The Problem of Mega Pixels

Hidden man in walkway

I love the capabilities of modern digital cameras, especially the wonderful sensors and great lenses available. But nothing is free, and I’m not just talking about the price of the gear. Having too many mega pixels can cause problems you may not anticipate.

Resolution is wonderful

I love extreme resolution. I’m not a fanatic about it, but I really appreciate it. I have not gone to 100+ MPixel sensors yet and I don’t normally do very large panoramas. Still, I get a thrill when I zoom in to 1-to-1 and see the great detail that is there. Then when I sharpen or contrast it more and the detail pops – wow!

Having large resolution allows me to create large prints. It is a necessary thing since I do this for a living. It is also something I really like to do. I don’t think an image is complete until it is printed. For me, a print is the physical expression of the image.

All things being equal, which they seldom are, higher resolution usually leads to sharper images. I love certain images to be “crunchy” sharp with great detail. It is part of my values that I can’t get away from.

Also, larger files allow for more cropping freedom. I try not to rely on this. It is much better to compose the image the way I want it at capture time. But sometimes it cannot be avoided. Maybe the image works better in a square format, or maybe I’m only carrying a lens that goes to 70mm and I want to shoot something I can’t get close enough to. In that case I have to “zoom” in post processing by cropping the image.

Or maybe I realize later that the real interest is in a smaller part of the frame. I have to crop the image heavily to salvage it. It’s not good practice, but I admit to doing it on occasion.

For me, a great print from a well executed, high resolution file is a joy.

Resolution is a pain

On the other hand, high resolution can be a pain. It increases the cost and time of all the downstream stages.

Every time I press the shutter it drops around 60 MBytes on my memory card. That is just the raw capture. It requires CFExpress or XQD cards to keep up. They are very expensive.

As long as I can process the image in Lightroom the size stays around this, but when I step into Photoshop each image balloons to several hundred mega bytes. And that is even without adding a bunch of layers.

Did you know that a Photoshop psd file (the native Photoshop format) cannot exceed 2 GBytes? Or that a tiff file cannot exceed 4 GBytes? I have found this out the hard way. Some of my images now have to be stored as psb files, the large file format version of Photoshop’s data.

Processing and editing time goes up with pixels. I use a powerful computer with 64G RAM and very fast Thunderbolt3 disks, but it can take seconds to do a simple stroke when I am masking or burning or dodging. I have seen multi GByte files containing one or more embedded smart objects take 2 minutes just to save to disk.

And you have to get to know disks in multiples of Terabytes. If you have a disciplined backup strategy, something I am fanatical about, then there are layers and layers of them.

I have bought in to the need of powerful and expensive equipment for editing and storing my images. The biggest problem, though, is the slow editing speed. This interrupts the flow of my mental process. I don’t like waiting on the computer.

Technique

One of the unfortunate truths they seldom tell you when you are looking at a shiny new high resolution camera is that it is harder to take good pictures with it. This is partially because of the geometries you are dealing with.

A full frame sensor, by convention, is 36 x 24 mm. My Nikon Z7 places 8256 x 5504 pixels in this space. That makes each pixel site 0.004 mm square. That is 4 microns from the center of one pixel to the center of the next. If you do not work in the world of integrated circuits or advanced physics you may have trouble conceiving these sizes. We do not directly encounter these dimensions in the real world.

As an example, human hair ranges from 17 to 180 microns in diameter. Therefore the thinnest strand of hair you can possibly find would cover over 4 of these pixels. An average sized hair, around 50 microns in diameter, would cover a strip of at least 12 pixels wide across the sensor.

A fun fact, but so what? The so what is that with each pixel being so small the problems of focusing and holding the camera steady are greatly compounded. Focus is critical and you almost have to rely on the very sophisticated focus system in your camera. Especially if it is contrast detection – meaning that it is searching for the best contrast, hence sharpest focus, measured directly on the sensor pixels.

And for the sharpest results, don’t even think of taking a picture without using a good tripod. I don’t know how steady you think you can hold something, but consider that for optimum sharpness the camera cannot move or shake as much as 0.004 mm while the shutter is open. I can’t do that, especially after coffee.

You need new lenses

Another sad truth is that to realize the full benefit of your high resolution sensor you need lenses designed to match it. Current lenses achieve resolutions significantly better than was the norm a few years ago.

The requirements for lenses for these new sensors greatly exceed the standard required for film or, say 6 – 10 MPixel cameras from just a few years ago. I have tried older lenses on my Z7. The results might be usable for some things, but nowhere up to the quality of something like a Z 24-70 f/2.8 designed specifically for the Z series.

So another cost and problem of trying to achieve very high resolution is that you need to use lenses that will achieve the quality you are seeking.

Why have lots of Mega Pixels?

With all those problems, why should you want to shoot high pixel images? Maybe you don’t. That is what I am leading to here.

Your gear should be chosen based on your intended use. These days many people will only post images on social media or put together a slide show of a trip or event. If they print at all it will probably be 8.5×11 inches (about A4 for you in the rest of the world). Quite honestly, a good 6 MPixel camera is all you would need for any of these things. Almost any mobile phone is great, except for the lack of lens choices.

I have images from a 6 MPixel camera in my portfolio.They are good files and the quality of the pixels is good. I just would not try to print them very large.

About the only thing that requires huge files is making large prints. This is a world I live in, but if you don’t then why bring these other problems on yourself? A good 12-16 MPixel camera is probably more than adequate for most people. They are smaller and lighter and cheaper. It is easier to take good pictures with them, it is easier to process them if you want to, and they require far less disk space. You can probably keep most of the images you want in online storage.

