High contrast film images coming out totally different from digital test shots

[peterB1966]

I did some test shots of a high contrast scene (I wanted the background blown out) on my Nikon D610 and once happy, converted the settings to equivalents on my Mayima RB67. To my surprise the shadow areas in two of the photos are way darker on film than on digital (the two photos of the white woman. Strangely the portrait of the black woman came out quite similar). Is this normal with film (I shot with TMY2 ISO400)?

The attached images will show the digital image (with exposure info) on the left, unprocessed but converted to mono in order to simplify comparison), and the right hand side is an unprocessed scan of the same same shots from film, with the keyword block showing the settings I used (as best as I/my notes can recall). You will see though that there appears to be no clipping in the shadow areas (sorry, the same histogram appears to have been posted in both of the first images, but nonetheless neither of them are clipped).

Disclaimer: I am new to using an MF camera, so I am not sure if I made mistakes in my my conversions from digital test shots to film (the discipline of note-taking is new to me as a digital user).

 

[Frank53]

I don’t if it’s a good idea to expose film based on a digital exposure.
What I do know that you cannot overexpose digital, because you lose information, while you can easily overexpose film 5 or 6 stops and still get a reasonably print. Digital is a bit “better” with underexposure.
Just use an exposure meter.
Regards,
Frank

 

[Lachlan Young]

Effectively what you are finding is that digital camera exposure is set up to protect highlights (with the benefit of lower noise in the shadows allowing them to be opened somewhat), whereas BW negs are intended to be exposed such as to deliver suitable shadow detail - because you can then use process and printing controls to adjust the highlight reproduction to taste. I wouldn't describe the situation you outline in terms of lighting 'high contrast', but rather a pretty standard setup where the background is to be high density on the neg such that it reproduces as white.

 

[grain elevator]

Are you sure the scanner (driver) isn't trying to protect the highlights you wanted blown out, in the process darkening everything else?

 

[peterB1966]

I don’t if it’s a good idea to expose film based on a digital exposure.
What I do know that you cannot overexpose digital, because you lose information, while you can easily overexpose film 5 or 6 stops and still get a reasonably print. Digital is a bit “better” with underexposure.
Just use an exposure meter.
Regards,
Frank

Thanks, Frank - I thought I would use my digital camera the way people used to use polaroids, for exposure testing. Of course I am metering as well, but due to the need to expose for the ambient light in front of the model as well as the blow out background, some test shots might save me time and money.

 

[peterB1966]

Effectively what you are finding is that digital camera exposure is set up to protect highlights (with the benefit of lower noise in the shadows allowing them to be opened somewhat), whereas BW negs are intended to be exposed such as to deliver suitable shadow detail - because you can then use process and printing controls to adjust the highlight reproduction to taste. I wouldn't describe the situation you outline in terms of lighting 'high contrast', but rather a pretty standard setup where the background is to be high density on the neg such that it reproduces as white.

So does that imply this is normal? I would have thought it was gonna show me what I metered for, not make subsequent decisions :-/

 

[peterB1966]

Are you sure the scanner (driver) isn't trying to protect the highlights you wanted blown out, in the process darkening everything else?

I have no idea as it was commercially processed and scanned, but it's a damned good question ... I will ask them.

 

[grain elevator]

I have no idea as it was commercially processed and scanned, but it's a damned good question ... I will ask them.

You have the negatives, right? What shadow detail is in them? There should be a banner "show us the negatives" that pops up every time someone is about to post a new thread

 

[peterB1966]

I have no idea as it was commercially processed and scanned, but it's a damned good question ... I will ask them.

You have the negatives, right? What shadow detail is in them? There should be a banner "show us the negatives" that pops up every time someone is about to post a new thread

Ha ha, ok yeah, I do have them.

Unfortunately I have no skill at reading their quality, nor do I have any idea of how to show them to you if I am not going to scan them, so I have simply shot them with my iPhone on a temporary iPad lightbox... hope that will suffice?

 

[MattKing]

After a bit of digital futzing:

I think what you are seeing is the result of a number of factors, not the least of which is that your meter reading is being overwhelmed by the backlighting.
The light that is actually hitting the front of your subject is barely enough to under-expose it. The digital tools can be used to eek out a usable image from the negative, but it is a stretch!
This is an example of a situation where an in-camera meter is barely suitable for what you need to do.
I expect you metered from the camera position. It would have been better if you had got closer and took a reading off of a small portion of your subject.
Notice how the exposure is better in the example where you are closer and your subject fills more of the frame. It was probably also helped by the fact that her skin tone is much darker - that will tend to cause the meter to suggest more exposure.
It isn't a result (so much) of your using a digital camera for metering. It is a result of using an in camera meter when an incident meter would be much better.
By the way, while having a background blow out into white is an effective tool, it is important to make sure that you have good light on the front of the subject as well. Unless reflectors are employed carefully, or there is an additional source of light reaching the front, this sort of lighting makes a lot of people look like they are deceased (or near to it)!
Really expensive models being the exception .

