I think the big difference will be in experience on the iPhone. Older systems were plagued by low resolution, crappy compression and lag / latency due to network routing issues (proxies etc.).
Judging from the videos, the iPhone 4G uses hardware to encode/decode the VGA video in H.264, which makes for low latency and high quality of that part in the processing chain, and by using standards like STUN/TURN, allows for mostly direct (p2p) connections between IP endpoints, thus using the best possible path for the media, with the result being low latency.
It will be interesting to see if Apple can coax the carriers into QoS for the video, once it travels over their 3G networks.
And it's also interesting that Apple chose pure SIP, rather then the IMS SIP that carriers have been trying to cook up over the last (almost) decade.
[EDIT] Digging some more, Apple is working with the carriers to get their networks ready for video calling, which probably means that they will offer QoS, for a price. Prepare to pay for the privilege...