Old lenses, diffusion filters and the elusive "film look"

[Yaeli]

Hi all !

So, for many reasons (the details of which I will spare you), I like film photography and what film gives me in terms of "look". I've also always sucked at editing, especially color, so film is great for that too - even though the "look" here is really dependent on the lab and the technician doing the scanning. But, I'm a lazy middle aged guy, with not a lot of time or money to spare, and I'm not really good at delaying gratification (I don't chimp, but I like to see my pictures right away when I get home). 6 months without a digital camera showed me one thing at least : I almost never shoot on film, and the rolls "mold" in my many SLR cameras. So basically, I want "the butter and the money from the butter and the dairywoman's buttocks", as we say in France (ie, I want to have my cake and eat it) : I want the look of film but the convenience of digital.

I noticed, sorting through all my old pictures, that the ones I like the most were, almost always, taken with my older EF lenses (especially the 85mm and 50mm F/1.8), and not with the - albeit great - RF 24-105mm F/4 L (too clinical and sharp for a "film look"). I'm using Mastin Labs presets, which, even though they're probably not 100% how the real films look like, have given me great results so far, and have taken from me the weight of having to create a look of my own. I'm happy to rely on them, even though I wish they all had (especially the B&W presets) a "pushed" version. And I've recently heard about diffusion filters, and I kinda like the look they give when they're not overpowering. And finally, I have a plethora of old analog lenses (OM Zuiko, M42, Minolta MC and EF) that I can adapt on my EOS R.

So, my questions are :
1) For that film look that I like, should I go for old lenses or new lenses + diffusion filter or even old lenses + slight diffusion ?
2) If I go for old lenses, will I see a big difference between older manual lenses and slightly more recent EF lenses ? - knowing that I shoot quite lot of moving subjects, for which AF might come in handy, and that I'm not sure that manual focusing can be as fast on my EOS R than it is on a SLR body with split prism.
3) For those who use presets like me, how would you achieve that pushed look ? I usually try to add contrast and a bit of grain, but it's nowhere near as nice as what I got with real film pushed...

Thank you beforehand for your answers
Yael

 

[wiltw]

I am not convinced that there was a 'film look', apart from certain films having certain characteristics (e.g. Velvia high color saturation and contrast).
Print Kodak color neg on Fuji paper might look 'different' from same color neg printed on Kodak paper, but was there a 'Kodacolor-X 'look' that could be identifiable reliably? OK, Kodachrome has someswhat warmer character than Ektachrome, but there is not a 'film color slide look', is there?

Film era lenses before the 1970s did not have the fancy multicoating, so there is some loss of contrast more apparent than modern lenses in the same demanding circumstances, but is that even apparent when the light conditions are average and typical (like inside a living room)? Lenses used to have 5 or 6-sided diaphram openings, and not the near-circular openings commonly found in most modern lenses, so you can tell the difference when highlights out of focus display the pentagon or hexagon shapes! But you get that same effect mounting same lens on digital camera as on film SLR.

I think folks perceive a 'film look' simply because they view 60 year old prints that have color shifted and lost contrast simply due to aging of dyes in the emulsions, not because the prints looked that way 60 years ago!

The only thing not easily simulated is film grain...digital simulation fails to make the density of grain change with tonality, as it can be seen on the film enlargement. The 'clumping' of grain alters that apparent spacing, less clumping in lighter tone areas.

 

[juan]

I was reading an article a few months ago that explained how the makers of modern glass lenses take advantage of the newest sensors ability to maximize contrast. This micro-contrast, for lack of a better word, results in what most people see as sharpness. It’s the glass, not the coatings. I realize I hate that look. To me, it’s unnatural and makes photos fungible.
On my Nikon, I have gone back to my old 1970s lenses. To me, they have a smoothness that’s lacking in modern lenses.
I’d say try your old lenses and see if you like them.

 

[radiant]

I have been photographing with old Canon FD lenses on my Fujifilm X-T3. One even has horrible fungus and some separation. But still those look pretty digital. What is shows that lens needs to be really really broken so that it has clear effect on the outcome.

I've been playing around with the anti-halation + grain simulations available, see the other thread about this:

Digital like analog

Hello. Long time digital shooter, dipping their toe back into analog. I was just a child when analog started fading away and don't have much experience in the medium. I'm curious about something from people with more experience in both worlds. In your opinion, what digital camera produces an...

www.photrio.com

and

Digital like analog

A much simpler and much more general scanner could consist of a backlight that could be selectively cycled through RGB with spikes according to the RA4 paper characteristic. Or even simpler, an electronic flash source, plus a removable filter that converts the output to the correct RA4...

www.photrio.com

 

[Yaeli]

Thank you for your answers !

