Why the hell do we need 400 megapixels? We don’t even need 100 megapixels, and I’ll prove it
posted Tuesday, January 23, 2018 at 2:10 PM EDT
Last week, Hasselblad announced a new version of their medium format camera that is capable of producing 400 megapixel images. Generally the first reaction to this kind of thing is of course, “oh wow, that’s wild!” But for me, it was a resounding, “Why?” Off and on over the course of digital photography, the industry has engaged in what many affectionally call the “resolution wars” or “megapixel wars.” Any increased emphasis on resolution in marketing materials from one manufacturer is quickly met with the same from rivals, and the megapixel rating becomes the main selling point of everyone's cameras. Never mind that these cameras likely have many other more impressive features to those paying attention. The megapixels always win out in marketing materials.
This is likely because the average consumer has no idea how or why sensitivity matters, the difference between on-sensor stabilization and digital stabilization, the tech behind autofocus engines, or how any number of other features that a camera has work the way they do. They just take those for granted or ignore them in favor of something they do understand: Photo size. Megapixels.
But if you’re reading this, you’re better than that. You actually do know the differences, and you shouldn’t be swayed by megapixels. But many of you are.
To explain my thinking behind this, we need to do a few things. First, let me talk about one of my favorite photos. I shot this photo out of a helicopter with a Phase One XF body and 100 megapixel sensor, so it has massive resolution. It occurred to me recently that I see this photo all the time, but it’s not where you might think.
It’s my cell phone background.
I took a $50,000 camera into a helicopter, captured a photo I consider to be one of my best, and the main use it gets is as an iPhone background. And you know what? This cannot be a story unique to me. I’m sure many of you have high resolution photos you’ve taken that also live digitally, and the resolution of the image plays little to no role in your enjoyment of the image.
So once I started thinking about resolution, I decided to test it: What would I need this kind of resolution for? I looked around at reputable printing houses and decided I would go with the largest print I could purchase from one of them. I chose WhiteWall because I remembered how awesome a photo that was printed in their “ultraHD” tech looked on the wall at WPPI two years ago, and decided to give that a shot.
After speaking with their customer service, I found that I could order a custom size from them that wasn’t listed on their options page: 70” x 48”. That is the largest size they can produce, so I figured that would be the best place to go to decide how important megapixels are and what they could get me.
While I waited for my order, I asked WhiteWall what I considered to be the most important question. Clearly I could order this print with my 100 megapixel image that calculated to a pixel width and height of 11,608 x 8,708, but what would be required by WhiteWall as a minimum?
They told me that given the assumption of a maximum resolution that makes sense of 300 dpi, calculating 300*48 x 300*72 you would come out to a maximum resolution of 14,000 x 21,600 pixels, far more than I provided.
Now you’re probably thinking, well that settles it! With only 100 megapixels, you weren’t even close to the resolution required to get the best quality out of the image you ordered!
Well, no. Not exactly. For starters, this is a gigantic resolution compared to what most, if not all, cameras can do. All higher dpi resolutions will be scaled down in WhiteWall’s production workflow.
Also, you have to consider the viewing distance. The larger a print is, the larger the typical viewing distance would be. You have to step back to see the whole picture. Here, we assume a distance of 2 to 2.5 times the image size (measured diagonally). Given the limitations of the human eye, a lower resolution of the printed image is fine, the larger the distance is.
To know exactly what kind of viewing distance I would get out of 100 megapixels to this size, I would have to wait to see myself.
Finally, WhiteWall also said that for a 70x48 inch print, they recommend at least 3,000 x 2,000 pixels. Given that they would be willing to print at something a quarter the size of what I delivered them, you can see how just knowing the maximum resolution they would accept isn’t nearly the whole story. With 3,000 x 2,000 pixels, WhiteWall said that would result in the minimum quality that they would accept, although it is not perfect and if you step closer to the print it would not really take advantage of ultraHD. For a good high quality ultraHD print, WhiteWall would recommend at least 4,600 x 3,066 pixels, so that’s what can be considered as the minimum dimensions in this particular case. Again, much smaller than what I provided.
Now, obviously, the more pixels you can provide a printer, the better the prints will be. But that “better” is an exponential curve. Eventually, your eye isn’t going to be able to see a difference.