 

[peterB1966]

After a bit of digital futzing:
...I think what you are seeing is the result of a number of factors, not the least of which is that your meter reading is being overwhelmed by the backlighting.
The light that is actually hitting the front of your subject is barely enough to under-expose it....

Thanks for your reply, Matt, but at the risk of sounding defensive, if you look at the digital version of this image on my first post, you will see it is not massively under-exposed (it appears more so due to the white background; see below for the processed images: digital on LHS, film on RHS), so my question was not whether it was under-exposed, but rather about why did an equivalent exposure triangle on film result in it being so unexposed when one would expect it to be the same or similar.

 

[gdavis]

I did some test shots of a high contrast scene (I wanted the background blown out) on my Nikon D610 and once happy, converted the settings to equivalents on my Mayima RB67. To my surprise the shadow areas in two of the photos are way darker on film than on digital (the two photos of the white woman. Strangely the portrait of the black woman came out quite similar).

According to your notes, the digital shot of the black woman was with a lower ISO, so it had less exposure than the film shot (assuming the shutter speed is correct, but you put a question mark by it). That might explain why those shots match better.

The RB67 has a bellows focus. The closer the subject, the more the bellows extends, the darker the exposure will be. This needs to be compensated for compared to your digital exposure.

As Frank noted, digital will be better at preserving underexposed shadows but will clip highlights. Film is the opposite and will clip underexposed shadows while preserving overexposed highlights better.

In the negative images, the density of the skin tones looks about the same in all three meaning the white woman will have about the same tone as the black woman. There should be more density in the white woman's skin.

 

[MattKing]

If we are referring only to the people (and not the background) both versions are technically under-exposed. The subject (excluding the background) has a relatively narrow SLR (Subject Luminance Range) and both the film and digital image are squeezing that SLR into the low end of the "curve". Digital images have a bit more latitude at the under-exposure end, so after the digital image processing happens they look a bit better then the film versions look after the scans undergo digital processing.
In both cases - digital and film - the technical quality of the images would be improved with more exposure, accompanied by post processing that darkens the final result. The backgrounds would be even more blown out, but if I understand things correctly, that would be okay with you.
I stress "technical quality", because it certainly is within your purview to intentionally squeeze that SLR into the lower part of the curve, in order to achieve the aesthetic result you have here.
As for the significant initial difference of appearance between the digital file and the result of the film scan, a lot of that is a result of the scanning step - these sorts of backlit subjects often require that any automatic or semi-automatic scanning controls be over-ridden. Otherwise, the scanning process will attempt to protect too much detail in the backgrounds, at the expense of the shadowed foreground.

 

[gone]

Digital is digital and film is film. You can use a digital camera as a light meter to get film's exposure, but that's about it. Think about it: film is a light sensitive emulsion, and the emulsions and degrees and types of light sensitivity are different for every brand and every type of film. Digital has a sensor that is exposed to light, but the way it reads that light is in no shape or form like a film emulsion. It just isn't going to work.

 

[peterB1966]

If we are referring only to the people (and not the background) both versions are technically under-exposed. The subject (excluding the background) has a relatively narrow SLR (Subject Luminance Range) and both the film and digital image are squeezing that SLR into the low end of the "curve". Digital images have a bit more latitude at the under-exposure end, so after the digital image processing happens they look a bit better then the film versions look after the scans undergo digital processing.
In both cases - digital and film - the technical quality of the images would be improved with more exposure, accompanied by post processing that darkens the final result. The backgrounds would be even more blown out, but if I understand things correctly, that would be okay with you.
I stress "technical quality", because it certainly is within your purview to intentionally squeeze that SLR into the lower part of the curve, in order to achieve the aesthetic result you have here.
As for the significant initial difference of appearance between the digital file and the result of the film scan, a lot of that is a result of the scanning step - these sorts of backlit subjects often require that any automatic or semi-automatic scanning controls be over-ridden. Otherwise, the scanning process will attempt to protect too much detail in the backgrounds, at the expense of the shadowed foreground.

Thanks - there's no doubt that I could have used more light on the the subject. There was someone else who also mentioned the scanner possibly compensating - I have sent an email to the people who did it, and will wait to see what they say: it would certainly go some way towards explaining the difference.

