In both cases there will definitely be performance gains to see from either. Hard drive performance is a function of a few things, such as:
- Drive bus type (IDE ATA66/100/133, SATA150/300/60, SCSI, FC, etc. Affects the upper bound maximum speed, and burst speed from the drive's cache)
- Drive rotational speed (Affects sequential read and write speeds and seek latencies. Faster platters = R/W head gets from point A to B that much quicker.)
- Drive size (The larger a drive, the greater the density of data per square inch, so read/write speeds are that much faster.)
- Number of platters (The fewer platters there are, the greater density (again). Also, noise, heat, and power consumption tend to be lower for a drive that only has to spool up one metal disk instead of five.)
- Drive firmware revisions (Any gains from these would pale in comparison to changes in the above, but for conciseness's sake, it's possibly a factor.)
Today's 1TB+ 7200 RPM hard drives are quite capable of outperforming older 10KRPM server hard drives simply due to the incredible amount of data flying under the drive read/write heads every rotation. Additionally, the price per GB sweet spot lives up in the highest capacity drives.
I believe that at the time of this writing, the greatest platter density is about 500 GB/in^2, which Samsung uses in their Spinpoint drives. These manage enough read/write throughput to actually saturate an ATA133 connection (finally!), so a 1TB+ drive might be choked a little under a SATA150 connection for bursty reads form the disk cache.
Either way I would definitely suggest grabbing a ~$100 USD 1.5 TB drive and connecting it via SATA150. SATA also has the bonus of being the cheaper interface, at the moment; equivalently sized IDE drives tend to carry a bit of a price premium now for being the older, deprecated technology.