Understanding underwhelming DSLR scan results

[GGfpc]

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

 

[Lew_B]

There are many people with tons more experience in DSLR scanning than me here, hopefully some will chime in… one thing I think could be an issue with your setup is the amount of stray light getting into the lens from the open areas of the light source outside of your film frame. I would suggest you need to mask off all the light from your source outside of the film carrier….all that stray light is going to kill the contrast in the scanning area. It’s like shooting directly at the sun
Like I said, not an expert here at all…Lew

 

[xkaes]

Good point, it should be done in a darkroom -- and mask off the areas of the light source that are open.

 

[Patrick Robert James]

Those are equivalent resolution, but the Fuji scans have a lot more sharpening applied to them which is what you are seeing.

 

[bernard_L]

the Fuji scans have a lot more sharpening applied to them which is what you are seeing

+1.
And do not try to match that amount of sharpening; it is just too much; hurts my eyes.

 

[madNbad]

The tripod on the tabletop is more susceptible to vibrations. A copy stand or old enlarger frame is a better choice for stability. There is far too much extraneous light falling on the negative.

 

[Alan9940]

The trouble with DSLR scanning is that most folks start into it thinking all they need is a simple setup (similar to what you have) and they'll get stellar results. Sorry, but that's not true!

For best results, you need a very sturdy setup that can hold the camera exactly parallel to the film plane. This is usually accomplished with a copy stand or old enlarger pieces custom tailored to this process. The film carrier should hold the film absolutely flat. Set the camera for raw capture, all manual control, daylight white balance, lowest ISO, and set the lens to f/8. Use a magnifier or zoom the preview to 10x and focus carefully on the grain of the film. Run some test frames to get your exposure correct (use shutter speed to accomplish this), then do a final check of alignment, if you touched the camera. Use a cable release, darken the room, and eliminate any extraneous light from your light source. Oh, and it's a good idea to use a lens hood.

I use a laptop tethered to my camera which gives me complete control of camera functions without the need to touch the camera, once it's setup. Since you'll probably be using relatively slow shutter speeds for the captures, it's a good idea to use live view to eliminate any vibrations from mirror slap.

Sounds like a lot of effort, but once you get the hang of it it's not bad. Have fun!

 

[Light Capture]

Half life size magnification is a good place to be for negative scanning. Since these are shot on APS sized sensor that's approximately where you will end up. Standard macro lenses usually perform well at this magnification.

There are several issues with the setup.
- 1/30s is in danger zone. Most if not all mirrorless cameras I tried had visible shake between 1s and 1/200s. Try electronic shutter if there's one on your camera.
- Tripod setup is finicky. It needs to be extremely stable to get good results. Even the table it's sitting on probably isn't stable enough to get results.
- Mirrorless cameras like yours have big advantages for scanning since you can scan remotely on most of them and can magnify image to 100% or beyond. Shake will be visible at these magnifications.
- Another issue is that it's extremely hard to make sensor parallel to the film when scanning from tripod. Edges will certainly be soft.
LED source needs to be of high quality even for BW scans (this is not the issue in your scans but can cause artifacts in some details).

Some techniques to get better results or eliminate some of these issues:
- Increase ISO to get you 1/250-1/500s speed and check results. There will be lots of noise but it should be possible to judge if shake is the issue or not
- Back off on magnification and use 0.25 life size to check if lens performs better there. Shake will have much less impact at smaller magnification. Once you get good results here, you can try increasing magnification. I don't have experience with Tamron 90mm so can't comment how it performs at specific magnifications.
- I'm using precision level on negative stage and on either (or both) lens mount and filter mount of the camera/lens combo. Any level should get you in ballpark but it will be hard to adjust precisely.
- Lab scan definitely looks oversharpened and grain on it still looks bit fake and mushy. Your results shouldn't look that sharp no matter what you do. Little bit of sharpening can be added.

Everything needs to be aligned and shake minimized for good results.

 

[Steven Lee]

@GGfpc I do not see more detail in Frontier scans, but I see a fairly aggressive sharpening applied on them.

If you are wondering if there's room for improvement, it's hard to tell without seeing the full sized scans. 100% patches are helpful, but not knowing the magnification makes it hard to see whether you can do better or not. If you ask me to guess, I'd say you are not far from the limits of your hardware (lens+sensor).

 

[GGfpc]

Does the extra non-masked light really make that much of a difference? I never considered that.

About the sharpening, I also sharpened my shot. It seems like everyone is in agreement that the lab scan is oversharpened but I'd like to know how I can achieve that look from a technical standpoint even if it's not astheatically pleasing. Is it just a unsharp mask?

I'll retry the process raising the shutter speed and see if I get better results.

 

[madNbad]

Does the extra non-masked light really make that much of a difference? I never considered that.

About the sharpening, I also sharpened my shot. It seems like everyone is in agreement that the lab scan is oversharpened but I'd like to know how I can achieve that look from a technical standpoint even if it's not astheatically pleasing. Is it just a unsharp mask?