 

[peterB1966]

Digital is digital and film is film. You can use a digital camera as a light meter to get film's exposure, but that's about it. Think about it: film is a light sensitive emulsion, and the emulsions and degrees and types of light sensitivity are different for every brand and every type of film. Digital has a sensor that is exposed to light, but the way it reads that light is in no shape or form like a film emulsion. It just isn't going to work.

In which case why would you use the digital camera's meter at all? I'm not just being facetious, it just seems strange to suggest that you can use the meter, but you'd be crazy to expect similar results!

 

[MattKing]

Thanks - there's no doubt that I could have used more light on the the subject. There was someone else who also mentioned the scanner possibly compensating - I have sent an email to the people who did it, and will wait to see what they say: it would certainly go some way towards explaining the difference.

Scanning is an art all its own, and lab scans are frequently found to be wanting.
I have a difficult relationship with the scanning part of a hybrid workflow. Sort of like a parent with a difficult teenager .
For clarity, and in response to the initial part of your question, one can use a digital camera to provide useful metering and exposure information, but you shouldn't expect it to be simple or automatic.
For instance, what metering mode were you using, and what metering pattern does your digital camera apply in that circumstance?
All of that information is information that you would want to apply when you interpret the data, in support of choosing the camera settings you intend to use on the film camera.
You do need to be careful with evaluative or matrix or similar in-camera metering systems - one really has no way of predicting what images in their data banks are being used!
Most likely if you use manual metering, become very familiar with the meter pattern of your camera and do some tests you will arrive at a situation where you can predict how best to use the digital camera's readings on your film camera - you will have "calibrated" that process. And you may very well be able to use those readings directly in a lot of different lighting situations. But the lighting conditions will matter, and I personally wouldn't start that calibration process with an inherently difficult first example - a strongly backlit subject with relatively low and low contrast front lighting.

 

[peterB1966]

Scanning is an art all its own, and lab scans are frequently found to be wanting.
I have a difficult relationship with the scanning part of a hybrid workflow. Sort of like a parent with a difficult teenager .
For clarity, and in response to the initial part of your question, one can use a digital camera to provide useful metering and exposure information, but you shouldn't expect it to be simple or automatic.
For instance, what metering mode were you using, and what metering pattern does your digital camera apply in that circumstance?
All of that information is information that you would want to apply when you interpret the data, in support of choosing the camera settings you intend to use on the film camera.
You do need to be careful with evaluative or matrix or similar in-camera metering systems - one really has no way of predicting what images in their data banks are being used!
Most likely if you use manual metering, become very familiar with the meter pattern of your camera and do some tests you will arrive at a situation where you can predict how best to use the digital camera's readings on your film camera - you will have "calibrated" that process. And you may very well be able to use those readings directly in a lot of different lighting situations. But the lighting conditions will matter, and I personally wouldn't start that calibration process with an inherently difficult first example - a strongly backlit subject with relatively low and low contrast front lighting.

OK, a lot of that makes sense. There was no small amount of hubris involved in trying such a difficult exposure, but I was under the impression that if I got the exposure right on my digital camera all I needed to do was transfer the settings across to the film camera.
I do have an incident meter, so I guess the proper step would be to get the exposure right with an incident reading off the subject, and then make the background at least 3 to 4 stops brighter...?
(And in the meantime I wait to see what the people who did the scans say to my enquiry.)

 

[MattKing]

OK, a lot of that makes sense. There was no small amount of hubris involved in trying such a difficult exposure, but I was under the impression that if I got the exposure right on my digital camera all I needed to do was transfer the settings across to the film camera.
I do have an incident meter, so I guess the proper step would be to get the exposure right with an incident reading off the subject, and then make the background at least 3 to 4 stops brighter...?
(And in the meantime I wait to see what the people who did the scans say to my enquiry.)

Here is an experiment to try.
Take a reading using the incident meter, and then use the results of that reading on both the digital camera and the film camera. For reference, also take a reading in the same way you did before with the digital camera.
Then process the digital file with an aim to matching the tone of, as an example, the armband on the middle subject - a nice middle grey.
See how the results compare to your first try. And note how the readings differ between the incident meter and the digital camera.
I'm guessing that the flexibility inherent in digital processing will mean that your digital results will be subtly different from the ones you showed us, but your negatives will be substantially different from the ones you showed us. With that lighting though, they will still be challenging for your lab to scan, unless they are prepared to provide a fairly high level of custom scanning services.
FWIW, based on what I see I could make decent prints from the second and third negatives you display above, and could probably make a decent scan of them as well, but I still would prefer the results if you had used some more front lighting .
And by the way, those are really well displayed negatives - much better than a lot of attempts to display negatives we see here.

