DSLR Camera Settings for Scanning of B&W Film

[Manny A]

I Realize the process probably holds many way of doing scan conversions. Opinions greatly appreciated.

I'm using a Canon R mirrorless camera with 100mm RF macro lens to scan B&W Negatives shooting in RAW. Utilizing Negative Lab Pro to convert negative images to positive. What are the best camera settings to use in order to do this? Should the camera be set to take snapshot of negative in color or B&W?

Thanks in advance

-Manny

 

[Alan9940]

I use a Canon 5D Mark II for scanning B&W film and I set the WB to daylight, aperture to f/8, shutter speed appropriate for a nice histogram. I shoot in raw mode tethered to a laptop and, like you, its NLP for the conversion.

 

[Manny A]

I use a Canon 5D Mark II for scanning B&W film and I set the WB to daylight, aperture to f/8, shutter speed appropriate for a nice histogram. I shoot in raw mode tethered to a laptop and, like you, its NLP for the conversion.

Thanks Alan.

I've been using f/5.6. ISO set to 200 so the shutter speed is slow around 1/10. You keep the camera in color mode verses B&W? Good point. I need to focus on the histogram in my settings.

I scanned a old set of Kodak T400 negatives recently that looked horrible. For some reason, really grainy with a sepia tone to them. I scanned them in color mode. Was wondering if this might be the problem.

-Manny

 

[Alan Edward Klein]

What is your light source? Don't you have to adjust the white balance to it

I use a Canon 5D Mark II for scanning B&W film and I set the WB to daylight, aperture to f/8, shutter speed appropriate for a nice histogram. I shoot in raw mode tethered to a laptop and, like you, its NLP for the conversion.

?

 

[Alan9940]

Thanks Alan.

I've been using f/5.6. ISO set to 200 so the shutter speed is slow around 1/10. You keep the camera in color mode verses B&W? Good point. I need to focus on the histogram in my settings.

I scanned a old set of Kodak T400 negatives recently that looked horrible. For some reason, really grainy with a sepia tone to them. I scanned them in color mode. Was wondering if this might be the problem.

-Manny

I use f/8 because IMO that aperture should provide enough DOF for slight bow, etc, in the negative. I keep my ISO at its lowest setting--100 in my case. My shutter speeds tend to hover around 1/15 sec for a good histogram for the style B&W negatives I make. I use live view and operate any camera settings I might need to change and fire the shutter from a tethered computer. The camera is mounted to a very sturdy copy stand.

Sorry, I really don't understand your question regarding color mode or B&W. If you shoot a raw file, it's color by default, right? Color being grayscale, of course, with a B&W negative.

 

[Alan9940]

What is your light source? Don't you have to adjust the white balance to it


?

I'd have to pull it out to give you the make/model, but it's a very high CRI light source that allows me to set the temperature of the light.

 

[koraks]

Don't you have to adjust the white balance to it

What gives? OP is scanning b&w film. Color temp of the light source won't matter.

When shooting raw, it can Furthermore be adjusted in post to taste.

 

[JerseyDoug]

I scan B&W film with a Fuji X-T20, but I expect any ~24MP mirrorless would produce similar results. I set the ISO to the native 200 for the camera, set the format to RAW, set the aperture to f/8 and set the shutter speed to AUTO. This gives me nicely centered histograms regardless of the density or contrast of the negatives.

I open and develop the RAW files in Affinity Photo where I change the profile to 16 bit gray, invert the image and set the black and white points just touching the ends of the histogram. If I want to print a contact page right away I do a batch job of the whole roll with Affinity Photo and set the black and while points with the Auto Levels function. I create the contact page with ContactPage Pro.

 

[L Gebhardt]

What gives? OP is scanning b&w film. Color temp of the light source won't matter.

When shooting raw, it can Furthermore be adjusted in post to taste.

It does make a slight difference in the noise levels. Ideally you want to use light with the color temp of the native white balance of the sensor, almost always around daylight, and expose to the right. That should give you the most information per pixel and the cleanest results. Because this is negative film noise will show in the highlights once reversed. That makes image noise more visible.

For color neg adjust the light to give even exposure through the orange mask in all three channels with the camera WB set to daylight.

 

[wiltw]

Aperture selection has to be the balance of

• sufficient DOF to deal with lack of flatness in the film being photographed
• large enough aperture to not induce diffraction effects...the peak performance of the lens is typically about -2EV from wide open

 

[Alan Edward Klein]

It does make a slight difference in the noise levels. Ideally you want to use light with the color temp of the native white balance of the sensor, almost always around daylight, and expose to the right. That should give you the most information per pixel and the cleanest results. Because this is negative film noise will show in the highlights once reversed. That makes image noise more visible.

For color neg adjust the light to give even exposure through the orange mask in all three channels with the camera WB set to daylight.

Doesn't the light have to be a specific Kelvin rating?

 

[koraks]

It does make a slight difference in the noise levels.

