I agree with cleantone on this. You don't get much improvement in accurate waveform reproduction in the audible frequency spectrum by going to the higher sample rate of 96 kHz, but the greater bit depth of 24 bits has easily demonstrable benefits. You'd be hard pressed to reliably identify the difference between two recordings of the same source, one of which was 24/96 and the other of which was 24/48 or 24/44.1. At the higher sample rate, you increase the likelihood of transfer rate limitations going through your USB port and to the hard disk. On Windows 2000 and later (any of the NT-based versions of Windows), you have one interrupt controller that handles all of the hardware interrupts, so a high traffic load on your USB port can affect how fast you can write to your hard disk and traffic to your hard disk can affect how fast you can take data from your USB port. So, you're more likely to get missing data in a 96 kHz recording than in a 48 kHz recording. If you've got a way fast machine, then it may not matter, but your run-of-the-mill laptop might not be able to keep up at the 96 kHz sample rate... at least not for an hour at a time.