Understanding underwhelming DSLR scan results

[Light Capture]

I'm starting to think that maybe my expectations are unrealistic. Based on my experience with digital, the lab scan seems sharp and the grain seemed finer, but I guess analog scans shouldn't look like that.
I'll try use faster shutter speeds and maybe give my lens a cleanup in the future, but from what I understood from everyone's responses, improving my scans significantly will become costly.

I think your setup is capable of producing better results than what you got. Needs some experimentation and getting vibration down to minimum.
From quick Photoshop analysis, your scans hold slightly more detail than lab scans.
Grain can be deceiving. Unless you're observing and measuring grain under microscope, it might not be represented accurately on scans.
Sometimes doubling magnification doesn't result in grain being twice as big leading to conclusion that there are artifacts either from lightning or resolving frequencies.
Another thing to try is to move negative further from the light source. Also closer to the light source and see if that affects grain appearance.
If you have a flash that can be triggered bellow negative that's worth trying. Flash will freeze any motion and will likely provide better quality of light.

 

[tnp651]

If shake is an issue, try using a speed light as a light source. Put it behind white Plexiglas for evenness. You'll still need a constant source to nail the focus but then turn that off for shooting.

 

[Pieter12]

I'm starting to think that maybe my expectations are unrealistic. Based on my experience with digital, the lab scan seems sharp and the grain seemed finer, but I guess analog scans shouldn't look like that.
I'll try use faster shutter speeds and maybe give my lens a cleanup in the future, but from what I understood from everyone's responses, improving my scans significantly will become costly.

What is an analog scan?

 

[runswithsizzers]

I considered trying to use a flash for copying film with a digital camera, but I wasn't able to think up a configuration that didn't seem awkward.

Now if my film copy rig was designed to work horizontally, it would be easy. But my copy stand and negative holders are designed to work with the camera mounted vertically and pointed down. Presently, I am using a <Skier Sunray Copy Box> as my light / negative carrier. Since my negative carrier is sitting on my light source, switching to a flash would mean devising some other way to hold the negative carrier. And then some way to switch between the focusing light and the flash without touching the negative carrier. It is doable, but adds considerable complexity compared to my present setup - which is working pretty good for me, so I never tried to incorporate a flash.

I started <this thread> to discuss options for how to set up the flash, and got a few good suggestions, if anyone is interested.

 

[250swb]

OP, maybe somebody has already spotted the problem but I think it's way, way too much sharpening, or at least sharpening that you could probably do better in some other software. You overall setup looks rickety but I don't see why it shouldn't work, and if your tablet is bright enough any daylight won't figure in the exposure. For levelling get a bubble level app for your phone and lay your phone flat on the screen. I would tend to use more than f/5.6, a good macro lens will be fine at f/11 and allow for any slight curvature of the film or not being perfectly level.

 

[Adrian Bacon]

Hi! I've recently started my journey in analog photography and I'm currently using a DSLR to scan my negatives.

I'm using a Canon M50mk2 (APS-C) with a Tamron 90MM F2.5 1:2 Macro Lens.

My issue is that my scans are noticeably less sharp than what I got from the lab that uses a Fuji Frontier SP3000. I've searched on the internet and it seems like I should be able to get similar results sharpness wise, but with better resolution from the DSLR.

I've attached some pictures. Left is DSLR and right is lab. All shots were taken at F5.6 (I compared the sharpness at several apertures) and 1/30s. I've sharpened in GIMP using the High-Pass filter technique.



In the IMGUR below is a picture of my setup and the original shots. I use a tripod and a Rollei Key Light to achieve the scan. It seems like my lens could use a cleaning (it looks worse in the picture than in reality), but could that really be the difference? It seems like I'm not able to focus clearly on the grain.



Any help is welcomed

what aperture are you scanning at? If you're doing it wide open, that's probably your issue... Also... Tamron? They're not known for being particularly sharp, and in the case of using a digital camera to scan film, the lens you use will make or break you. The second thing to look out for is your shutter speed. Unless you have a very sturdy setup or a camera with 5 axis IBIS, your shutter speed will hose you up. The third thing to look out for is how you treat your pixels after the fact does affect perceived sharpness.

 

[PinkPony]

I'm starting to think that maybe my expectations are unrealistic. Based on my experience with digital, the lab scan seems sharp and the grain seemed finer, but I guess analog scans shouldn't look like that.
I'll try use faster shutter speeds and maybe give my lens a cleanup in the future, but from what I understood from everyone's responses, improving my scans significantly will become costly.

Here's my current setup if it serves for inspiration: https://bitmaster.se/posts/2023-01-03-light_box_v2/

I do think that stray light can lower contrast just as it does when printing in a dark room.
I try to mask out light outside the negative carrier.

