Some Thoughts on the D800 vs. Film
Some people have been saying "the new Nikon D800 is better than film". The other day, a reader asked me again if I thought this was true.
Looks like it's time for another chapter of Film vs. Digital, as we roll on into 2014.
When someone says a digital camera has "surpassed film", I would ask, "In what way?" Are they talking about resolution? Dynamic range? Acutance?
There's no question the D800 has high resolution. It offers a full-frame (FX) sensor with some 36 megapixels. While 10 or 12 MP is plenty for what most consumers are doing, some things do require more. I can think of a couple situations where the extra resolution really comes in handy. First, it really helps for jumbo enlargements. Second, it's important for macro capture of film. Scanning has always been a resolution bottleneck, and this is especially true for larger pieces of film. If a 120 frame has anywhere from 50 to 100 megapixel equivalent, then a 12-megapixel sensor is going to lose a lot of detail if you use it to digitize that film. In fact, when I've used a 12MP camera for large format (4x5) transfers, the detail loss was extremely obvious. Of course, you could argue that reality has more megapixels than large format, so maybe more megapixels ain't so bad overall... a whole 'nother debate.
Anyway, here's the current issue. The D800 with its 36 megapixels has been said to surpass the resolution of a 35mm negative.
I won't disagree with that, even though a few people have said the better 35mm film stocks can approach that number. If you had a film that could do 100 line pairs / mm, that comes out to almost 35MP. I believe Velvia 50 can do 80 or 81 lp/mm, though I've read figures that place it as high as 100 lp/mm. If you do the math for 35mm Velvia 50, you get somewhere just over 23 MP, up to as high as about 35 MP.
It seems the widely-accepted "best case" figure for 35mm film is 24 to 25 megapixels. So, for the sake of this article, let's say that the D800's resolution surpasses that of a 35mm film frame.
The problem with trying to peg film vs. digital to any particular measure is that it's more complex than that. Though you can start to resolve the film grain with a good scan, many people do not realize that there are smaller and smaller levels of "grain". Actually, not even levels. It's a continuum. Dye clouds and silver halide clusters go all the way down to the nanoparticle level. There is no quantized unit of size for these. Now, even though you've already resolved nearly all the detail by the time you can see the biggest grains, the analog nature of grain could be influencing the picture in ways that go beyond detail resolution. And even that's only part of it.
I've written articles about the dynamic range, and that's a huge component. (The issue is actually dynamic highlight range, or DHR.) Maybe it's not even the only thing that sets the two media apart. In fact, I know it isn't. (By the way, another reader emailed me and asked whether Camera vs. Log was shot using RAW on the D7000. He wanted to know if perhaps there'd be more highlight range that way. Well, don't you worry none, because for better or worse we're going to do Camera vs. Log Part II... soon as the weather gets nicer).
The D800 is widely said to have better highlight range than any other DSLR on the market. I'm even hearing people say its DR is "better than film". Actually, no. The resolution, yes; but not the range. If you want the proof, here it is. Just as you've seen here at 120studio.com, this photographer did the kind of test you'll probably never see on the "benchmark" sites. And guess what he found? The Fuji S5 has much better highlight range than the D800, as I expected. (You know the guy knows what's good... he shoots large format, too.) I've probably said before, at least a couple times, that the S5 was the best thing that digital ever threw at the dynamic range problem, but even that probably wouldn't have done any better than a little roll of Fuji Superia that expired back in '07. The new Kodak Portra has even more range than Superia.
I assure you my purpose is not to offend the many photographers who enjoy digital cameras. (Most of you know that I enjoy digital sometimes too, and this camera is a little gem.) Here's my philosophy on this. Right now we're talking about a category of DSLR's that can easily cost $2,000 and up. Before a person drops that kind of money on something, they should at least know its limitations. Unfortunately there has been a lot of technical jargon that tries to cover for the glaring fact that digital cameras still don't do so well with highlights. I've often thought the silicon photodiode would run into a wall sooner or later (although the Fuji S5 shows us there is a lot of headroom that manufacturers just aren't using). Now, Fuji and Panasonic are developing an organic CMOS sensor that sounds very promising, but we'll talk about the implications in another article (no, it's not the "end of film", but I wouldn't be surprised if the media said that.)
We can know the D800 does not match film in terms of highlight range, because the D800 can't even come close to the Fuji S5. In turn, the mighty S5 was still not an equal to color negative film. Right now the king of dynamic range is probably Kodak Portra film, and thank God that we can still buy it. That is a blessing. Contrary to what some were expecting, Kodak didn't toss their film division in the wastebasket, and the chemists & engineers who worked at the plant never took that rocket to Saturn that I was talking about. Someone realized that Kodak film is indeed awesome. It seems they were too smart to be taken in by the "film is dead" hype. And so we have Kodak Alaris, named because it's more agile than the original Kodak. They're not sitting on their hands, either. As my readers keep showing, film has a growing user base.
