This really belongs in the Cakewalk forum. However:
The audio driver bit depth can not be set any higher than the capability of your sound card. For ex., if you are using an SB Live, you would have to set the bit rate at 16, since the card can not record at 24 bits. If you have a 24 bit card, you can set it to either 16 or 24 depending on your preference (iow, you can always set it lower than the card's capabilities, but never higher).
The file bit depth, however, refers to the bit rate that you want to store the audio at. This does not have to match your sound card. So, for ex., you can record at 16 bits, but still have a 24 bit file (even though initially only 16 bits would be used). Some people argue that even with a 16 bit card, it is good to work with 24 bit files if you plan to do any DSP to the file. However, remember that if your card is only capable of 16 bit playback, you will be hearing a truncated version of the mix.
When you playback a file, the recording program keeps a certain amount of the audio readily available in buffers. It feeds the audio from the buffers rather than reading it directly from the hdd to help smooth the playback (sort of like a cache, if you are familiar with that.) The number of buffers and the size of the buffers determine exactly how much audio is kept available to the program. The more buffers you have and the larger they are, the smoother your audio playback will be. However, the downside is that this will also increase your latency. Generally you want to keep the number and the size of your buffers as small as possible without getting interruptions (dropouts) in your playback.
