Andy and Eyal are each part correct, as I see it. Harmonizers separate the formants so that vocal pitch can be shifted while preserving the character of the sound. In addition, many harmonizers allow some formant processing so that, for example, a male voice can be made to sound like a female voice, and vice versa.
Many samplers and sample playback modules, e.g., those from Emu, allow rhythmic loops to be expanded or condensed in time while preserving the formants.
On the other hand, I do not think that vocal harmonizers can control time (since they operate with a real-time input). Most samplers with "beat munging" do not allow much control over the formants in real time.
So, Variphrase technology seems new in that formant, time, and pitch can all be controlled in real time. As a result, you can take any two vocal phrases in different keys and tempos, change the melodies so that they are harmonized, change gender, and sync the phrases all using real-time controls. So I think it is fair to say that Variphrase technology goes beyond what else is available.
Above, I am considering Variphrase as implemented in the VP-9000 (of which I have seen a video demo). I am not sure which of these capabilities have been retained or omitted in the VA-7.