Now at this point my D800-owning friends might be thinking I've just completely dissed their favorite camera. Far from it. I'm just being realistic: it's a tool with its own set of pros and cons. And yes, I would use a D800 gladly. Might even do a review of it on here one of these days.
Are there any other important differences between a modern DSLR and a piece of film?
Let's find out.
A Sophisticated Computer
We know the D800 still gets trounced in highlight range. As I said, it's not really that much better than a D5100. (Great camera outside of that particular measure, by the way.) So, okay, what else is there to compare?
To me, the Nikon D800 shots have more of that same "Nikon DSLR" look that we've all come to know and love. Especially in landscape shots, I can see it immediately. They're like more-detailed versions of what you'll see from a Nikon D5100 or D7000. The pictures have that same "look" to them, in much the same way that Canon T3 shots have something in common with those from a Canon 5D MkIII. Be glad they do this. Diehard Nikon or Canon DSLR users expect a certain color rendition, a certain familiarity, that carries over into new models. When you have a popular formula, why change it? (In case you don't remember 1985... yep, the "New Coke" was pretty gross. At least I thought so.)
Canon and Nikon have good "recipes", but they're not film. Even if they cranked 'em up to 100 megapixels, it wouldn't matter. It's not megapixels. Appearance-wise, a 110 film shot has more in common with a 35mm film shot than it does with something from a point-n-shoot digicam. So we know it's not resolution. It's not even dynamic range alone. That's part of it... but it really seems like there is something more going on here. Digital cameras yield pictures that look more detailed than ever, but at the same time they look synthetic. Not every time, but often enough that it's noticeable. Even without that jaggy, pixelated look that we saw in the early days, digital photos somehow manage to look computer-generated. I don't mean crudely computer-generated. Not at all. They look like the product of a very sophisticated computer, because they are.
Acutance (edge contrast) may be part of it, but I don't think that fully explains the synthetic look. By the way, you might remember that many people were thinking digital "surpassed" film resolution way back at 8 or 10 megapixels. I think one reason is that a lot of us were using scanners that couldn't even approach the potential of the film. The D800 is at about the level of detail where your best possible 35mm scan would be. It says a lot that it took digital this long to get there, but again, that's just one variable out of several.
Supposing we do eventually identify all the reasons that make them different, there's something else to consider. Film photographers use a rather simple mechanical device to make pictures. The Nikon D800 produces files on a memory card, by way of incredibly complex microelectronics. Neither method is invalid, but I know which one I prefer personally. (And yeah, I still like that you can take nice pictures with cameras that have so few parts.)
It really comes down to the fact that a Nikon D800, like its predecessors, is not Ektar, Portra, Provia, Superia... or Velvia. Even the Fuji digital cameras aren't Velvia, despite the fact that they offer "Velvia" modes. Even if you could make a 99.9% approximation, it would still be a digital facsimile. Some people wouldn't care, others would. Right now we're not even close to 99.9% approximation, though.
Put some slides on a light pad (pick up an 8x10 size here, or check out my review of this handy "mini" one). Or, even better, view the slides through a projector (help out my website buy purchasing your vintage one through this link). There's no substitute for film. People are emailing me a lot now, observing this same thing for themselves. Many of them have picked up film for the first time. Many others have returned to it after a long time using digital. Either way, they like what they're seeing. And obviously, no one says anyone has to throw away their digital camera.
Now, the very astute observers will note something about digitized (scanned) film. The fact that we can digitally acquire film-- and it mostly retains the "film look" on the computer screen-- means that theoretically, the elusive "something" about film can be quantified, translated into the JPEG / TIFF / RAW data spaces. In other words, a piece of computer software can at least encompass that data set without changing it substantially. That means in theory that a computer could emulate the film look.
Here's why that's probably not going to happen any time soon. Yes, the computer is able to import a complex data set that already exists. Today's DSLR does not have the means to generate that data set "de novo". The differences happen at the film vs. sensor level. The DSLR generates a close approximation, but we don't even fully know what's missing. The differences may be too numerous to nail exactly. It seems every time I turn around, I find yet another difference that I'd never thought about.
Let's look at a comparison to illustrate what I mean. Here's a case that eliminates a few variables. I'm not even going to include the dynamic range element here. We're using nearly monochrome light. It's not a bright light, either. (Here, I turned up the contrast to approximate that of the Velvia. Slide film has a reputation for being contrasty, but low-light monochrome makes it even more so.) Now, mind you, I did everything I could to get the digital shot to look like the film shot. And it's pretty close... but yet, it's not.