It matters to the extent that the light source is ideally matched to the sensor - although sufficient light output would be a priority since that's a threat to noise just as well. It does not matter to the best of my knowledge what you set the white balance to on the camera during exposure, provided the camera is recording RAW files. My initial wording wasn't careful enough; note how Alan seemed to refer to the camera white balance setting, and not so much the color temperature of the actual light source. The latter matters (a bit, at least), the former not so much IMO.

Doesn't the light have to be a specific Kelvin rating?

Daylight = 5000K through 6000K generally, depending a bit on which standard you go by. Anything in that range (and even quite a bit outside of it) will work fine with a modern mid- to high-end digital camera.
Here's a good start to get the basics down: https://www.cambridgeincolour.com/tutorials/white-balance.htm

 

[250swb]

I Realize the process probably holds many way of doing scan conversions. Opinions greatly appreciated.

I'm using a Canon R mirrorless camera with 100mm RF macro lens to scan B&W Negatives shooting in RAW. Utilizing Negative Lab Pro to convert negative images to positive. What are the best camera settings to use in order to do this? Should the camera be set to take snapshot of negative in color or B&W?

Thanks in advance

-Manny

Use the lowest ISO that you have, set the camera to delayed exposure or use a remote, and shoot in Adobe RGB even for B&W scanning. If you are using a good macro lens (it seems you are) then use an aperture that will ensure enough DOF for absolute sharpness across the frame because some films bow more than others even if you have some ANR glass to hold them flat. With a dedicated macro lens you won't get diffraction so don't be afraid to use a small aperture, f/16-f/22.

 

[grain elevator]

With a dedicated macro lens you won't get diffraction so don't be afraid to use a small aperture, f/16-f/22.

Huh? There's no way around physics.

 

[Alan Edward Klein]

It matters to the extent that the light source is ideally matched to the sensor - although sufficient light output would be a priority since that's a threat to noise just as well. It does not matter to the best of my knowledge what you set the white balance to on the camera during exposure, provided the camera is recording RAW files. My initial wording wasn't careful enough; note how Alan seemed to refer to the camera white balance setting, and not so much the color temperature of the actual light source. The latter matters (a bit, at least), the former not so much IMO.



Daylight = 5000K through 6000K generally, depending a bit on which standard you go by. Anything in that range (and even quite a bit outside of it) will work fine with a modern mid- to high-end digital camera.
Here's a good start to get the basics down: https://www.cambridgeincolour.com/tutorials/white-balance.htm

The reason I mentioned white balance and light source is because the OP never mentioned his light source or its rating. It certainly will have an effect on color film scanning if not BW.

 

[Alan9940]

The reason I mentioned white balance and light source is because the OP never mentioned his light source or its rating. It certainly will have an effect on color film scanning if not BW.

A high CRI rating is best for color scanning, but I don't think it matters at all for B&W. I use a cheap LED panel to scan 8x10 film without issue.

 

[koraks]

The reason I mentioned white balance and light source is because the OP never mentioned his light source or its rating. It certainly will have an effect on color film scanning if not BW.

Yes, that's a valid concern. For B&W it doesn't matter much; to the extent it does matter, it's mostly splitting-hairs territory. Color is a slightly different story.

A high CRI rating is best for color scanning

Maybe. Given the discontinuous nature of both the color dyes in a color film as well as in a typical CMOS or CCD sensor, I doubt that a highly continuous light source is necessary per se. It probably works OK in practice and helps to keep things simple, but I bet that with a discontinuous light source and proper calibration, the same color fidelity is possible. It just requires a bit of analysis of the spectral sensitivity peaks of the camera sensor and the spectral density of the dyes involved to work out the requirements on a light source. I'm sure this has been done extensively, but I haven't yet come across credible attempts on this forum yet. It would make for an interesting exercise, albeit of moderate (at best) practical relevance.

 

[Alan Edward Klein]

Yes, that's a valid concern. For B&W it doesn't matter much; to the extent it does matter, it's mostly splitting-hairs territory. Color is a slightly different story.



Maybe. Given the discontinuous nature of both the color dyes in a color film as well as in a typical CMOS or CCD sensor, I doubt that a highly continuous light source is necessary per se. It probably works OK in practice and helps to keep things simple, but I bet that with a discontinuous light source and proper calibration, the same color fidelity is possible. It just requires a bit of analysis of the spectral sensitivity peaks of the camera sensor and the spectral density of the dyes involved to work out the requirements on a light source. I'm sure this has been done extensively, but I haven't yet come across credible attempts on this forum yet. It would make for an interesting exercise, albeit of moderate (at best) practical relevance.

Considering all the problems photographers have with scanning negative color film to get the colors right, shouldn't there be a set standard for the lighting to eliminate that as a potential issue? What do scanners use for lighting like my Epson's V600 and V850?

 

[koraks]

Considering all the problems photographers have with scanning negative color film to get the colors right, shouldn't there be a set standard for the lighting to eliminate that as a potential issue?

It won't hurt, but it won't help much either. 99% of the problems with color correction of color negative scans people run into are unrelated to the quality of the digital recording in the first place, let alone the light source used.