I'll retry the process raising the shutter speed and see if I get better results.

Your meter is reading the extraneous light in addition to the light from your source. Plus, the extra light is washing across your negative. If you take the simple steps of darkening the room and using a remote shutter release, you will see improvement. Camera scanning is a learning process but some things are necessary, a good stable stand that holds the camera parallel to the base and level, a consistent light source and a way to hold the negative flat. This is a good place to find answers and welcome to the forum!

 

[Helge]

Looks like a focus issue.
You should probably consider placing the film between AN glass or naphtha moistened glass, setting the aperture to 8 or 11 to increase DoF (a good trade off vs. diffraction) and shooting at higher shutter speeds.
Get a more powerful light if that is what is holding you back.

Also, what film where you shooting?

 

[momus]

Lab scans tend to be over sharpened for some reason. You are using something like PS to edit the images after scanning, right? You need to do that w/ any scanner, I had to do that even on my expensive Nikon Cool Scanners.

 

[GGfpc]

Looks like a focus issue.
You should probably consider placing the film between AN glass or naphtha moistened glass, setting the aperture to 8 or 11 to increase DoF (a good trade off vs. diffraction) and shooting at higher shutter speeds.
Get a more powerful light if that is what is holding you back.

Also, what film where you shooting?

What does the glass do? I'm shooting Fomapan 200. I have used F8 but it's less sharp. My lens seems to be sharpest at f2.5 which I find super weird. I used f5.6 since it gets me decent DoF with better sharpness than f8. I use the mirror trick to make sure everything is aligned.

Lab scans tend to be over sharpened for some reason. You are using something like PS to edit the images after scanning, right? You need to do that w/ any scanner, I had to do that even on my expensive Nikon Cool Scanners.

Yes, all the pics have been edited although I didn't take much care with my scans since I just wanted to match the lab for this test

 

[Pieter12]

Vibration from the relatively slow shutter speed and a long lens can introduce subtle softness. Are you using a cable or other remote release? Self-timer works, too. 1/30th might be too slow, you could increase the ISO also you can use a shutter speed of at least 1/125. It looks like you might be shooting tethered. If so, take the laptop off the shooting table, pressing the key or pad to fire the camera could introduce vibrations. I am not familiar with the camera, but I will assume it is mirrorless--otherwise, use mirror lock-up.

 

[Helge]

What does the glass do? I'm shooting Fomapan 200. I have used F8 but it's less sharp. My lens seems to be sharpest at f2.5 which I find super weird. I used f5.6 since it gets me decent DoF with better sharpness than f8. I use the mirror trick to make sure everything is aligned.

Glass keeps the film flat, which is usually the biggest challenge. 5.6 might get you more absolute sharpness in an ideal situation (full open is the best for the hypothetical perfectly corrected lens, which is partly achieved with some very good lenses in the center).
But in flat macro work it’s very easy to get out of ideal. Some of the frame might be as good as it gets while other parts are soft.

 

[cerber0s]

Did you shoot raw or jpeg? Is there any noise reduction added in camera, or in post processing?

I also find that over exposing by 1.3 stops works wonders when photographing negatives.

 

[runswithsizzers]

Steven Lee makes a good point in post #9. If your examples are showing crops at approximately 100%, then there is probably room for some improvement, but not a huge amount. And, as others have said, the lab scans are grossly oversharpened and they are not a good goal to aim for.

My first negative copying rig was similar to yours, and based on my experience, the first thing I would troubleshoot is vibration.

How are you releasing the shutter? It is almost impossible to press the shutter buttion without introducing motion blur. You need either a time delay shutter release or a cable release. And try getting rid of the table - set the tripod on the floor, preferably concrete. It is not unusual for someone walking around in the house to cause enough vibration to cause motion blur. If your tripod allows the center column to be reversed, that might possibly be more stable? (Just guessing; I always used mine with the column reversed.)

The second thing I would investigate is your focusing technique. Are you using autofocus, or are you focusing manually? If manually, does your Canon have some kind of zoomed in focusing aid?

If you are copying with the base side up, you might try flipping the negatives so the emulsion side is up, and see if that helps. (don't forget to flip them back to normal in post processing.)

Next I would consider your post-processing sharpening. There are many different sharpening routines, with many different settings. It shouldn't take anything fancy, but it can take some time to figure out the best settings.

That is most of the low hanging fruit. The next level involves spending money.

 

[Pieter12]

Did you shoot raw or jpeg? Is there any noise reduction added in camera, or in post processing?

I also find that over exposing by 1.3 stops works wonders when photographing negatives.

Many cameras feature a bracketing option. Once you have gone to the trouble of cleaning and positioning a negative, bracketing is a simple bit of insurance, preferrably using shutter speeds. And shoot RAW or DNG or whatever you camera offers--that will give you a bit more adjustment latitude.

