Peer-to-Peer immersive voice communication

[Dannythm]

This thread is for the devs to see because it's really (and I mean really) technical, but any discussion will be welcomed as well of course.

So I was looking for papers to use as reference for my current work, and I bumped into this one, and as it's more or less recent, (it was presented in ICPADS 2011 conference, in Tainan, Taiwan), I thought maybe the devs didn't knew about it:

Peer-to-Peer Immersive Voice Communication for Massively Multiplayer Online Games (It's a IEEExplore link, so unless you have an account or access from some University you cannot access the full text, but I'm sure Square Enix has plenty of engineers who have IEEE accounts).

So in short, this paper presents a system to include voice chat in MMO's without being a source hog, quoting from the introduction of the paper:

Text chat is the most common way for players to exchange messages with each other by keyboard typing, which will distract players from the gameplay. With the richer and richer game content media and gameplay styles, voice chat may be a better way of communication.

[...]

Currently, most MMOGs are based on the client/server architecture, which has the inherent limitation on scalability due to the constraint on available network bandwidth and computing resources. In such a case, additional voice transmission will seriously degrade the scalability of the MMOGs. Third-party voice communication software, such
as TeamSpeak, Skype and Ventrilo, may be a transitional solution to meet the needs of voice communication in MMOGs. However, when using these kinds of software programs, users have to manually solicit members of a collaborative team or a special community to form a conversation group. These third-party voice programs lack the mechanism to synchronize with the states of games. When the game state changes, the users have to manually adapt their contact list of the voice communication program to fit the dynamic conditions of the games. Moreover, the
programs are also based on the server/client architecture and thus suffer from the scalability problem.

[...]

In this paper, we propose a peer-to-peer voice communication scheme, called immersive voice communication (IVC), to provide MMOG players with the immersive experience to hear the voice of neighbor players within the area of interest (AOI). For a player, its AOI is the area of a circle centered at it of a fixed radius, and it only interacts with other players within the area. By observing human behavior, we have found that a person usually intentionally ignore the surrounding sound when he/she
focuses on a specific voice. Therefore, a player’s AOI neighbors are classified into two categories: the listener, who focuses on listening to the player, and the overhearer, who just overhears the talk of the player. In is noted that the latter consumes much less bandwidth than the former, which can be utilized to reduce the traffic load. We also develop a relation model for a player to determine which AOI neighbor to listen to according to the criteria such as the distance, orientation, and social relationship among the players. IVC is constructed on the basis of the peer-to-peer (P2P) overlay network. It can be integrated into existing MMOGs to render them with the voice communication ability without causing too much additional loads to them.

I won't quote more, because the rest is just technical stuff about how they achieve that and some early experimental results that look really promising, if anyone wants more info I'll gladly explain the stuff a bit more deeply, but with those fragments from the introduction I think it will be enough to give the general idea.

Anyway, I would for sure LOVE to see this implemented in FFXIV ARR, but as always, with a toggle on/off option, because we know not everyone is always as polite as they should be, but this would add a lot of immersion and depth on the game.