First we'll look at the digital picture. This was shot with a 12-MP DSLR. The point here is not resolution but overall appearance.
So, here's the digital picture.
Now, the Velvia 100 35mm film:
Even though this is a digitized (scanned) version of slide film, and even though we're dealing with a sub-800-pixel JPG, it still looks different from the DSLR shot. The differences are subtle, but they're there. I did my best to make the digital look like film, but something is still not the same. (The Velvia was scanned with an old Epson 1640, because my V500 died. When I get the chance, I promise I'll do a macro-cap and upload bigger files to compare.)
At this stage, someone may say "So it looks a little different. So what?" That's the point, though. The DSLR doesn't provide quite the look I want. (No mounted transparencies, either.)
A couple things here:
1. Highlights, shadows, and tones render a bit differently between the film and the digital. This cannot be emulated adequately with a software filter.
2. The green is different. This is one of the colors that Velvia does well but digital renders "not quite right". It just so happens the human eye is very sensitive to green, and we humans can usually discern different shades of green very well. So, if you get green wrong in a picture, well... it's going to affect the impression given by that picture. And there's a lot of green stuff in nature.
Now, here's another interesting point.
3. If you look carefully, notice that the hue varies at different places in the DSLR shot. This was basically a monochrome light (green LED). That's a relatively pure green. There is one sharp peak around 565 or 570 nanometers. There is no reason I can think of why part of the picture should look more yellow-green than another! There's also no reason why a monochrome LED would produce white, but the DSLR rendered some spots as white. That's a problem.
The Velvia registered it as pretty much all one color (as it should), but if you look carefully, the DSLR is showing a lot more hue variation. I even checked in my software. That said, here's the important part. It doesn't matter whether one is "technically superior" to the other. It doesn't even matter whether one is more "accurate" than another. What matters is whether you like it. Here again I prefer the way the film handles the color. Slide film is just great stuff.
All this digitized film looks pretty good, but even that's not as beautiful as the slides in real life. The color is better, they're more vibrant... everything. That's because even though a computer scan can encompass "most" of the data, it still loses something. And let's not forget that slides and negatives have that process, that tangibility, that archivability. If every other advantage went away, that would still be an enormous one.
Just a little digression here. We're moving into a new era of technology which, despite its various benefits, has at least one major disadvantage: intangibility. This leads to an impermanence, a rootlessness, that can impact our society in more ways than one. (Probably in ways we don't even recognize yet). Like many readers, I remember a time not too long ago when our music and our photographs were all stored on analog media. Then it started to go digital. Now, we're moving into a time when some people are not even storing their data files on their own hard drives. There may come a day when a Nikon D800 will have more in common with my 1980's film camera than it will with what people are carrying around to take pictures then. Big business will gladly sell you a gadget that is cheaper to make and can't even do half of what you want. This modern-day regression toward "mainframe computing" is their dream come true. Then, they can give you even less... and make you pay a subscription fee for it. Some of these guys want everyone hooked on dumb terminals and paying a monthly fee.
If you let these bean-counter types have their wish, your camera will be a dumb terminal, too.
I remember hearing people start out with 640x480 camera phones, and already they were saying "Who would ever need film?". This kind of thinking then prompted industry executives to say "Yeah, you know what? Why do we still need to make transparency film, when the person can just make some digital snaps on his camera phone?"
"Yeah, I mean, like... apples and oranges are both fruits, so why can't everyone just make pies with oranges?"
Well, possibly because they're not apples.
(Slide film. No bounce, no strobist setup, just TTL flash.)
Once again, photography is about art. Well, it could be about delicious food, but either way, it's all about the subtleties. Ever heard of subliminal advertising? We can be influenced by data that we don't even recognize. Again, subtleties are everything. You can pick up a one-megapixel digicam and be an artist, but your art will have a different look from that of an artist who uses, say, a 110 film camera, or black & white 35mm, or something else. It goes back to that analogy I've used about various paints and paintbrushes. The different choices of film and digital cameras could be like the choice between acrylic, watercolor, oils, tempera, etc, etc... and with different brush sizes to boot.
(Quick question... which one do you think would be the oil paints?)
You know, I just thought of something obvious. Digital has been trying to match the appearance of film since what, 1997? We're going into 2014 and it's still not there. The resolution is there (D800), but still something is missing. Not just highlight range, but something else, too. That should tell you something about the quality of film. Again, I'm not saying digital can never get there, but so what if it does? (It only took 'em the better part of two decades...) It makes me want to pick up some extra rolls of film and go on a photo walk, just because.