 

[GGfpc]

Steven Lee makes a good point in post #9. If your examples are showing crops at approximately 100%, then there is probably room for some improvement, but not a huge amount. And, as others have said, the lab scans are grossly oversharpened and they are not a good goal to aim for.

My first negative copying rig was similar to yours, and based on my experience, the first thing I would troubleshoot is vibration.

How are you releasing the shutter? It is almost impossible to press the shutter buttion without introducing motion blur. You need either a time delay shutter release or a cable release. And try getting rid of the table - set the tripod on the floor, preferably concrete. It is not unusual for someone walking around in the house to cause enough vibration to cause motion blur. If your tripod allows the center column to be reversed, that might possibly be more stable? (Just guessing; I always used mine with the column reversed.)

The second thing I would investigate is your focusing technique. Are you using autofocus, or are you focusing manually? If manually, does your Canon have some kind of zoomed in focusing aid?

If you are copying with the base side up, you might try flipping the negatives so the emulsion side is up, and see if that helps. (don't forget to flip them back to normal in post processing.)

Next I would consider your post-processing sharpening. There are many different sharpening routines, with many different settings. It shouldn't take anything fancy, but it can take some time to figure out the best settings.

That is most of the low hanging fruit. The next level involves spending money.

I'm shooting tethered with a 10s delay to avoid the vibration. I never thought about doing it on the floor. I was worried about dust on the negative. I've used the center column in reverse but I thought it would be less stable because the camera needs to be held "up" to be 90º even though the center of gravity is lower, but I'll try again. I'm copying the emulsion side at the moment.
Could you give me some hints of different sharpening routines?

 

[runswithsizzers]

I'm shooting tethered with a 10s delay to avoid the vibration. I never thought about doing it on the floor. I was worried about dust on the negative. I've used the center column in reverse but I thought it would be less stable because the camera needs to be held "up" to be 90º even though the center of gravity is lower, but I'll try again. I'm copying the emulsion side at the moment.
Could you give me some hints of different sharpening routines?

I expect your tabletop setup is probably good enough, depending on how solid your floors are, and assuming no one is stomping around in the house. I once lived in a shabby rental house where my stereo turntable would skip whenever someone walked through the room. Moving the turntable to a different wall (90 degrees left or right) solved the problem because of the direction the floor joists ran.

Does your digital camera have a focusing aid? My Fuji has a Focus Assist button that zooms in for focusing and it allows me to choose a red (or blue) indicator that shows where there is good focus.

But, it sounds like you have been doing everything right to optimize sharpness, so you may be getting near optimal results from the gear you have.

I can't really help much with sharpening routines, especially not for GIMP which I do not use. Before I got the NLP plug-in for Lightroom, I inverted my b&w negatives in Photoshop. There is a "High-Pass filter" available in Photoshop, which I often used with good results - but I don't know how similar it is to GIMP's version. Of course, Photoshop has about 5 or 6 different sharpening filters, and some of those have numerous settings. It would probably take me about half-a-day to try all the permutations in Photoshop.

Now that I use the NLP plug-in for Lightroom, I sharpen using Lightroom Classic's Detail tab which has only 4 sliders. The Detail and Masking sliders help to avoid oversharpening the fine details (so-called grain) - which is what is causing your lab scans to look so gritty and speckled. I think I may slightly prefer the results I was getting from Photoshop's High Pass filter(?), but for my Lightroom workflow that would take a lot more effort for a questionable improvement.

Honestly, since I started using my present lens (Rodenstock APO-Rodagon D 75mm f/4.0 duplicating lens), I don't often feel the need to do any additional sharpening other than what Lightroom applies to my RAW files by default.

 

[GGfpc]

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.

 

[xkaes]

Does the extra non-masked light really make that much of a difference? I never considered that.

That light is going straight into your lens. You may not see it in the viewfinder, but it's clearly hitting the front of your lens -- messing everything up.

 

[runswithsizzers]

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.

Unrealistic expectations can result from the method used to evaluate your results. For example, we don't know how much your sample images are magnified. Are the crops showing a 100% view? 200%?

While doing any sharpening, I zoom in to 100% magnification. And if I am comparing two images, again I may take a look at 100%. But I don't see any point in ever evaluating my results at any magnification greater than 100%. And practically speaking - depending on the resolution of your image and your intended use - sharpness which may look a bit iffy at 100% can look perfectly acceptable at normal viewing sizes.

Since you are working with RAW files, remember that sharpening done in GIMP/Photoshop, etc. is Capture / Input sharpening. When you output to your final use - be it making a digital print or making a JPEG for screen viewing - then you will have a second opportunity to sharpen again. Output sharpening may be more important than what you see when looking at full sized RAW images on your computer screen, because output sharpening is done after resizing, and using settings specific for the intended use.

 

[George Mann]

Undersharpened vs. oversharpened (user/operator error). Never underestimate what a properly calibrated Frontier is capable of.