@wiltw : you raise very valid points concerning the grain and the fading of old photographs, and I agree with you on both. I should also probably have said "filmic look" and not "film look". As you very well point out, there are plenty of different films, there's no one film look. And I agree with you, if there is a film look, it is not defined by the look of aged photographs. If you look at pictures taken by Meyerowitz, Leiter or Weber, you can clearly see that nothing is faded of lacking in contrast, so that's not what I meant.
As for the existence of a look, I tend to think that we agree, even though we don't phrase it in the same way Let me explain : what I mean by filmic look is precisely the characteristics of film in general, and of some in particular : grain, of course, but also highlight rolloff, non-eccentric color palette (except for Aerochrome and a few like that), etc...
Take another example : If I say : "Old Mustangs have a certain look that I prefer, compared to the new ones", how is that different from saying : "old Mustangs have certain characteristics (boxier shape, round headlights, etc...) that I prefer, compared to the new ones" ?
What I can tell you is that, when I used to have Instagram, I would scroll through the pictures and, 8 times out of 10, when I stopped on one that I liked, I then noticed that it had been taken on film. So I think there is something identifiable that makes them stand out (to my eyes at least). And that identifable something is the characteristics that you and I mentioned, I think. I hope I'm making myself clear
Edit : there is also the whole question of the lenses, the perspective and DOF of MF and LF (and the movements, in the case of LF). All of that makes a difference also, I think.
As for lenses, I agree with you on the coatings and aperture blades. But I assure you that the same picture taken on my 85mm f/1.8 (design dating back to the 90's) and on my RF 24-105 L at 85mm look incredibly different, even in average light conditions. The RF, as @juan pointed out, is incredibly sharper and crispier.
I don't mean to dismiss what you say, on the contrary I agree with you (or at least, I think so ^^).

@juan : that is exactly my opinion also. That's why I will go back to using my older EF lenses. I'll give a try at the older manual lenses that I have, but I'm afraid that, for the kind of photography that I do (portraits on the fly outdoors, skydiving, and documentary), the lack of AF will make me miss a lot of shots...

@radiant : thank you for you input and for the links, I will go and check that out !

 

[Yaeli]

@radiant : So, I went and checked the links you gave and, from what I understand from the discussion, most people agree that the way film images look cannot be replicated with digital. You can come close, but grain and halation cannot (for now) really be reproduced. Is that what I was supposed to get from those discussions ? If so, I 100% agree with that. I believe that film has certain characteristics (thanks @wiltw for saying it this way !) that cannot completely be replicated with digital. It's even clearer if you factor in the different printing papers, developpers and processes : I've yet to see a digital image that even remotely looks like a tintype or a lith process. My point is not to achieve that perfect replication, but to come close enough that my images will be more pleasing to my eyes than their original digital version. My goal is to, first, take a more analog approach to shooting (I already don't chimp, because I like to stay in the moment, but I'm also planning on shooting less - and particularly not with burst mode, on taking more care with composition and direction and quality of light, etc...) and then to use the softwares to emulate what can be emulated from film, knowing full well that it won't ever look like real film. I know myself, I won't have neither the discipline nor the patience to shoot on film (I've tried and I'm just too lazy and don't like the constraints it gives me), so I'm trying to find a decent middle ground. I hope it makes sense
P.S : one of the French photographers whose work I like most on Flickr uses various combinations of developpers and papers and processes and achieves a sense of atmosphere and emotion that would be quite absent from identical shots taken on digital : https://www.flickr.com/photos/chomski69
And if I take one of the photographers I like most, David Burnett, 99% of his shots that I like were taken on film, even the recent ones, like this one : https://media.newyorker.com/photos/5e18fad2c59629000922a1c1/4:3/w_3759,h_2819,c_limit/burnett-08.jpg

 

[Yaeli]

Hi all !

So, here's what I'm gonna do, in case someone is interested. I'll make the experiment myself. I'll buy 1 roll of Tri-X, one of Portra 400, one of HP5 and one of Ektar. I'll shoot them alongside my EOS R, with the same lens and same settings and same subject and lighting conditions. Then I'll edit my Raw files with Mastin Labs presets for these rolls, and then I'll scan the negatives and will do the comparison. That way, I'll know (and maybe some that are interested too) if film is really "worth it" (worth the effort and money, for ME that is) or if I can get away with digital emulation. It'll take a while, but I'll let you know. Cheers

 

[grat]

Well, grain is very compact and highly irregular, whereas the bayer sensor on most cameras is very compact and very regular. So that's going to be an issue, although you can add artificial grain in post, and you could, in theory, capture an image of an 18% grey card, scan at very high resolution, and capture the grain "detail" to be applied to other photos.

If you look up the spectral response curves for a given film, you could probably duplicate that and create a LUT file, if you're very good with photo editing software.

The final difference is going to be that as a rule, film is exposed to maximize shadow detail, and digital is exposed to limit highlight clipping, so you'd need to constrain your shadows / highlights to fit in the overlapping region. And I have no idea if that sentence made any sense.

I suspect with a great deal of effort, you could produce something "close" to the Film Look-- but it might just be easier to shoot film and scan your negatives.

 

[radiant]

So that's going to be an issue, although you can add artificial grain in post, and you could, in theory, capture an image of an 18% grey card, scan at very high resolution, and capture the grain "detail" to be applied to other photos.

Grain doesn't work that way. It is not uniform "noise". Grain makes the "tones" we see. Less grain = darker, more grain=lighter.

