It won’t have passed you by that you can’t turn around these days without a camera company bringing out yet another model with even more pixels than the last.
On the face of it, it would seem a good thing and in one respect it obviously is. Resolution.
The more photo sites there are, the more little bits of data there are to make up a print. The more the little bits of data the bigger the print can be (for a given print resolution). The more the merrier. If only it were that simple. You see, it’s not just the number of pixels that govern the quality of the image … and call me old fashioned, but isn’t it quality that is the goal?
So, it’s not just the pixel count. It’s also the physical size (the area) of the sensor. To over simplify, if you squeeze more photo-sites than you had before onto a sensor, they either have to be smaller than before, or closer together, or both.
Simply said, if a photo-site is smaller, it’s light gathering capability (i.e. sensitivity) is lower. To compensate for this the signal must be amplified more. There is always a certain amount of electrical interference but increasing amplification exacerbates it. This increases ‘noise’ which exhibits itself as visible artefacts or pixels which have values other that those appropriate to the image. This noise, put crudely, has an effect opposite to that expected from increased resolution. Another source of ‘noise’ is the close proximity of one photo-site to another. More interference.
There are solutions, which to a greater or lesser degree ameliorate this problem. One way of approaching this is to process out some of this noise in the camera. It is important to understand that noise is random. That is, if one photographed the same scene twice (no changes whatsoever) the noise pattern in both frames would differ. One can use this phenomenon to our advantage by superimposing these image files. The ‘in camera’ solution I mentioned fakes this random duplication, using a pre loaded program and it works reasonably well (if available).
Another solution is to shoot a number of identical images and cobble them together (not quite as simple as I have made it sound). This is staggeringly effective and hugely improves image quality. It demonstrates perfectly how much ‘noise’ is being generated in your camera when previously you may not have been aware of it!
Anyway, all this may be making you wonder why camera companies jam so many pixels onto small sensors in the first place if it creates a counter productive effect. Well, that’s the easy one. Marketing. Simple as that. Sales.
Ask yourself, what do people want? In the main, they (non professionals) want small cameras. You can’t have a small camera with a large sensor. I may explain this on another occasion, but suffice it to say, if you did put a large sensor in a small camera one would end up with all sorts of problems to do with perspective and lens issues.
So, a small sensor is required. Easy. Now populate this small sensor with an appropriately small number of photo-sites. Well, you would think so wouldn’t you? This is where the other requirement of the buying public tends to come to the fore. ”Mine’s better than yours!”
The “mine’s better that yours” syndrome works like this. The potential camera buyer is scanning the shelves in the shop and sees two ‘point and shoot’ cameras of similar price. One is 10 mega pixels and the other is 12 mega pixels. He knows his next door neighbour had a 10mp camera. Choice made. In addition to this, he naturally thinks that 12 is better than 10 anyway. Why wouldn’t he? Camera companies and their marketing people are well aware of this and the rest is history.
Here’s another bit of fun. One day, if I have nothing better to do, I will work some numbers with reference to lens quality and sensors. Suffice to say here though, the resolution of my mobile telephone’s lens would have to be many times (big number) better than my best professional lens, in order for it to resolve the number of pixels on the phone’s camera sensor. That should be all you need to know about the marketing meeting that happened before the designers went to work.
The next time you come across a mobile phone that has an equal or higher pixel count on it than a quality single lens reflex (SLR) camera, I doubt you will think, for one second, that the quality of the image it can produce will be even remotely comparable … or will you?