That trick is basically to make windows use a larger cluster size when formatting the drive. This "cluster size" is basically the smallest chunk of data that can be read from the hard drive (by windows, this is not the same as a drive's block size which is always 512 bytes).
Using the Z:64 switch tells windows to format the drive using 64K clusters. That means that every file exists in 64K clusters on the hard drive. The downside is that a lot of space is wasted if you have a lot of smaller files. For instance, a 32 byte file is going to take up 64,000 bytes on the disk. Likewise, a 65K file is going to take up 128K on the disk. Now when it comes to audio you're usually talking about (relatively) small number of very large files, so the amount of wasted space is really pretty small. The supposed benefits of using larger clusters then, is that disk access will be faster because larger chunks of data will be read/written at once.
Personally I've never seen benchmarks to prove that larger cluster sizes are better. I mean it sort of makes sense, but when you consider the buffering that's already going on by both the drive, perhaps the controller, and the windows file subsystem, I'm not sure it's really that beneficial. But who cares, it doesn't hurt anything!
Now the "Z:" thing is what you'd use if you were formatting from the command line. When you format from the disk management snapin thing, it'll ask you for "allocation size", and a whole boatload of options. I would choose either 64K or "default"...or you could spend many hours benchmarking your various options and blah blah blah, but then you might just forget to record some damn music
Slackmaster 2000