Digital Photographic Sensors
In the realm of digital photography there are tons of things that can change the quality of the picture the camera takes. Most of these things take place right within the camera. To start with, if you’ve got a lousy lens, it doesn’t matter if everything else is great because light hitting great stuff will have been ruined before it got there. From the other direction, the same problem can exist. If your sensor sucks, it makes no difference how pristine the image is when it hits it. Besides that, the quality with which the camera plays with the data it gets has a huge bearing on the final output as well. That can mean how well the camera does its ISO noise reduction, how good the JPEG compression algorithm it uses is (or if it’s used at all), methods of analog to digital conversions, and so on.
The two main types of sensors in a digital camera are CCD (Charge-Coupled Device) and CMOS (Complementary Metal Oxide Semiconductor). The differences between the two are huge, and arguments about which is better abound. The basic idea of both is to collect light and spit out a bunch of data in the form of a photograph. The similarities between the two almost stop there.
A CCD typically has a single “drain” point. Through that spot all of the data for the entire picture must be passed. The kicker is how that data ends up in that spot to be drained and the form the data is in when it is drained. Each row of the image actually cascades across the entire sensor into a holding row (which is sometimes the last row in the image itself) and then across that row to a single dot. All this shifting of charge is done as actual voltage signals based on massive input from a battery and signal strength data from each pixel of the image. Also, as voltages are shifted across the sensor, data of previous pixels can remain slightly and end up added into the latest shifted row. This can result in smudging within the picture or famous side effects such as “purple edges.” The advantage is that all data is uniformly interpreted through a single drain which has what I like to call a “single opinion” about how each color looks. The disadvantage is that to access any single pixel the whole sensor must be flushed. When the data leaves the chip it is no where near ready to be stored in an image file destined for a printer or computer as all we have is a stream of voltages. These voltages need to go through conversion chips to handle all the steps of converting them into digital data.
...
