- Consumer Electronics & Personal Gadgets»
- Portable Electronics
How many pixels are enough?
Digital images are made up of pixels, dots of a single color.
A one pixel image looks like this.
A four pixel image looks like this.
Pixels do not have size.
They can be scaled up or down to print images at different sizes.
A megapixel is a million pixels.
To figure out the number of megapixels in an image, multiply he number of pixels in one dimension by the number of pixels in the other dimension.
My current digital camera is a Panasonic G2, which captures images that are 4,000 pixels long by 3,000 pixels wide. Each image contains 12,000,000 pixels (4,000 x 3,000), or 12 megapixels.
The resolution of an average computer monitor is between one and two megapixels.
A one to two megapixel computer monitor can not display a 12 megapixel photo without downsizing the image.
When downsizing, the computer averages the values of groups of neighboring pixels into single pixels. This produces a smaller, lower resolution image which can be viewed on an average monitor.
The above image shows the effect of downsizing.
Each small square inside each image is a pixel.
There are 100 pixels (10 x 10) in the lefthand image, and 25 pixels (5 x 5) in the righthand image.
Neighboring pixels in the lefthand image were averaged to make the righthand image. Some detail was lost.
Another tactic computers take with oversized images is to do nothing at all.
In that case only part of the image will be visible on the monitor, with the rest accessible through scrolling. Large images are often displayed this way on the web.
This picture of my wife is 800 x 600 pixels, which is about as large as you will see on most websites. (You may have to click on the picture to see it at full resolution.)
800 x 600 pixels is exactly 480,000 pixels, or a little less than half a megapixel.
That's good news if you are buying a camera to post pictures to the web. All cameras have more than enough resolution to post pictures to the web.
When buying a camera to post pictures to the web, instead of megapixels, pay attention to features such as ease of handling, or if you are an enthusiast photographer, maximum aperture.
How many pixels do you need for prints?
Printing is where the question of "How many pixels?" gets a little confusing.
There are a number of misunderstandings in cyberspace related to how many pixels you need to print at different paper sizes, most of them related to the number of dots per inch that the human eye is supposed to be able to resolve. Usually a figure of 300 dots per inch is cited.
The most common advice is that if you want maximum print quality, you need to print at 300 dots per inch, to take advantage of the maximum capability of the human eye to resolve fine detail. For the purpose of printing, dots and pixels are considered to be equivalent, so images should be printed at 300 pixels per inch for maximum print quality.
This is good advice, and it is true for images that are viewed closely, images that you hold in your hand or put your nose up to. However, it is not true for images that are viewed from farther away, which I will discuss later in this hub.
To determine the maximum print size for an image that is to be viewed closely, divide the number of pixels in each dimension by 300, which calculates the number of inches that can be printed in each dimension at maximum quality.
For example, in the case of the Panasonic G2, the long side of an image is 4,000 pixels, so 4,000 / 300 = 13.3, which means that you can print up to 13.3" on the long side and still retain maximum print quality. The short side of a G2 image is 3,000 pixels, so 3,000 / 300 = 10, which means that 10" is the largest you print the short side at maximum print quality. Therefore the largest print size at which you can expect maximum print quality for the G2 would be 13.3" x 10".
According to this formula, if you print a 4,000 x 3,000 pixel image at 13" x 19", a common print size, the print will not be as sharp as a smaller print when viewed closely.
Consider the above comparison.
The basis for all three images is the leftmost image which is 100 x 100 pixels.
It is scaled up to 200 x 200 pixels to make the center image, and 400 x 400 pixels to make the rightmost image.
Which image looks the sharpest?
I'm sure you would say the leftmost image. The center image is somewhat pixellated, and the rightmost image is very pixellated, seen from up close.
Now step back from your computer screen. As you get farther away, it becomes impossible to make out any detail in the leftmost image, and the rightmost image is not pixellated anymore. At some point the rightmost image becomes the best image. In fact, the rightmost image begins to look better than the leftmost image did at first.
