This is hardly a new idea, but there are certain problems with it.
Consider things like reports for Negative Attitude, Offensive Language, ect. With typing, the Tribunal can visually see the offences or lack thereof and thus make a judgement. With a voice chat system, thats a lot more complicated.
First off, if there was any hope of enforcing the same standards that are present in typing chat, all the voice chats would have to be recorded all game. Then, tribunal proceedings would involve tediously listening to about an hour of people raging, ect.
Secondly, keeping said recordings is borderline illegal with even some constitutional problems if it is not EXPRESSLY made clear that they are being recorded. This leads to legal issues, ect.
And if they're not being recorded, the Tribunal and report system is dead anyways.