "little different" is to use a glass holder vs a non-glass holder, assuming both have their height optimized.

 

[flavio81]

You;re comparing to the V850 without sharpening. Here's after sharpening. (Epson on the right) Who can tell the difference to the Howtek drum?
Screen shot compares scanners at 100 percent after sharpening by Alan Klein, on Flickr

Yes, I can see the difference, moreover if you click on the link and see the above mage at full size.

And, as I hinted above, here due to the film being large format, the impact of grain will be minimal. But apply sharpening to a 35mm frame or 6x6 frame scanned with drum scan vs V850 and the difference will be dramatic.