Again while waiting for the print’s arrival, we at Imaging Resource did some thinking: Aside from some very, very specific use cases, what would 400 megapixels actually be good for? Beyond scientific or perhaps archival photography (photographing art and cultural heritage objects, for instance), we couldn’t really think of any. Roger Clark calculates the human eye's resolution as equivalent to 576 megapixels. But that would be for the entire visual field, and in practice, a photo is only going to occupy a small percentage of that area. Even a huge photo would only occupy a segment of your total vision.
Another argument for more pixels is that it would give the photographer the option to "crop in" and focus the subject. Well, yes, but you have to consider the circumstances of using a camera like this. If you're shooting with this Hasselblad, it is extremely likely that the client has requested it. And if this is the case, you're working with a budget with a client who is providing a brief. You would be shooting specifically for that brief. If you found the need to crop in at any point, odds are that 1) you messed up and need to hide the mistake (bad) or 2) the client provided a poor brief and should be paying for a reshoot. In either case, cropping in would be avoided on set by using the right optic, and if you do crop in you're likely changing quite a bit of the original flow of the image. Good shooters who are photographing for a brief get as much right in camera. Cropping later changes perspective, and isn't ideal.
From our own experience in testing print quality from many cameras, 20 megapixels is more than enough to make a 30x40 inch print that looks great at normal viewing distances of a foot or two. Even a billboard doesn't need much more resolution, unless you're going to walk right up to it- which you would not do. It's all a matter of how large the printed image is in your field of view; it's rare to view an image in a way that it takes up more than 30% or so of your field of view horizontally and vertically.
But what if one did? Would that then make a difference? That was what I wanted to test with this giant print.
After hanging the print and lighting it, I asked eight friends to come over and take a look at it. Every single one of them was first in awe of the size, and then of the quality. It looks really, really good. Here is the kicker: It looks really, really good when your nose is basically pressed up against it.
That’s right, you can enjoy this image from as far away as my back office wall, or as close as you can physically get to it.
And that’s kind of what I was expecting.
A few weeks ago I was lamenting that I couldn’t think of a reason to use 100 megapixels, and now this week that has extended to 400 megapixels. Today, with modern printing technology from arguably one of the best photo printers on the planet, 100 megapixels is more than enough to enjoy a six foot by four foot ultraHD print from centimeters away. This is a totally fringe use case, as actually viewing photos from that close doesn’t make any sense, for one thing, but it’s also extremely rare to encounter an image you would even consider doing this to. For what it’s worth, as a frequent museum-goer and art exhibition attendee, I’ve seen hundreds of well-printed photos up close. This photo I created with WhiteWall is by far the largest that can be viewed from so close that I’ve ever encountered.
And again, you can not perceive a loss of quality from any close distance to this image.
A friend of mine called me right after Hasselblad introduced their 400 megapixel camera and wanted to talk about it. I said that I thought it was pointless, and that it unnecessarily shifts the conversation away from what should be important to imaging. Yes, there are absolutely use cases for this camera as mentioned, but you’ll find folks gushing about it who would never, ever use it for those very specific use cases. They’re looking at 400 megapixels from the same frame of mind that they use when seeing 42 megapixels. And that, to me, is wrong.
My friend then asked if companies should stop innovating, since I clearly was unimpressed with this achievement. I responded that he was missing the point I am making.
I love this print. I look at it every day. But I likely could be enjoying this image had I taken it with the new Nikon D850, or the Canon 5DS R. Or a Sony A7R III. Or any number of awesome cameras that exist on the market today. The number of pixels I provided the printer with vastly exceeded the minimum needed for me to enjoy this image. Why the hell do we need 400 megapixels, when I didn’t even need 100 in order to create the largest print I’ve ever seen and enjoy it with my nose basically pressed up against it? The thing is, we don’t.
I told my friend that camera companies should be focused on improving the quality of the pixels, not the number of them. Improving other aspects of camera design such as high ISO quality, body weight and feel, stabilization, and dynamic range are all, to me, more important than the number of pixels that can be produced from a camera. There is still quite a bit to be innovated towards, and it doesn’t have to involve megapixels.
Let’s focus on making our cameras better, not bigger.
Thank you to WhiteWall for providing the print for this experiment.