What do scanners use for lighting like my Epson's V600 and V850?

Older ones used a CFL, which is generally a pretty crude and discontinuous spectrum. Modern ones use LED, which is in most cases likely a lot better, but still far removed from an ideal black body radiator. I think it's an apt illustration how the light source indeed isn't the bottleneck, given the fact that you can get excellent color scans with such scanners.

 

[xkaes]

With a dedicated macro lens you won't get diffraction so don't be afraid to use a small aperture, f/16-f/22.

You will, but it varies with the lens (focal length) and the magnification. The higher the magnification, the smaller the effective aperture, and the greater the diffraction. Of course some of the diffraction is hidden with a 3D subject, but that really doesn't apply to film. As mentioned, most lenses have their best resolution at about two stops down from open, and with film that should provide enough DOF if alignment is correct.

In any event, this is a moot point -- just run a few simple tests.

 

[Alan9940]

Maybe. Given the discontinuous nature of both the color dyes in a color film as well as in a typical CMOS or CCD sensor, I doubt that a highly continuous light source is necessary per se. It probably works OK in practice and helps to keep things simple, but I bet that with a discontinuous light source and proper calibration, the same color fidelity is possible. It just requires a bit of analysis of the spectral sensitivity peaks of the camera sensor and the spectral density of the dyes involved to work out the requirements on a light source. I'm sure this has been done extensively, but I haven't yet come across credible attempts on this forum yet. It would make for an interesting exercise, albeit of moderate (at best) practical relevance.

Can't argue with any of this, but given the choice between a high CRI rating and not I'll choose the former every time because I like to delude myself!

 

[koraks]

Besides, the high CRI light definitely won't hurt either!

 

[wiltw]

You will, but it varies with the lens (focal length) and the magnification. The higher the magnification, the smaller the effective aperture, and the greater the diffraction. Of course some of the diffraction is hidden with a 3D subject, but that really doesn't apply to film. As mentioned, most lenses have their best resolution at about two stops down from open, and with film that should provide enough DOF if alignment is correct.

In any event, this is a moot point -- just run a few simple tests.

Diffraction is an absolute factor based ONLY on the f/stop used, as it is light rays bending around the diaphram leaves that accounts for diffraction amount.

'Effective aperture' is not a real physical fact, it is a concept only which serves as a means of thinking about the spread of light across a larger image circle, at macro distances...the light striking film/sensor gets dimmer only because the image circle is bigger and bigger as magnification gets higher...the number of photons coming thru is fixed, dimming light at the larger circle has nothing to do with the amount of light coming thru the hole (diaphram).

Because the magnification of the blur circle grows larger as magnificaiton of the subject gets higher. it is the APPARENT amount of blur which is more easily seen at higher magnification...no different that seeing diffraction easier at 20x24 than what you can perceive at 8x10 or at 4x5 print size. Similarly 8X enlargement of 135 (to make 8x10 from 135) allows you to use a smaller aperture on 4x5 (to make 87x10), because 4x5 neg is simply enlarged less, not because it has 'less diffraction'!

 

[250swb]

Use the lowest ISO that you have, set the camera to delayed exposure or use a remote, and shoot in Adobe RGB even for B&W scanning. If you are using a good macro lens (it seems you are) then use an aperture that will ensure enough DOF for absolute sharpness across the frame because some films bow more than others even if you have some ANR glass to hold them flat. With a dedicated macro lens you won't get diffraction so don't be afraid to use a small aperture, f/16-f/22.

Apologies for a clumsy phrase. I was trying to refer to the trade off between DOF and diffraction, in that diffraction won't degrade the image as much as DOF can improve it. As in a soft corner caused by bowed film or minutely misaligned camera won't be improved at f/5.6 but it will be improved if f/16 causes it to be sharp, hence DOF trumps a tiny amount of diffraction. Overly worrying about diffraction when people are using a home setup risks throwing the baby out with the bathwater.

Of course there are many variables, lens, distance, aperture, film flatness, rig alignment, etc. but if you are using a modern digital camera with focus peaking it's worth going from wide open to stopped down and check across the negative you are copying how much detail 'peaks' and if it's uniform. If there is nothing to be improved in hardware or flattening the negative further then whatever the first aperture is that gives most focus peaks across the plane then that's the aperture to use and not be driven by dogma. And while diffraction is a physical thing and gets trotted out as a 'rule to obey' it can't be avoided that we are making digital images from the negative, and with that comes software and programmes such as Capture's 'Diffraction Correction' sharpening, not that I've tried it because my negative scans are sharp enough.

 

[Manny A]

The light source I use is a Eyesen LED panel. Color temp 10,000k - 12,0000k. Is this adequate?

Over the weekend, scanned a few more negatives and found that setting exposure 1 stop to the right made a noticeable difference in the scan quality. A lot less grain, more contrast in the image. Settings: ISO @ 200, f/8 , 1/4s.

Thanks

-Manny