How Cameras Sensors Work And The Benefits of HDSLR Sensors for HD Video

By now, you have probably heard that more megapixels in your camera is not necessarily better for still photography. If not, check out Ken Rockwell's article explaining the megapixel myth. This is especially true for video, which at full HD only uses 1920 by 1080 pixels – or about 2 megapixels. However, one reason why DSLR cameras are clearly better than point and shoots is the large, advanced sensors that continue to improve. But the reason they are large is not just because they're packed with pixels but because they have larger photosites.

Digital Camera Basics

But lets back up a bit. Exactly how does a camera work, anyway? Basically, they work similarly to your eyes; light photons are focused into a central part of the back of your eyeball, where it is converted to an electrical signal and processed by your brain. Similarly, cameras work by focusing light onto a sensor. The photons smack the sensor, then are converted to an electrical signal that can be stored to create an image.
Sensors are made up of photosites – an array of square, light-sensitive spots – that collect photons like a bunch of square buckets on a tennis field would collect water. The more the buckets are exposed to the rain, the more water they collect just as the longer the shutter on your camera is open, the more light exposure you will have available for your image. The buckets are not placed directly next to each other; in fact, they may only cover 25 percent of the tennis court, so much of the rain that falls is lost. In camera sensors this is true as well – much of the light falls between the photosites and is lost. Continuing the analogy, the buckets capture not only rain water but also debris, so the water is not pure water but contains contaminates. These contaminates are called “noise” in images processed by camera sensors. (Note that the noise is increased with higher temperatures, so if you experience overheating in video mode, you'll probably also have noisier images.) Read more about signal noise here.

Why HDSLR Cameras Excell In Low Light

How do we limit the noise? By increasing the amount of usable signal we can work with. This can be done by increasing the size of the photosites (or buckets). Capture more light and you'll have more signal, creating a higher signal-to-noise ratio. This is why DSLR cameras are better than point-and-shoots and why HDSLR cameras are often better than professional camcorders at capturing images in low light; they have larger photosites and larger sensors. And they continue to get better. Camera companies are reconstructing the way everything is organized and using different materials so more light can reach the sensor. Read more about how digital camera sensors here.

ISO Comparison with Canon EOS 400DISO Comparison with Canon EOS 400DWhen you crank up the ISO for low-light conditions, the camera amplifies the signal to use more of the signal that is captured by the camera. When more signal is amplified, more noise is also amplified, resulting in a grainier, noisier photo. Of course, if more signal is available from larger photosites, less noise will be included. In video this is called “gain,” but camcorders suffer from noisier images because the image sensors are smaller and have less signal to work with. Knowing how digital cameras work will help you make wise decisions when purchasing a camera and creating images.

Which ISO?

Generally, it is advised to stay below 1600 ISO when filming video. You can push this in newer cameras, but many still try not to breach 1600 when filming to produce the cleanest images. When using Canon cameras, Shane Hurlbut, ASC and others have found that you should stick to ISO's of 160 ISO, 320 ISO, 640 ISO and 1250 ISO for the greatest signal-to-noise ratios (see Get Native With Your ISO at The Photoletariat).


Mar 16, 2011
by Anonymous


I have had so many people try to explain to me how a camera actually works and what puts certain of them above others, but it wasn't until I read this article that I actually "got it". Thanks!