None of the images are better or worse than the others. They can't be. They are all based on the same information, the leftmost 100 x 100 pixel image. What makes them look good or bad is the viewing distance.
Now lets go back to our old friend, the four pixel image.
We can "print" our four pixel image any size we want, but the image will not be sharper or less sharp in any of the cases. It can't be. The same four pixels are in each print. Our brain receives exactly the same information in every case. Different viewing distances might appear to make one image sharper than the other, but it is an illusion.
Now for a mental exercise, think of a 300 x 300 pixel image. You can print this imaginary image at 1" x 1", the supposed standard for maximum sharpness, or you can make each pixel 1" in width, and print the entire image as a 25' x 25' billboard.
Which print will be sharper?
If you are at the right viewing distance for each image, they will look the same. They have to. They contain the exact same information.
The first image has 300 pixels per inch, the standard for maximum sharpness.
The second image has 1 pixel per inch, but it is just as sharp, when viewed from the correct distance!
Dots per inch don't matter
Dots per inch don't matter.
Viewing distance does.
If all of the information in your image capture makes it into print, then that's the best print you can make.
If a print looks pixellated, then the viewing distance is wrong.
In that case, either make a smaller print or increase the viewing distance.
If you choose to make a smaller print, please understand that you are not making a smaller print because the larger print was less sharp. The larger print was just as sharp as the smaller print, that is, it contained exactly the same information.
In a sense, you control viewing distance with the size of the print.
Viewers will realize that they are to stand back from a large print, in the same way that realize they are to stand back from a large television screen.
Yes, the odd person will stick their nose up to a huge print, just as the odd person will stick their nose up to a 103" television screen.
But such people are the exception, not the rule.
You shouldn't print your photos, or arrange your living room furniture, with them in mind.
Get a good camera and take good pictures
I have two Olympus point and shoot cameras.
One is a seven megapixel C7070 from 2005.
The other is a 12 megapixel Tough 3000 from 2011.
The C7070 takes far better pictures, and makes far better prints.
The Tough 3000 is waterproof and great for kayaking, but the C7070 has a great lens.
All of the megapixels in the world won't make a difference in a print if the pixels don't contain good information.
You are better off with fewer megapixels and a camera with a good lens, then lots and lots of megapixels and a substandard lens.
The Nikon D800E has 36 megapixels. It is the highest resolution consumer camera ever made.
Yet if if I put a crummy lens on the D800E, or if the shutter speed is too slow, or if I don't focus, then I can get a blurry picture out of the D800E as easily as I can out of any other camera.
So the photographer is also more important than the number of megapixels.
Yes, all else being equal, the D800 will take better pictures than the C7070, but there is a limit to how much better.
If I were replacing my C7070 today, I'd get this camera.
There is a limit to detail.
Consider a pattern of a checkerboard on a wall.
You take a picture of it with the first two megapixel Kodak camera that I ever owned (which prints wonderfully at 8x10, by the way--I never tried it larger).
You take a second picture of it with the just released Nikon D800E, which has 36 megapixels.
Which picture is better?
There is a limit to the detail that you can capture in the scene. There are only black and white squares. Since the two megapixel Kodak captures all of the detail in the black and white squares, then its image will be just as good as the image from the far more expensive and capable Nikon. The images from both cameras will print equally well, at any size.
Yes, that is an artificial example, but there is a limit to detail in real life too.
If you are taking a picture of a model, then a single hair in an eyelash might be the smallest detail that you can reasonably hope (or want) to capture. Once you have a camera that is good enough to do that, and today's 12 megapixel cameras are good enough, there isn't much more detail that you can hope to get out of the scene that will be worth viewing in a print.
Sure, you can buy a 36 megapixel camera and then maybe, with a magnifying loupe, you can tell in a print if the end of the eyelash hair was split or not. But nobody looks at prints like that, and you are fooling yourself if you think they do.