Here's a direct link to some sample scans with the setup: https://bitmaster.se/posts/2023-01-03-light_box_v2/sample_gallery/
The photos with 4/3 aspect ratio is 645 negatives, the others 135 format.

There's no tracking, adds or other horrors behind the links above.

 

[Huss]

Hi! I've recently started my journey in analog photography and I'm currently using a DSLR to scan my negatives.

I'm using a Canon M50mk2 (APS-C) with a Tamron 90MM F2.5 1:2 Macro Lens.

My issue is that my scans are noticeably less sharp than what I got from the lab that uses a Fuji Frontier SP3000. I've searched on the internet and it seems like I should be able to get similar results sharpness wise, but with better resolution from the DSLR.

I've attached some pictures. Left is DSLR and right is lab. All shots were taken at F5.6 (I compared the sharpness at several apertures) and 1/30s. I've sharpened in GIMP using the High-Pass filter technique.



In the IMGUR below is a picture of my setup and the original shots. I use a tripod and a Rollei Key Light to achieve the scan. It seems like my lens could use a cleaning (it looks worse in the picture than in reality), but could that really be the difference? It seems like I'm not able to focus clearly on the grain.



Any help is welcomed

If you are not able to focus on the film grain you will not get a sharp image.

When I first tried my hand at film scanning I had a similar set up like you. Dslr, manual focus lens, tripod over the image. Awful results.
Get rid of that tripod and get a copy stand. They sell new on ebay for about $150. Do whatever you need to do to be able to focus on the film grain. If that is not sharp, you’re wasting your time. I use an AF macro lens and it instantly focuses on the grain.
Shoot at base ISO for max quality. Stop down to f8-f11. When you shoot macro your dof is wafer thin, and that will accomodate for any film curl/bend.
Use a delayed shutter release/self timer.
Make the room dark.

 

[Matroskin]

I'm also one of the guys who is not glad with DSLR/mirrorless camera scans if compare to Lab scans or Nikon 8000/9000 scans.
My setup is Fuji XT-4 + 80mm macro lens + kaiser slimlite plano + lomography film holder. Parameters usually are 1/30, f/8, ISO 400. Every time I focus manually and use remote control to shoot.
Finally what I don't like - I don't see grain just digital noise from camera and the dynamic range seems to be lower comparing to scanner.



PSActually, DOF at f/8 is quite high (~8 mm) and it's almost impossible to miss.

 

[runswithsizzers]

DOF at f/8 is quite high (~8 mm) and it's almost impossible to miss.

Can you please share with us where this number came from?

 

[Matroskin]

Can you please share with us where this number came from?

Why not This is medium format film shot without stitching

 

[brbo]

You have to know when those numbers apply...

In negative scanning you don't have even half a mm of error margin.

 

[eradman]

In my experience:

• Run a series of experiments to find the best aperture
  (at close focus your effective aperture is increasing and depth of field is very thin)
• Use your camera's native ISO
• Invest in a macro lens to minimize distortion
• Use a solid surface to and self-timer to minimize vibrations (even loud noises can reduce sharpness!)
  A tripod is fine if it's very sturdy

 

[runswithsizzers]

@Matroskin,
If I understand correctly, you are using a Fuji XT-4 digital camera to copy medium format film with an 80mm lens set to f/8.0? And, at a distance of 40 cm from the camera sensor to the film, the width of the 120 image just fills the full height of the sensor (no more, and no less), right?

If so, then I would calculate your magnification as: height of sensor / width of 120 negative = 15.6mm / 56 mm = 0.28x. Compared to using the same crop sensor camera to copy a 35mm negative, which would require a magnification of: 15.6mm / 24mm = 0.65x. The higher magnification required to copy 135 film is going result in much less DoF compared to copying 120 film.

The PhotoPills website offers two Depth of Field Calculators. Your screen shot appears to be from the one titled simply, "DEPTH OF FIELD (DOF) CALCULATOR" at <this web adderss>. According that calculator, your setup should have a DoF = 0.61 cm = 6.1 mm as shown in your screenshot.

But PhotoPills also offers a "MACRO DEPTH OF FIELD (DOF) CALCULATOR" located <here> The macro one is a little more complicated to use, and I can't say I completely understand it. But unless I'm using incorrect inputs, the macro calculator is showing your DoF as closer to 3.0 mm. However, that calculation also shows the magnification as "0.38x" not the expected 0.28x. If I change the focusing distance to 46-47 cm, then I see a magnification of 0.28x-0.29x and the DoF changes to about 5.0 mm. Less than "~8mm" -- but still a generous amount -- if you are copying 120 film at 0.28x.



However, if you want to copy 135 film with a crop sensor camera at 0.65x, then the DoF drops to about 1.25mm, and things like alignment and focusing start to get a little more tricky:


For a second data point I played around with some calculators offered on this website <https://www.cambridgeincolour.com/tutorials/macro-lenses.htm> and I got results similar to PhotoPills calculations.

