Obviously this is a bit moot, as fifteen seconds of black does not a gripping clip make. Compressing a picture takes more effort, but the principle is the same. Whatever the method or algorithm used (I'll call them Codecs - it's short for COmpressorDECompressor) we need to get that frame into 833 bytes.
I've heard about two pass encoding before. What's with that?
Two pass is a much better method of compression. Basically, it's the webvid equivalent of 'measure twice, cut once'. On single pass encoding the codec will compress a frame, and then move onto the next one and so on. This means if something does not change from one frame to another it still gets entered into each frame's data. In two pass, the codec watches the clip once noting any places where things do not change much from frame to frame, and removes that data from subsequent frames. This makes the file smaller, and makes static images (backgrounds, close up pads etc) less blotchy (we'll see this later). Two pass is a good thing. Use it whenever you can.
Image size
There is a real trade off between image size and file size. If the frame size is too small people will find it hard to see what is going on, and too large means the file will be too big. I have always found that 320x240 is a good compromise, and that's what I've used for the examples.
The different Codecs
I'm going to discuss four compression methods, Real, Window Media, MPEG and Quicktime. To be fair, there are different versions of most of these, and we will talk about them at the time. To make some comparisons I recorded a ten second clip and encoded it in several different ways so you can see what produces the best results for you.
The original clip is a 45Meg uncompressed AVI at 24bit. Bear this in mind when you see the file sizes of the clips...
Nope, I'm lost Nope, I'm bored At last, something moving