But human nature being what it is, we can’t discount the lust factor. Pixel lust. Just like I know people who do some woodworking and have a workshop outfitted with an array of near commercial quality equipment. An expensive overkill, but if they have the space and money to burn, why not? You might need it someday.

If you want to be logical and save some money and time, resist the lust for lots of mega pixels. You won’t need them.

Its an OK problem to have

Some of us are convinced we need them. Some of us just want the biggest and best. Many are just caught up in the hype of shiny new products.

If you are going to have a high mega pixel camera, be aware going in of the costs and problems. But if you “need” it, go for it! The results are marvelous if you use the tools well.

I love the results I get so much that I forget about the size and processing problems. I love the results so much that I gladly learn the required techniques to achieve them. They make all of my images better.

Cameras and gear have advanced to the point where many of us cannot achieve the maximum they are capable of. But that is an astounding problem to have. What an embarrassment of riches! If we are the weak link in the process, we can learn and improve. We get better and our results get better.

It’s a great time to be a photographer.

What have your experiences been with high resolution photography? Let me know!

What is DPI?

An extensively processed image

DPI is simply an acronym for “dots per inch”. It should be a simple concept, but people sometimes get twisted up in knots over it. Our printer manufacturers have not helped the situation.

What are “dots”?

Unless you are reading a printer spec sheet, dots just mean pixels. A pixel is a “picture element” – the smallest piece of a digital image. By convention, a pixel is a triplet of red, green and blue values. That sounds very technical, but it just means they are 3 values carried around together, say something like 95, 134, 47. By convention each value can have a range of 0 to 255. That is not representative of what camera sensors really do anymore, it is the convention. The convention comes from the 8 bit representation of color values way back in early times. The maximum value 8 bits of binary data can represent is 255. The practice has been established and perpetuated by Photoshop over the years. The actual data range we use is a subject for another post.

The image that gets stored in your computer is a grid of pixels. The camera I am using most often right new creates an image that is 8256 x 5504 pixels. That makes a lot of data!

A fuzzy quality metric

Your camera does not know the concept of DPI. As a matter of fact, DPI is a fairly useless term unless you are printing an image.

Have you ever had someone tell you they need a file that is 8×10 inches at 300 DPI? Unless you are sending the file to whoever does your photo printing, it really means they do not understand what they need.

At best DPI is a metric for the quality of an image viewed or printed at a certain size. The more pixels you have in a given distance, the better the image should look, in general.

But in most people do not even know the pixel resolution of their screen or printer. If you save a file in jpg format there is a good chance the default is 72 DPI. This was considered the “normal” screen resolution – way back. The main monitor I use is 219 DPI. And it is several years old. And I don’t really care, because that is a number that is never important to me. I never use it for anything.

DPI really doesn’t matter for the screen

One of the reasons DPI doesn’t mean much for most of us is that our computers scale images for viewing on our screen. And they usually do a really good job. But what most apps do is map the pixels available to the pixels of the screen. So if you look at a file that says it is 10″x10″ at 72 DPI, you will see a 720×720 pixel image. It will be however large 720 pixels measures on your screen.

When you view an image on your screen all that really matters is the resolution. The DPI number is generally ignored.

Size matters

If DPI is not as important as many people think, then is resolution unimportant? No. Absolutely not. The number of pixels you have to work with is always an ultimate limitation of what you can do with the file. As is said in many things, size matters.

With plenty of pixels imaged through good glass you have the flexibility to print large, or to crop tightly or to create images of astounding detail. Also, massive numbers of pixels gives sharper edges and smoother gradients.

DPI for printers is a whole different thing

The major printer manufacturers have confused the issue for us. A printer ad may proudly proclaim it does 4800 DPI! This is technically correct, but not helpful. They are talking about the density of ink drops they can lay down on the substrate (paper). But printers do not print pixels.

A drop of print ink is not a pixel. You do not send the printer an image scaled to 4800 DPI!

A printer takes the pixels available in a given area and transforms them to densities of the subtractive colors needed to come close to reproducing the colors and gradations contained in the original pixels. This is a complex technology and I will not attempt to explain it here. Sufficient that you remember a drop of ink is not a pixel.

Where DPI means something

When you know how large of a print you wish to make then DPI becomes meaningful. The number of pixels available combined with the desired print size give us the DPI. DPI is a measure of the amount of information available for an inch of print.

For optimum printing the guideline is supplying a source file of around 250-360 DPI. This gives the printer driver enough information to do the ink transform we talked about above.

My Canon printer, for instance, has a “standard” resolution of 300 DPI. It can print well with a range of values, but this is considered optimal. This means that one of my image files of 8256 x 5504 pixels could be used to print an image of 27.5″ x 18.35″ with no scaling or loss of resolution. That is the size of this image at 300 DPI.

If I want to print one of my images at a more typical size, say 18″x12″ I could scale a copy of my master file down to 300 DPI. Or not, because Photoshop or Lightroom is perfectly capable of scaling it down when printing with little discernible degradation. One person even says that higher DPI gives better results.

If I have an image with insufficient pixels I could just try to print it. Printer drivers do amazing things. Or I could scale it up in Photoshop, which also does an excellent job within limits.

It’s those limits that you have to be able to estimate. If you have a 2 MPixel image and you want to print it poster size, well, your results probably will not meet your expectation. No free lunch.

Conclusion

So what about DPI? Don’t sweat it unless you are printing. Only if printing is it a meaningful metric. And it is only meaningful when you are taking about a particular image printed at a particular size.

There are plenty of technical issues to stress out about. This should not be one of them, unless you are producing high quality prints.