Video: 'Why We Still Love Film' by NBC

(10:23)

One thing I disagree with is the idea that film slows you down, and that this is supposed to be a good thing. I don't know that I want to be slowed down by my capture medium. OTOH, I do like the whole ritual of using film cameras.

Another point of disagreement: the idea that film will die because it's bad for the environment. Nonsense. If it dies it will be either because it gets too expensive, or because digital evolves to be a superior medium.

....or because digital evolves to be a superior medium.

Already there Karim. Has been for almost a decade now.

Show me the film with a 12 stop 'dynamic' range, or that has a genuine sensitivity above 1000 ISO and with almost no grain.

I don't know whether to laugh out loud, or feel deeply sad, when people shoot film with the sole intention of scanning it to a digital file.

Already there Karim. Has been for almost a decade now.

 

Show me the film with a 12 stop 'dynamic' range, or that has a genuine sensitivity above 1000 ISO and with almost no grain.

 

I don't know whether to laugh out loud, or feel deeply sad, when people shoot film with the sole intention of scanning it to a digital file.

No current DSLR has a dynamic range over about 9-10 stops at iso 1000. Show me a DSLR that has 18 stops like some film can achieve. Nothing wrong with scanning film with a quality scanner. I don't know whether to laugh or cry at someone trying to create a b&w image with a Bayer interpolated color sensor that throws away a lot of rez for b&w. See...works both ways. For someone that talks about digital being better all the time...you sure appear on the film forum a lot.

...Show me a DSLR that has 18 stops like some film can achieve...

You've got me curious. What film is that?

You've got me curious. What film is that?

Portra 400 and TMax. Even with Kodak P3200 and TMax developer at a rating of 3200, it is easy to get between 10-11 stops.

A number of us where in on a test at th Large Format photography forum years ago and we did TMax film in TMax developer and achieved between 17 and 18 stops of range.

For me though, I use different films because of the look.

18 stops? So if you start with an exposure of a gray card zone v at f/8, 120 sec, to get zone 18 or 13 stops, would use an exposure of f/8 at 60 seconds. to get pure white. Most people get pure white, zone 10, by opening up 5 stops, f/8 at 1/4 sec.

If a normal snow scene exposure might call for f/8 at 1/125 sec, if I wanted to raise the snow to pure white I would give a 60 second exposure at f/8? I rather fear everything else in the scene would also be blown to pure white, people's faces, etc.

18 stops? So if you start with an exposure of a gray card zone v at f/8, 120 sec, to get zone 18 or 13 stops, would use an exposure of f/8 at 60 seconds. to get pure white. Most people get pure white, zone 10, by opening up 5 stops, f/8 at 1/4 sec.

 

If a normal snow scene exposure might call for f/8 at 1/125 sec, if I wanted to raise the snow to pure white I would give a 60 second exposure at f/8? I rather fear everything else in the scene would also be blown to pure white, people's faces, etc.

18 stops

No current DSLR has a dynamic range over about 9-10 stops at iso 1000.

Nobody claimed that. And neither does any film at such an ISO speed.

A number of us where in on a test at th Large Format photography forum years ago and we did TMax film in TMax developer and achieved between 17 and 18 stops of range.

Yeah right!

Only with extremely dusty bellows that lose about 9 stops through lens and camera flare.

A directly exposed step-wedge isn't a picture. Any scene with such a brightness range would be severely compressed by lens and body flare in any practical camera. Just work out the tiny percentage of stray light from the highlights it takes to pollute a shadow down to even 12 stops of contrast.

The point I was making is that the supposed superiority of film is completely lost in an 8 bit JPEG, sRGB space reproduction - possibly down to 6 bits when viewed on many devices. So those people (many of whom call themselves 'analogue' photographers without having ever set foot in a darkroom) that claim to see nuances of difference between 2nd or 3rd generation digital film copies are completely fooling themselves, and attempting to fool others, with such nonsense.

And I have no objection to anyone using film, other than on environmental grounds. It was Karim that made the inflammatory remark about digital being inferior.

Ansel Adams when formulating the Zone system said that the useful zones on film that could be transferred to paper were zones 1 to 9. Anything above that results in blocked highlights. Take a photo of a bride in the shade. If a band of sunlight falls on her white dress creating a sunlit triangle on that dress. the dress might be in zones 8 and 9 but the triangle might be in zones 12 and 13 for the cloth and lace shadows. When printed the lace and cloth zones 8 and 9 will look good but the zones 12 and 13 would just look like a white triangle. If one tried to burn in the blocked highlights one would just get a white blob on the dress.

I enjoy looking at my large 8x10 inch negatives. There is a beauty there with the higher blocked zones that I can see on the negatives that can not be captured on a print.

