I am just finishing my final year project. Spent the whole year messing around with a way of creating games with voice recognition. Did it all in VB.Net too.

VOIP sounds good. When you choose what platform to develop with, just use the best tool for the job. Don't purposefully make life difficult for yourself because of your universities skills alignment. So what if all the lecturers hate Microsoft and love C++. You do what you think is best, for your project. You can even white a whole section of your final report slating C++ and say why C# is so much better if you like, as you need to state why you chose to develop with whatever you did over what you didn't.

I think the best projects have been the ones that are scalable where little success is good, but more success is better. That way the project scales well depending on the time you have. And if you find you can do something cool while researching or developing then your project can be adapted quickly.

Best bit of advice I can give however is to tell you to do a little bit of footwork before September. Get a very basic proof of concept working. Then you will know if VOIP is what you want to do or not. And you won't wast time understanding the very basics of your project's underlying technologies.

P.S. Also remember that when you go for job interviews having a cool final year project that you can talk about could help get you the job.