Too bad we can't include E100VS and Elitechrome here, but Kodak is still making great color negative films. (Links to Ektar 100 later on in this page). I'll eventually do a whole article on Kodak's current films, because I really enjoy them and I know a lot of other people do, too.
Live And Kickin'
We live in an interesting time for photography. Though it's not as frequent now, the "film is dead" script has been rolled out regularly for a good ten years or more. (Tired yet, fellas?) Up until not very long ago, the mainstream media still felt compelled to portray film photographers as "hold-outs". And yet, film is somehow still on the scene. Still relevant, still inspiring people... and connecting a whole new generation with the best of photography.
No matter what the future may bring in digital improvements, I will always think of film as a "forever" technology. It's right up there with the bicycle, the ballpoint pen, and the record player. Just because something new has come along, that doesn't automatically mean it has to replace these things. I have no use for a world of virtual flying cars where we eat digital food and don't exist.
Today a lot of digital photographers often shoot film just to mix it up a bit, or to enjoy its unique aesthetic, or maybe to archive some important shoot. However, it was only five or ten years ago that this sort of thing wasn't so much the case. There was, at times, almost an animosity toward anyone who liked film. Maybe not even "almost". It seemed every group had some guy who would start arguments to prove the superiority of his new $4,000 purchase. These guys didn't take a lot of pictures, because mostly they were too busy spoilin' for a fight. It was actually a lot like the tubes vs. transistors thing that happened in audio, years ago. There are differences in the sound we might not even know enough to measure, but they affect the overall quality. And so, after some years of bickering, eventually a lot of people realized that tube audio is a valid thing. Very much so. (Same thing with vinyl LP's, which by the way can also be digitized like film.)
For the most part, the film vs. digital debate is not so strident any more. Things seem to have improved. (Maybe because they have both been around long enough now.) There are important differences that should be preserved, meaning they both deserve to have a permanent place in the repertoire.
It's as though we've assimilated each other. I've really started to see this happen since maybe 2012, and it's a good thing.
If you get the occasional strange look while toting a film camera today, it's from people who aren't really that into photography that they'd know the real score. They just sort of assume "no one" uses film, because that's the basic idea that's been marketed to the consumer. Then again, the same idea has been pitched to pro photographers, too. I mean, what are the digital-only publications going to do? "Well, here's a review of the latest DSLR, but if you really want image quality, go shoot 4x5 film." Their sponsors might not like that too much. (If they make film too, like Fuji does, they could actually benefit from that approach... "Here, try this digital camera. You like that, right? Now try this..." )
(And of course Nikon still makes the awesome F6 film camera. They don't exactly need to be shy about it.)
The Nikon D800 equals or surpasses the "megapixel equivalent" of 35mm film, but I don't think that really changes anything. DSLR resolution has been compared favorably to 35mm for several years now. The D800 brings detail resolution to a point where there's no longer any doubt. No worries, though. I really mean that. First of all, we've seen that the highlight range problem is still an issue with the D800, even if it's just a tad better than what DSLR owners have become accustomed to. Second, the "film look" is a complex interplay of a bunch of different factors. Not just highlight range, but also color response and a few other things. Not only that, but film offers a whole process and a means of archivability that digital cannot replace. When you add up all the different sub-types of film, from slide film to Fuji Instax, there are so many different "looks" you can achieve with film that it's amazing.
Enjoy film, celebrate film, use it! And if you want a full-frame DSLR that is in fact quite impressive, apart from the usual dynamic range issue... by all means, pick up a Nikon D800. Obviously, the camera is quite useful. Get this lens for the D800, and you'll have an enviable rig for macro capture of film. At 36 megapixels, that's what I see the D800 being best for: film capture. (I bet Digital never thought it would be assimilated by Film.) Other people see the D800 as a primary working camera, and if that works for you, then so much the better!
We're really fortunate now that as we go into 2014, we have the choice of both kinds of photography. There are some great film emulsions on the market, and of course there are some highly worthwhile digital cameras, too (not the least of which is the D800). Even with all this, my thinking is still the same. I still want the look of film, the process of film, the tangibility of film... the whole thing.
That's "forever" technology. And yeah, digital can ride along too, but only if it behaves. (Speaking of which... the question is not whether a Nikon D800 is better than film. The question is whether Nikon is even as good as Canon... :-D )
I hope this article brought a smile to your face, or you found it helpful. As always, you can really help me out by using the links on this site to pick up your photo gear, or anything else. It helps me keep up this site.
Keep enjoying photography, and thanks for reading!
||4x5 (20 sheets)
||4x5 (10 sheets)
||4x5 (10 sheets)
(Portra 400 is also available in 220 rolls!!)