Only 8-10 stops for digital? I get that with 10-bit, S-35 video, using HLG or Log-x compression. The recorded signal is compressed exponentially mostly above the 50% level, while the overall exposure is reduced by 2 stops, giving you about 6 stops of headroom. This maintains detail in highlights, to which digital imaging is particularly weak, at the expense of deep shadows, which have less useful information. A bride in the shadow, hit by streaming sunlight - no problem - along with texture in a tuxedo in the same shot.

With modern digital sensors, you can do the same thing for up to 15 stops, without the benefit of log gamma (which doesn't apply to raw images). You can underexpose to contain highlights then boost the shadows with little or no added noise.

B&W film can reach 18 stops by a similar mechanism. Film sensitivity decreases as the exposure level increases, adding a natural compression. You expose to the left in order to get detail in the shadows and let the highlights take care of themselves.

Things have changed since spreading egg white emulsions on glass plates ;)

"... I don't know whether to laugh out loud, or feel deeply sad, when people shoot film with the sole intention of scanning it to a digital file."

Why would you say something like this? The same characteristics of various films, i.e. the Kodachromes, Velvia, Ektachrome, various color negative film stocks, C41 black and white, traditional black and white, all come through in a good scan and in the digital print. They may be somewhat subtle to various degrees, but they are there. If you're saying you can achieve the same nuances in an all-digital workflow, fine, but why criticize those who utilize and may prefer another method?

There is something strange about those 18 stops shown above. If the exposure was set to get the face tones on zone 6 then snow on the background would normally be on zone 9 which in the diagram above looks like gray to me not white with some detail.

We have all seen this:

Q - When I shoot a photo of my son in a white snow suite building a snowman on a white field, the photo comes out gray. what should I do?

A- The camera meter is "fooled" by all that white and darkens the scene down to middle gray, zone v. You should open up about three stops to get the correct exposure.

According to the 18 stops shown above the answer would be:

A- The camera meter is "fooled" by all that white and darkens the scene down to middle gray, zone v. You should open up about nine stops to get the correct exposure.

I don't think so.

Why would you say something like this? The same characteristics of various films, i.e. the Kodachromes, Velvia, Ektachrome, various color negative film stocks, C41 black and white, traditional black and white, all come through in a good scan and in the digital print.

No, they don't.

The colour space of most transparency film is partly outside of the sRGB space, and the inverse of the way that digital display works.

All colour film is 'subtractive' and uses cyan, yellow and magenta dyes to simulate other colours. This means that the most saturated colours in film need a fairly high density. With film you can have saturated colours or bright (low density) colours, but you can't have both unless they're pure cyan, yellow or magenta.

The red, green and blue digital additive system allows for both saturated and bright colours, but some cyans and greens are poorly represented in the common sRGB space used by JPEGs.

Therefore it's not possible to properly display the properties of slide film in a digital reproduction, while negative film has already gone through an individual 'interpretation' process in order to be seen as a positive. Also, any digital representation can be reproduced exactly by another digital representation - meaning that a digital reproduction of a film image could be exactly simulated through manipulation of a digitally-shot original. And without the delay, expense, faff, (often) pretentiousness and waste of resources that shooting film entails.

If you want to shoot film and keep its original appearance, then wet print it, or find an old slide projector and screen it to your audience at first hand. Just don't pretend that a scan compressed and uploaded to the web is an accurate reproduction of the original bit of celluloid and gelatine.

B&W film can reach 18 stops by a similar mechanism.

Maybe, but most cameras and lenses can't.

Let's take a near best-case example: We have a perfect lens (no flare or light losses) pointed at a bright object against a perfectly black ground - say, the moon on a night and at a location with perfect seeing. That image is then projected into a real camera, with the moon taking up just 10% of the frame area. During exposure, hopefully the film or digital sensor will be the most reflective surface in the dark chamber. With a reflectance, say, of 18%.

Now let's run a few numbers. The moon, taking up 10% of an 18% reflective surface, re-radiates (10%*18%) 1.8% of its light back into the dark chamber. The dark chamber in turn has a 1% reflectivity (being extremely optimistic about the quality of flocking or black paint). So the total percentage of light bouncing around the dark chamber and hitting the shadows of the image is: 10% * 18% * 1% = 0.018%. This is 1/5556th of the moon brightness, or as a contrast ratio 5,556:1, or as an exposure range it's 12.44 stops.

The situation gets even worse with a daylight pictorial scene. Then we have a whole frame brightness that averages around 18% and our least shadow brightness then becomes 18% * 18% * 1% = 0.0324%, or 3,086:1, or just under 11.6 stops.

Remember, we haven't yet factored in any internal lens flare, nor the extraneous image circle outside of the frame boundary. That's at least another 60% * 18% * 1% * 1% of stray light bouncing around.

You may well disagree with those hypothetical reflectance figures, but even if you halve or quarter them, you're still a country mile away from getting an 18 stop brightness range onto the film/sensor surface.

No, they don't.

The colour space of most transparency film is partly outside of the sRGB space, and the inverse of the way that digital display works.