 

[glbeas]

Another factor to consider is the exposure response or “curves” if you will of the film versus digital. Film has a toe at the bottom end where the density builds slow then as it gets to the midrange its close to a straight line then rounds to a plateau or shoulder as it approaches overexposure. Digital shot at the same values generally has a more straight line response from shadow to highlights till it starts blocking due to maxing out the sensor. So even if you get one end or the other of the tonal scale to match its going to look different.
Certain films like Tmax have a much straighter response than the older type of emulsions so if you want to work this way you might try those films and see how you like it.

 

[wiltw]

One thing that you might be seeing about digital vs. film in the shadow areas is that film sensitometry follows what is called 'the S-curve' while digital is fundamentally limear in its response, unless you tell your software to use a S-shaped curve applied to the photo.

Another factor to consider is the exposure response or “curves” if you will of the film versus digital. Film has a toe at the bottom end where the density builds slow then as it gets to the midrange its close to a straight line then rounds to a plateau or shoulder as it approaches overexposure. Digital shot at the same values generally has a more straight line response from shadow to highlights till it starts blocking due to maxing out the sensor. So even if you get one end or the other of the tonal scale to match its going to look different.
Certain films like Tmax have a much straighter response than the older type of emulsions so if you want to work this way you might try those films and see how you like it.

Indeed, thanks for bringing that up...I was going to do that but illustrate with a graph with superimpose an S-curve and a linear response, but I could not fiture out how to execute the S-curve drawing. In postprocessing of digital images, there is a default linear response to brightness levels, and you need to deliberately apply an S-surve shaped response, to more closely mimic films.

 

[pbromaghin]

OK, a lot of that makes sense. There was no small amount of hubris involved in trying such a difficult exposure, but I was under the impression that if I got the exposure right on my digital camera all I needed to do was transfer the settings across to the film camera.
I do have an incident meter, so I guess the proper step would be to get the exposure right with an incident reading off the subject, and then make the background at least 3 to 4 stops brighter...?
(And in the meantime I wait to see what the people who did the scans say to my enquiry.)

I hope you aren't saying you will point your incident meter at the subject, but that you will hold the meter very close to the subject while pointing at the camera.

 

[Craig75]

Are they shot with flash or continuous light?

They look.like flash with catchlights in pupils but my flash game is very weak so that could be garbage.

The shutter speeds are too fast for a lot of film cameras + flash

 

[peterB1966]

Here is an experiment to try.
Take a reading using the incident meter, and then use the results of that reading on both the digital camera and the film camera. For reference, also take a reading in the same way you did before with the digital camera.
Then process the digital file with an aim to matching the tone of, as an example, the armband on the middle subject - a nice middle grey.
See how the results compare to your first try. And note how the readings differ between the incident meter and the digital camera.
I'm guessing that the flexibility inherent in digital processing will mean that your digital results will be subtly different from the ones you showed us, but your negatives will be substantially different from the ones you showed us. With that lighting though, they will still be challenging for your lab to scan, unless they are prepared to provide a fairly high level of custom scanning services.
FWIW, based on what I see I could make decent prints from the second and third negatives you display above, and could probably make a decent scan of them as well, but I still would prefer the results if you had used some more front lighting .
And by the way, those are really well displayed negatives - much better than a lot of attempts to display negatives we see here.

Quite a bit to respond to here! First of all, thanks for all the input, and the feedback on my negatives - was wondering if they were ok or if they were making me look a total rube (to use an Americanism)

Yes, I think it's a great idea: I will use the incident meter to do a test shot using my grey card as a yardstick, at 1/400th as well as 1/200th just in case the PC cable is causing issues at higher syncs (which I doubt, based on some of the other photos from the same roll). Thanks.

 

[peterB1966]

Another factor to consider is the exposure response or “curves” if you will of the film versus digital. Film has a toe at the bottom end where the density builds slow then as it gets to the midrange its close to a straight line then rounds to a plateau or shoulder as it approaches overexposure. Digital shot at the same values generally has a more straight line response from shadow to highlights till it starts blocking due to maxing out the sensor. So even if you get one end or the other of the tonal scale to match its going to look different.
Certain films like Tmax have a much straighter response than the older type of emulsions so if you want to work this way you might try those films and see how you like it.

Yes, I was concerned if that played a role at all, which is why one of the reasons I posted here, and why I mentioned which film I used. Currently testing a roll of Tmax, so this should be interesting (and possibly more satisfying for my coddled digital sensibilities!)