In a picture of a model, viewers are looking at the model's face. In a picture of a bear in the woods, viewers are looking at the bear. Viewers are not focused in on the model's eyelash hairs, or twigs on the branches in the background behind the bear. That's why pictures often look better with the blurred backgrounds. Nobody cares about details that don't have to do with the subject of the picture.
There is a limit to the detail that you can capture (or should even want to capture) and display in a print.
Don't lose megapixels to poor framing
I'm a former math teacher.
I know better than anybody that you don't want to talk about math.
But we have to talk a little bit about length and area if we are to understand the difference between cameras with different megapixel ratings.
A typical 12 megapixel camera captures images that are 4,000 x 3,000 pixels. (Panasonic G2)
A typical 16 megapixel camera captures images that are 4,592 x 3,448 pixels. (Panasonic G3)
Not much of a difference, is there? With the 16 megapixel camera you get 592 extra pixels in one dimension, and 448 pixels in the other. Big whoop.
The reason that the increase in pixels along each dimension is so small is that megapixels increase with the square of any change in length, because megapixels are area.
Think about our 4 pixel image. It was 2 x 2 pixels. If we make a 3 x 3 pixel image, it has 9 pixels, more than twice as many as the 2 x 2 pixel image! How did that happen? A 3 x 3 pixel image doesn't seem more than twice as big as a 2 x 2 pixel image. But we increased the length of each dimension by a factor of 1.5, from 2 to 3. If we square this increase, 1.5 x 1.5, we get 2.25. So there are 2.25 times as many pixels in the bigger image as in the smaller image. Finally 2.25 x 4 = 9.
All right, enough math, I know you're sick of it. The point is this: There is not that much difference between a 12 and a 16 megapixel camera. The extra pixels that you get in each direction don't add up to a whole lot of extra resolution. Don't lose that extra resolution by poor framing.
Let's say you are shooting with your 12-megapixel camera and you see a pelican. You carefully frame the pelican to fill up the entire frame, and take a picture.
Somebody else comes along with a 16-megapixel camera and sees the same pelican. A little overconfident from the high resolution of their new camera, they sort of loosely center the pelican in the middle of the frame, with a lot of empty space on all sides, and take a picture.
Do you see what happened? The second person thought they had a resolution advantage by buying a 16-megapixel camera, but they didn't have that many more pixels in any one direction than the 12-megapixel camera did. And they wasted whatever small advantage they did have by only using a small portion of the available pixels in the frame. The better framed picture from your 12-megapixel camera ends up with more resolution than the poorly framed picture from the 16-megapixel camera..
The moral of the story?
If you are going to buy a high-megapixel camera, at least get close enough to your subject so that they are filling the frame.
Otherwise, you might as well be shooting with a much lower megapixel camera, because you are donating most of the megapixels that you have to empty space.
The bottom line, how many megapixels do you need for prints?
12 megapixels are enough.
OK, I admit it, I'm just like you. I want a 16 megapixel camera too. And there are some great deals on 16 megapixel cameras today.
But 12 megapixels really are enough.
I print at 8" x 10" a lot, and 12 megapixels is overkill for 8" x 10".
I print at 13" x 19" a lot, and 12 megapixels is tack sharp at 13" x 19".
I know of people who print regularly at 20" x 30" with 12 megapixels.
If you make wall-sized prints that people put their noses up to, then by all means pick up a 36-megapixel camera. It will make a difference in your photography.
If you only think that you are going to make wall-sized prints that people put their noses up to, but you haven't actually done it yet, in fact you haven't done much printing at all yet, then take a deep breath, step back, buy a 12-megapixel camera, and spend a few years taking and printing a lot of pictures to get a feel for what your equipment is capable of. In the future, if you need to upgrade, you will know what specification of your current camera is lacking and what you want to upgrade it to.
In the end, the most important factor in making good prints is learning how to frame and take sharp pictures. If you use a tripod, fill the frame with your subject, and put your time in waiting for the right moments, you can print at almost any size with almost any camera.