Screencast recording consists of two parts: capturing the screen and then encoding the end result into the desired codec. As it stands now, the freescreencast.com screencast record uses Windows Media Encoder for the screen capture, which has some issues.

The first part, the actual capture, concentrates on efficiency, quality, and speed. This phase needs to capture a nearly perfect copy of the screen but must do so in the most processor effecient manner possible. Very little processing of the stream takes place; therefore, it may generate a large temporary file, which is fine as a trade off.

The second part, the actual encoding, is where the compression takes place: colors are reduced, only those parts of frame that have changed are included into the stream, etc. An extra benefit of the two part capture is that one screencast recording session can then be encoded multiple times: with different quality settings for different file sizes or even encoded with different codec all together, i.e. FLV (flash) and AVI.

The end result is a whole new section of code for capturing and storing the intermediate version. This eliminates a dependency and increases flexibility, a win-win situation.

My holy-grail of compression quality is to allow for screencasting of photoshop / paint shop pro / gimp sessions with very little to no artifacts introduced. This is an area ripe for the benefits of screencasting, but it has to be right before I introduce it to that segment of the market.