Digital video : Podcasting

To convert analog video into a digital format, the process is similar to digitizing audio. The incoming analog signal is sampled at discrete intervals, and the values are stored. Each sample is called a picture element, or pixel. To faithfully represent all the information in the video signal, it was determined that each line would be sampled 720 times. Combine this with the 480 visible lines, and you end up with a screen resolution of 720×480.

When we're digitizing video, we're storing values for each and every one of the pixels, of which there are quite a few:

720 * 480 = 345,600 pixels

Multiply that by 30 frames per second, and we're looking at over ten million values that have to be stored every second. For each one of these values, we have to allot a certain number of bits to store the value. Even if we try to limit the number of bits we use to store each value, we'll still end up with a very large file. We'll find later that this becomes even more of an issue when we're encoding for a podcast. In order to send files over the Internet, we must reduce them to a manageable size, and doing so with video requires some compromises.

One of the main compromises that can be made is the encoding scheme used to assign values to each pixel. You can assign this value in a number of different ways. Because all colors can be made out of red, green, and blue, one approach measures how much of each color is present in a pixel and then assigns a value accordingly. This is known as RGB encoding, which is the default method used on computer monitors. Different types of RGB encoding are named according to how many bits are used for the value, such as RGB 24 and RGB 32. RGB encoding can be very high quality, but it isn't the most efficient way to encode video information.

We learned that our eyes perceive light as being composed of luminance (brightness) and chrominance (color content). Our eyes are very sensitive to brightness and not so sensitive to color. Encoding luminance and chrominance information is a much more efficient way of encoding, because the color information can be compressed, and we won't notice the difference. We can get the same video quality, but with a lower data rate. Table 1 illustrates the bit rates and files sizes of some common video encoding schemes.



When digitizing, the best approach is to capture at the highest possibly quality. Starting with the highest possible quality gives you more flexibility during the editing phase and provides better raw material for your video podcast. However, the quality of your video capture may be limited by your equipment. If you're working with the DV format, it is encoded using the YUV encoding scheme and compressed at a 5:1 ratio. If you're trying to capture uncompressed video at full resolution (640×480 or above) you must have very fast hard drives and plenty of storage. Aim for the highest possible quality, and settle for whatever works within your limitations.