 

All colour film is 'subtractive' and uses cyan, yellow and magenta dyes to simulate other colours. This means that the most saturated colours in film need a fairly high density. With film you can have saturated colours or bright (low density) colours, but you can't have both unless they're pure cyan, yellow or magenta.

 

The red, green and blue digital additive system allows for both saturated and bright colours, but some cyans and greens are poorly represented in the common sRGB space used by JPEGs.

 

Therefore it's not possible to properly display the properties of slide film in a digital reproduction, while negative film has already gone through an individual 'interpretation' process in order to be seen as a positive. Also, any digital representation can be reproduced exactly by another digital representation - meaning that a digital reproduction of a film image could be exactly simulated through manipulation of a digitally-shot original. And without the delay, expense, faff, (often) pretentiousness and waste of resources that shooting film entails.

 

If you want to shoot film and keep its original appearance, then wet print it, or find an old slide projector and screen it to your audience at first hand. Just don't pretend that a scan compressed and uploaded to the web is an accurate reproduction of the original bit of celluloid and gelatine.

You do realize scanners can scan in 16 bit and higher with other color spaces than sRGB and output to better file type than JPG...thus invalidating everything you just said. You may want to read up more on the hybrid workflow...because what you wrote is woefully shallow on the topic.

Maybe, but most cameras and lenses can't.

Flare would tend to limit the low end of the characteristic curve, with relatively little effect on the dynamic range. Most of the dynamic range of B&W film is in the upper reaches of density. That said, it is the abscissa, exposure, not the ordinate, density, which is a measure of dynamic range.

Published curves for Tri-X start near zero and stop at a density of about 3.0, at an exposure range of 3.5 log(lux-s), the latter corresponding to a dynamic capture range of 12 stops. At that point, the response is almost linear, which means the shoulder of the curve is unknown. When completely fogged, Tri-X reaches a density of at least 5.0, which represents 17 stops off the film. In order to have a capture range of 18 stops, you would need an exposure range of 6 log(lux-s), which is not unreasonable, if not actually practical. Processing which reduces the slope (gamma) is practical, however, and would have a profound effect on dynamic range.

Flare would tend to limit the low end of the characteristic curve, with relatively little effect on the dynamic range. Most of the dynamic range of B&W film is in the upper reaches of density. That said, it is the abscissa, exposure, not the ordinate, density, which is a measure of dynamic range.

 

Published curves for Tri-X start near zero and stop at a density of about 3.0, at an exposure range of 3.5 log(lux-s), the latter corresponding to a dynamic capture range of 12 stops. At that point, the response is almost linear, which means the shoulder of the curve is unknown. When completely fogged, Tri-X reaches a density of at least 5.0, which represents 17 stops off the film. In order to have a capture range of 18 stops, you would need an exposure range of 6 log(lux-s), which is not unreasonable, if not actually practical. Processing which reduces the slope (gamma) is practical, however, and would have a profound effect on dynamic range.

I learned about processing various b&w films when large DR is needed from a workshop with Bruce Barnbaum. He spoke at length on the topic when we discussed how he captured a range of 15-16 stops when photographing in Antelope Canyon back in the 80's and 90's. Issues where how to meter and expose the film...and then process to capture the range. One is definitely not limited by lens flare, etc in this regard. His work, experience, and final output where proof of that.

Getting it on film is one thing. Putting the results on paper, which has a dynamic range of 6 stops or so, is an art.

Or more so a craft (which shouldn’t be underrated). Art is doing it with some sort of passion and/or expressiveness, one or two stops beyond dynamic range. :)

No, they don't.

The colour space of most transparency film is partly outside of the sRGB space, and the inverse of the way that digital display works.

 

All colour film is 'subtractive' and uses cyan, yellow and magenta dyes to simulate other colours. This means that the most saturated colours in film need a fairly high density. With film you can have saturated colours or bright (low density) colours, but you can't have both unless they're pure cyan, yellow or magenta.

 

The red, green and blue digital additive system allows for both saturated and bright colours, but some cyans and greens are poorly represented in the common sRGB space used by JPEGs.

 

Therefore it's not possible to properly display the properties of slide film in a digital reproduction, while negative film has already gone through an individual 'interpretation' process in order to be seen as a positive. Also, any digital representation can be reproduced exactly by another digital representation - meaning that a digital reproduction of a film image could be exactly simulated through manipulation of a digitally-shot original. And without the delay, expense, faff, (often) pretentiousness and waste of resources that shooting film entails.

 

If you want to shoot film and keep its original appearance, then wet print it, or find an old slide projector and screen it to your audience at first hand. Just don't pretend that a scan compressed and uploaded to the web is an accurate reproduction of the original bit of celluloid and gelatine.

This does not agree at all with my experience and I have been passionate about the whole process for over fifteen years. I see the respective characteristics, as previously mentioned, as do other experienced people I've associated with.