Copyright Michael Karbo, Denmark, Europe.
Chapter 9. Optical resolution
A digital camera has a maximum optical resolution. This is the resolution a camera was originally expected to work with. But you can also take photographs with a lower resolution. And with some cameras you can, in fact, obtain image files with a higher resolution than an optical. If you want to understand how a camera juggles around with exposures in different resolutions, then it is difficult to avoid the term interpolation. This is the method the camera’s computer uses to, among other things, guess the image’s colours.
Use optical resolution
It is the camera’s menu system that gives us the option of selecting between images with different resolutions.
A 5 MP camera can, for example, take image files in 5, 4, 3 and 2 MP resolutions. The advantage of low resolution is that there can be more exposures on the memory card. But if you have bought a 5 MP camera, you shouldn’t use it to take 2 MP exposures. It is important to learn how to make use of the camera’s full potential – instead of restricting the possibilities!
The opposite can also be the case: that the camera can take exposures in a higher resolution than the image sensor purely optically is designed for. As we shall see in a little while, this can be taken advantage of.
Selecting resolution. Here with a 3 MP camera, which gives you the choice of taking photographs in 1, 2, 3 and 6 MP resolutions. As 3 mega pixels is the camera’s ”real” resolution, then it would be natural to select either this or the higher resolution of 6 MP (Fujifilm S5000).
If you work with digital images in different resolutions, the term interpolation appears often. Purely theoretically, interpolation is the calculation of an unknown value based on two or more known values. In practice it means that a computer can enlarge the resolution of an image by converting it into new pixels. The term is maybe recognized from scanners, which are often described with very high – but interpolated – resolutions.
So what does interpolation have to do with digital cameras? Quite a lot, in fact, because a camera’s computer produces lots of interpolations. It does it in several different situations, where calculations are necessary:
These conditions are all based on interpolation, just like all image-processing programs use interpolation for lots of operations. So let’s briefly take a look at what, in fact, happens.
The picture’s resolution can be changed
Interpolation is usually described as a mathematical procedure, which can make a qualified guess. This guess could be the calculation of new pixels. If we have a digital image with a particular resolution, then we are sure of every single pixel’s colour value.
Figur 30. There were originally 9 pixels at the top. They are going to be resampled, so that the image will fill 25 pixels. This is done in two steps. First, the ”blank” pixels are put in the spaces in between. Then the colour value for each of the blank pixels is calculated. The result is an image with more, but “artificial” details.
But what happens, if we want a bigger version of the image? The program will interpolate so than every time a new pixel is put in, its colour is calculated on the basis of the surrounding pixels.
Colour values are calculated for unknown pixels, which are put in to increase the resolution.
When interpolation is used to enlarge an image, more colour nuances appear and because of this also more image details. The new details are, however, “artificial” – they have been created by a mathematical interpolation.
If an image is to be scaled down, so that the resolution is reduced, then you simply just remove pixels.
A camera’s software can itself reduce resolution. Selecting exposure in one of the smaller formats does this. During the exposure the camera takes a normal image in the optical resolution of the image sensor. Afterwards the camera’s software removes to the best of its ability image details while, at the same time, retaining the image’s original appearance as much as possible.
The process is called down sampling or down sizing, and it is easily done. The problem is, however, that image details disappear in relation to the original image. A high resolution can, you see, never be recreated from a reduced version.
Figur 31. This camera has an optical resolution of 5 MP, but the user can select a resolution of M1 (Medium 1). This gives 2 MP-images of 1600 X 1200 pixels. The only thing you get out of that is that the camera’s software throws 60 % of the image details away (here Canon G5).
Some cameras can produce image files, which have a higher resolution than the image sensor is built for. A digital exposure has been enlarged with the use of interpolation. The camera’s software calculates new image data, which is added to the existing pixel mass. The new pixels are “artificial” in contrast to the original pixels, which come directly from the camera’s image sensor.
Image quality is best when the interpolation is done with image data, which comes directly from the image sensor. This data is called ”raw”, and is found in a camera just after an exposure has been taken. If it is interpolated – before it is converted into image files – the result will be much the best, and very much better than if a computer’s image program is used to ”resample” the photograph, which is what interpolation is called when you do this.
Interpolation of ”raw data” is the best way to achieve really high resolutions.
In some cameras you can select an image format in a moderately enlarged resolution. Here the resolution is enlarged with circa 15% for every interpolation. The result gives, for example, 8 MP images from a 5 MP camera.
Cameras produced by Fujifilm have a special image sensor called Super-CCD. It is, so to say, designed to work with twice as much resolution in the image files. This is why you can always select image files in very high resolutions in Fujifilm’s cameras. The high resolution is in principle interpolated and, therefore, “artificial” but it can give extra image details if they are needed.
Interpolation is best when it is the camera’s raw data that has been processed. So if your camera can produce images in an interpolated high resolution format, then it is worth trying it.
Figur 32. Optical resolution can be increased by a camera’s software. Here the quality ”Enlarge Size” corresponds to a resolution of eight mega pixels even though the camera’s image sensor only produces five million ”optical” pixels (Olympus 5060Z).
Finally, the more advanced image-makers use RAW formats, which many cameras support.
RAW files contain the image’s raw data, and the trick is that they can be interpolated and processed in a computer. They are, therefore, widely used by professional photographers, whom are able to enlarge them with high quality interpolation in the brilliant program Adobe Photoshop Camera Raw plug-in.
What are we going to do with all of these pixels?
It is true that an artificially enlarged resolution won’t give any more genuine image details in the exposure. But if you don’t have a problem with room on your memory card or hard disk, then an image file with a really high resolution is often the best choice.
There are masses of pixels to work with in the image program and you can produce very big paper copies. RAW files, especially, can produce images in enormous resolutions with a very high quality.
Does the future hold cameras with a 100 MP resolution? It’s possible. Even though there are expenses tied to raising a camera’s resolution. Cameras and objectives will be more expensive to produce while at the same time there will be problems with so-called ”digital noise”. Light sensitivity will be reduced if resolution is raised. If we have larger image files then we will need larger memory cards and more computer power. Finally, the camera will use more power.
But it looks as if technological development keeps on winning over these sorts of boundaries and points towards still larger resolutions.
Figur 33. The camera’s resolution can in different ways be increased by interpolation. Interpolation based on raw image data that has come directly from an image sensor can give a very good result. The terms “raw data” and RAW file formats will be explained later on in the book.