 

[Matroskin]

In my experience:

• Run a series of experiments to find the best aperture
  (at close focus your effective aperture is increasing and depth of field is very thin)
• Use your camera's native ISO
• Invest in a macro lens to minimize distortion
• Use a solid surface to and self-timer to minimize vibrations (even loud noises can reduce sharpness!)
  A tripod is fine if it's very sturdy

Yep. I did everything except native ISO. Maybe that's the key

runswithsizzers,​

thank you for the detailed answer, but I really doubt that the reason is DOF. I realize the closer the distance the lesser DOF. Previously I used f/11 and had the same "problem". Apparently the reason I use "high" ISO. So I'll do tests and come back with results

PS earlier I had out of focus issues and that was because I didn't adjust focus every time before shooting. And now I do it very thoroughly


PPS Never tried to scan 35mm.

 

[eradman]

Yep. I did everything except native ISO. Maybe that's the key

runswithsizzers,​

thank you for the detailed answer, but I really doubt that the reason is DOF. I realize the closer the distance the lesser DOF. Previously I used f/11 and had the same "problem". Apparently the reason I use "high" ISO. So I'll do tests and come back with results

PS earlier I had out of focus issues and that was because I didn't adjust focus every time before shooting. And now I do it very thoroughly


PPS Never tried to scan 35mm.

About focus: the choice between manual and autofocus should probably be based on the equipment you have.

I usually use autofocus (Fuji X-E4 + XF 60mm f/2.4 ASPH)

 

[Matroskin]

About focus: the choice between manual and autofocus should probably be based on the equipment you have.

I usually use autofocus (Fuji X-E4 + XF 60mm f/2.4 ASPH)

I use manual
Could you share results? 100% crop

 

[Adrian Bacon]

My setup is Fuji XT-4

Fuji x-trans sensors are sub-optimal for scanning film. They might do OK for BW, but for color negative, you want more red and blue samples than what the x-trans CFA pattern gives you. It really cuts down on your color resolution.

 

[eradman]

I use manual
Could you share results? 100% crop

Sure, here is a test "scan":

http://raticalphotography.com/articles/images/macro_aperature_test_f8.png

(film is 645, HP5+)

 

[runswithsizzers]

thank you for the detailed answer, but I really doubt that the reason is DOF. I realize the closer the distance the lesser DOF. Previously I used f/11 and had the same "problem". Apparently the reason I use "high" ISO. So I'll do tests and come back with results

Yes, when you say DoF is unlikely to be the cause of the softness we can see in your posted example, I agree. But as someone who is copying 35mm film, I thought your statement about having "~8 mm" of DoF didn't sound right, so I wanted to figure out why.

If I had to guess why your camera scans are soft, I would be more inclined to troubleshoot motion blurr caused by minute vibrations first. You don't say how you are supporting the camera, but 1/30th sec. is going to require a very solid support.

Another possibility - assuming your 80mm macro has OIS - some say OIS should be turned off for tripod work - other say leave it on. You might play with that.

Personally, I am a little skeptical that your "high" ISO is causing you to loose much sharpness; 400 is not that high on a Fuji. However, if you have boosted noise reduction in post processing, that can cause loss of sharpness.

Here are a couple of images taken with Kodak T-Max 400 and copied with my Fuji XT-1 (16MP). These have been converted with NLP and no additional sharpening other than Lightroom defaults. They are crops from Lightroom at 100%. With your Fuji XT-4 and larger negatives, you should be getting sharper images than I am. I should mention I am using a pretty good lens for this purpose - a Rodenstock APO-Rodagon D 75mm f/4.0. But if your 80mm macro is a Fuji lens, I don't think that is causing you any problems.

 

[Matroskin]

Sure, here is a test "scan":

http://raticalphotography.com/articles/images/macro_aperature_test_f8.png

(film is 645, HP5+)

thank you. Looks good. Is it one frame?

Yes, when you say DoF is unlikely to be the cause of the softness we can see in your posted example, I agree. But as someone who is copying 35mm film, I thought your statement about having "~8 mm" of DoF didn't sound right, so I wanted to figure out why.

If I had to guess why your camera scans are soft, I would be more inclined to troubleshoot motion blurr caused by minute vibrations first. You don't say how you are supporting the camera, but 1/30th sec. is going to require a very solid support.

Another possibility - assuming your 80mm macro has OIS - some say OIS should be turned off for tripod work - other say leave it on. You might play with that.

Personally, I am a little skeptical that your "high" ISO is causing you to loose much sharpness; 400 is not that high on a Fuji. However, if you have boosted noise reduction in post processing, that can cause loss of sharpness.

