How Much Dynamic Range Do You (Really) Need?

At first it was megapixels. 4 were better than 2 (right?), 10 were better than 6 (right?), and 24 should be better than 12 (right?). I won't talk about megapixels today (feel free to read this, however), nor about ISO and noise - the next thing people began fretting about. Today I will talk about Dynamic Range which, together with something people refer to as 'Color Depth', is the most recent number people fret about.

To an extent, it's human nature. People need simple answers to complex questions (to digress and offer a sociopolitical comment, think why governments need to convince their citizens about the presence of one, specific "bad guy" group that needs to be dealt with, or why it is easy to blame immigrants or ethnic minorities about unemployment, or other issues).

In terms of photography, people need specific, simple answers (in other words: a single number) that lets them evaluate whether camera A is better than camera B. And so we have Dynamic Range measurements, or Color Depth measurements.

But let's take a breath here, shall we?



I won't go into the technical aspects of these things (and hence perpetuate the number infatuation), but I will ask you to do something very simple. Something that we tend to forget to a puzzling extent, considering that photography is a purely visual medium. What I will ask you is this: trust your eyes.

Trust your eyes and examine the following two images (click to enlarge if you want):

In fact. if you follow AmateurNikon on Twitter, Facebook, or Google+ - and you should! ;) - you might remember that some time ago I asked you to take a survey. The survey displayed the two images above, together with the question "Which one do you prefer?". The possible answers were top, bottom, I don't see any difference/I don't know.

The results were as follows:
~35% responded that they preferred the top image.
~30% responded that they preferred the bottom image.
~35% responded that they either don't see any difference or that they don't care either way.

So, what's going on here?
The top image has 235610 unique colors; the bottom image (that has undergone deterioration in Photoshop) has 202221. That's a whooping 15% decrease. To put things in perspective when we talk about Dynamic Range or Color Depth measurements, the difference in percentage you see above would correspond to the difference between 14.4 EV and 12.2 EV and 25.1 bits depth and 21.3 bits depth, respectively.
Do you want a point of reference?
14.4 EV and 25.1 bits of color depth are the numbers given for the Nikon D610 by DxOMark.
13.7 EV and 24.2 bits of color depth are the numbers given for the Nikon D7100 by DxOMark.

The survey results suggest that with even such an outrageous difference in numbers, a viewer would either not notice the difference or will not automatically choose the "better number" as the better image. In addition, as most of the readers of this website are photographically savvy, they don't even represent the average Joe/Jill who simply views an image and doesn't care about colors or dynamic range.

Let's now paint some more devils on the wall, shall we?

How about this image below, do you like it, is it a "better" image than those other two (again, click to enlarge)?

How many unique colors do you think exist in this image?
Take a guess...
Come on, take a guess...
No?
Well, let me remind you that the images above had about 235k and 202k respectively.
This one?
24k. As in, 24000 (and some hundreds). If you don't believe me, run it yourself through some color counter. In fact, you know what? This number isn't even telling you the whole truth. You see, in reality, what you see above has only 255 colors. I kid you not. What I did was this: I tone-mapped the original image, then reduced the color numbers to 256. To save it as a jpeg I transformed it into RGB 8-bit space, which is basically responsible for the 24k+ colors in the count (they're just roundings and approximations created by Photoshop algorithms, not original information; the 255-color .gif version looks exactly similar to the 24k .jpeg version).

So, this last image has 99.991% (even if you wanna go by the 24k figure, it's still 90%) less unique colors compared to the previous two. Think about that for a moment.

I can imagine some objections already.
a) "This isn't the same thing as having a sensor that captures more dynamic range"
b) "This is a special case, in some landscape/portrait/ambient light/other situation, it would be different"
c) "This is an image post-processed in a specific way"
d) "There are monitor/printing/jpeg compression factors"
e) [....] send a message or add a comment below with yours ;)

Guess what. What all of the objections above (or similar ones) mean is this: That we, photographers, are all different. We have different needs, different subjects we (usually) want to photograph, different procedures we follow. Different outputs, different audiences. Different scopes. A camera (and its associated lenses, flashes, software, etc.) are only a medium; a tool. The end goal is always a photograph and never the tool itself. Can you imagine Michelangelo, Botticelli, or Picasso spending agonizing moments in their workshops regarding whether they had the latest and greatest brushes? Can you imagine them (if we suppose they were contemporaries) to argue whether a brush had this more many hairs per square inch, or it was elastic by this more compared to another brush?

There's a photo taken with an ancient D40 (and the seriously consumer 55-200) , which is nowhere near the D7100 - not to mention a D810 - in terms of dynamic range, color depth, or whatever number you wanna compare it by. So what? This image won me 2nd place in a local competition and was published on several media. 

In some photography forum, someone was asking some time ago "Is the Nikon D5300 enough for an amateur?". Someone replied, perhaps a bit too directly but certainly rightly: "The most awesome photos in the history of mankind were taken with lesser equipment than this. Who do you think you are?"
I see people fighting and agonizing whether the D7100 (or the upcoming D7200) is good enough for them. Others wonder whether the D610 is good enough for them. Some even say that, heck, they know almost nothing about photography, but they'd rather buy a D810 so that they know it won't be the equipment that fails them.

If you recognize yourself in these patterns, don't let me stop you. If you absolutely must make Nikon richer, go ahead and buy the most expensive camera you can get your hands on. You will have a camera with 25.7 bits of color depth. Great, congratulations. Pat yourself on the back, good thing you didn't get that other camera, that had 24.2 bits of color depth.
Now, show me your portfolio.