 

[grat]

I understand that. But using layers and blending, you can accomplish some of that effect. Ideally, blend it with the "detail" layer of the photo using the color layer to control the blending, so that it affects the contrast, while being controlled by the color (or luminance, if you prefer).

 

[Yaeli]

Thank you for your answers

Yes, as you both mention, grain remains a problem. I might be wrong here, but I also tend to think, looking at my pictures taken on film, that grain is also different depending on the area of focus. It seems to me that it is less visible where the focus is and more visible in the out of focus areas. But I might be wrong....

@grat : I don't necessarily want the exact perfect "filmic look". If I can, and because I'm a lazy guy, I'd simply like to get close enough for my taste. But if my experiment shows me that I can't get close enough, then I'll do as you mention : I'll shoot film and scan the negs

 

[wiltw]

Hi all !

So, here's what I'm gonna do, in case someone is interested. I'll make the experiment myself. I'll buy 1 roll of Tri-X, one of Portra 400, one of HP5 and one of Ektar. I'll shoot them alongside my EOS R, with the same lens and same settings and same subject and lighting conditions. Then I'll edit my Raw files with Mastin Labs presets for these rolls, and then I'll scan the negatives and will do the comparison. That way, I'll know (and maybe some that are interested too) if film is really "worth it" (worth the effort and money, for ME that is) or if I can get away with digital emulation. It'll take a while, but I'll let you know. Cheers

keep in mind to NOT USE the linear response that is often the default with digital processing, but use an s-curve to mimic the response of emulsions.

 

[radiant]

I don't necessarily want the exact perfect "filmic look". If I can, and because I'm a lazy guy, I'd simply like to get close enough for my taste. But if my experiment shows me that I can't get close enough, then I'll do as you mention : I'll shoot film and scan the negs

Yes it is a long way to "close enough" - you need to apply all kind of stuff. LUTs and noise isn't just enough. That is what my Fujifilm X-T3 does. It gives quite a nice vibe but it is still quite far from analog look (even with old lenses).

 

[Yaeli]

Thank you for your answers

@wiltw : I'm not sure I understand what you mean... I had planned on using my EOS R for "scanning", and then using a plugin in LR to convert the negs. Does that "linear response" still apply to that process ?

@radiant : well, that's what I want to discover. I even wonder if you can get close enough to a light sensitive material such as film, with a random distribution of particles of silver that differ in size, using a digital sensor with organized cells of the same size... Maybe it simply doesn't capture the light and colors in the same way, and thus cannot be replicated...

 

[radiant]

@radiant : well, that's what I want to discover. I even wonder if you can get close enough to a light sensitive material such as film, with a random distribution of particles of silver that differ in size, using a digital sensor with organized cells of the same size... Maybe it simply doesn't capture the light and colors in the same way, and thus cannot be replicated...

One of these grain simulations is actually superimposing the digital image or interpolate while running. Then run the grain emulation on that data. After that - resize back to normal size. The problem is that it takes memory and more importantly a lot of processing time. That way you can sort of emulate the grains.

There is some good analysis here: https://www.filmscanner.info/en/Aufloesung.html - 40-60 million dots of 135 film. And I believe they are referring to the grain "clumps" we see as "grains". Well, again: actually the spaces between.

That just gives some insight how detailed the grain simulation should be to achieve the same look, I think. And that is just ridicilously slow if you are using for example gaussian distribution / distance grouping to achieve the clumping effect.

 

[guangong]

Wouldn’t it be easier to simply use film?

 

[radiant]

Wouldn’t it be easier to simply use film?

This discussion is in 100% digital forum.

 

[Yaeli]

Thanks for your answers

@radiant : I'm afraid this is a bit too technical for me I'm not gonna go into that much detail, I just want to see if I can get close enough for my taste. It won't be the real look, I know that. But (and that also answers the question of @guangong ) if I can get something that satisfies me, that looks close enough to film, but lets me choose between color and b&w depending on my wish, change my ISO whenever I want, and review my photos right when I get home without any additional time or cost needed, I know (again, because I'm lazy) that I will choose this, even if it's not real film. That being said, if the presets are way off, which is a possibility (having shot "real" Ektar and Gold, the lab scans didn't look much like the presets), I'll probably try some other presets first, but then, if that also fails, I'll go to film. I have a feeling that this is gonna be the case in the end, that I won't be satisfied with the presets, but I have to make sure first

 

[wiltw]

Thank you for your answers

@wiltw : I'm not sure I understand what you mean... I had planned on using my EOS R for "scanning", and then using a plugin in LR to convert the negs. Does that "linear response" still apply to that process ?

The sensitometric curve for film is an S-shaped curve, but digital is Linear in its response. RAW processing is often set to a Linear curve by default, like this...

...whereas a more S-curve like shape (which is more similar to film emulsion) could look like this setting in Lightroom

 

[Yaeli]

@wiltw : Ooooh, THAT S curve Yeah, of course, I had planned on tweaking it Sorry, I didn't know you were refering to the tone curve (stupid French guy here )