Here are a couple of images taken with Kodak T-Max 400 and copied with my Fuji XT-1 (16MP). These have been converted with NLP and no additional sharpening other than Lightroom defaults. They are crops from Lightroom at 100%. With your Fuji XT-4 and larger negatives, you should be getting sharper images than I am. I should mention I am using a pretty good lens for this purpose - a Rodenstock APO-Rodagon D 75mm f/4.0. But if your 80mm macro is a Fuji lens, I don't think that is causing you any problems.

View attachment 328308 View attachment 328309

I almost forget about stabilizer.. It is also has IBIS and I'll consider it as a possible reason in my nest tests. And actually 1/30 is pretty fast (I shoot with handheld medium format camera without blurring) and of cause I use tripod. May be in future I should invest in some kind of copy stands..
PS I also use NLP..

 

[eradman]

thank you. Looks good. Is it one frame?

Yes, although with bracketed exposure to provide more flexibility in processing (shooting JPEG fits my workflow better than RAW)

 

[runswithsizzers]

thank you. Looks good. Is it one frame?

No, two different frames from the same roll. (I have that roll posted <here> if you want to see the complete images.) Those two 35mm negatives were taken with my SMC Pentax 35mm f/2.0 (M-series) which is probably not my sharpest lens.

I almost forget about stabilizer.. It is also has IBIS and I'll consider it as a possible reason in my nest tests. And actually 1/30 is pretty fast (I shoot with handheld medium format camera without blurring) and of cause I use tripod. May be in future I should invest in some kind of copy stands..
PS I also use NLP..

I think I would put OIS / IBIS at the top of my list for testing.

And I would put your tripod second on the list. Unlike hand-holding your medium format camera, putting a longish lens on a tripod, and zooming in close to your subject is going to amplify the tiniest bit of camera movement enough to cause motion blur - which 1/30th sec is unlikely to stop.

 

[kino-eye]

Hi! I've recently started my journey in analog photography and I'm currently using a DSLR to scan my negatives.

I'm using a Canon M50mk2 (APS-C) with a Tamron 90MM F2.5 1:2 Macro Lens.

My issue is that my scans are noticeably less sharp than what I got from the lab that uses a Fuji Frontier SP3000. I've searched on the internet and it seems like I should be able to get similar results sharpness wise, but with better resolution from the DSLR.

I've attached some pictures. Left is DSLR and right is lab. All shots were taken at F5.6 (I compared the sharpness at several apertures) and 1/30s. I've sharpened in GIMP using the High-Pass filter technique.



In the IMGUR below is a picture of my setup and the original shots. I use a tripod and a Rollei Key Light to achieve the scan. It seems like my lens could use a cleaning (it looks worse in the picture than in reality), but could that really be the difference? It seems like I'm not able to focus clearly on the grain.



Any help is welcomed

Are you using an extension ring? You need one: a Canon APSC 3:2 sensor needs ~0.62x (1:1.62) magnification (greater than 0.5x, or 1:2) to fill the sensor with a 135 frame. (Sony/Nikon/Fuji APSC sensor is slightly larger, and needs ~0.65x [1:1.54].) According to this online calculator you need a 12mm or longer extension ring to achieve that on your M50 with the Tamron 90mm 1:2 lens, which has a reputation for good sharpness. (A Sony et al. crop sensor would need minimum 14mm.) Without it, you can't focus close enough. For those pushing the OP to f8 or narrower, the DoF is probably already large enough that we didn't realize the softness is slightly missed focus! (Although I agree the postprocess sharpening also plays a big role in perceived sharpness in this comparison.) I would stick with f5.6 on the crop sensor, though f8 probably won't hurt much if at all.

If you don't have an extension ring handy, you can fill the frame of an MFT camera with that lens alone at 1:2. Or just move your M50 farther away until you can focus on the grain, and crop to the smaller image as needed--losing some pixels in the process, of course. Or spring for a new 1:1 macro lens . . .

 

[kino-eye]

If you are not able to focus on the film grain you will not get a sharp image.

Agreed, start here before chasing down other gremlins. I have used a tripod and remote release (app on my phone) with good results. A copy stand is more convenient for faster set up and take down, but not required if you're careful.

I'm also one of the guys who is not glad with DSLR/mirrorless camera scans if compare to Lab scans or Nikon 8000/9000 scans.
My setup is Fuji XT-4 + 80mm macro lens + kaiser slimlite plano + lomography film holder. Parameters usually are 1/30, f/8, ISO 400. Every time I focus manually and use remote control to shoot.
Finally what I don't like - I don't see grain just digital noise from camera and the dynamic range seems to be lower comparing to scanner.



PSActually, DOF at f/8 is quite high (~8 mm) and it's almost impossible to miss.

You don't specify which lens you're using, but the softness in your comparison suggests you may be facing the same thing: insufficient magnification and pushing beyond the minimum focus distance. That seems less likely if you're filling the frame with a medium format negative though . . .