multiplayer bullet time

The Agents vs Max Payne

In ye olde days of MPHQ, multiplayer for Max Payne was an issue. First the possibility of making a mod comprising a multiplayer mode was seriously discussed in the forums. [Then the topic became a nuisance, and finally a running-gag played on ↵n00bs.] Quickly it became clear, that it was impossible to achieve for certain reasons: from the technical side lack of access to the source code was a powerful argument, the problems arising with bullet time another. The essential knack lies in the very concept itself. A player going into bullet time gets the decisive advantage from it, because the gameworld around him/her slows down, but the player still can act—aiming the weapon has to be named here—in real time. Imagine a multiplayer game featuring bullet time. Say six players are together in one map. Every instance one of them goes into bullet time, in consequence the other five only can act in slow-motion. There will be no fluent gameplay possible anymore. Nobody would want to play a game like that. Dramaturgically it would be ideal if gameplay would not slow down for the other 5, but the actions of the bullet-timer would appear to them superhumanly fast. But then the gameserver would have to handle at least two time-lines and ultimately achieve the feat of looking into the future to be able to synchronize the actions inside and outside bullett time. Impossible! Impossible? Not quite, as it seems. The NewScientist reports on Finnish researchers [who else?] ↑Jouni Smed, Henrik Niinisalo and Harri Hakonen having developed a way to achieve the bullet time effect in real-time multiplayer-games. Like in The Sting and in IRC-black-magic, lag is the word of the hour:

Smed’s solution is to exploit something called a local perception filter (LPF). This is software that compensates for the natural communication-time delays which occur in networked games by rendering objects and players at slightly out-of-date locations.

In locally networked games, time delays can be as much as 10 milliseconds, while transatlantic games suffer a latency of around 60 milliseconds. However, the use of LPFs means players do not notice any time lag because events are ever so slightly slowed down until the game catches up with itself.

Using a test-bench game called ↑MaxMazeDemonstrator [download link there! | .zip | 80KB], Smed and colleagues found that they could also artificially introduce delays of up to a few seconds, allowing one player to slow down their environment and gain a strategic advantage, while game-time appeared normal to their opponent.

The three heroes have published a paper called “Realizing the bullet time effect in multiplayer games with local perception filters” (2004, 2005). The abstract reads:

Agent Smith in bullett timeLocal perception filters exploit the limitations of human perception to reduce the effects of network latency in multiplayer computer games. Because they allow temporal distortions in the rendered view, they can be modified to realize the bullet time effect, where a player can get more reaction time by slowing down the surrounding game world. In this paper, we examine the concepts behind local perception filters and extend them to cover artificially increased delays. The presented methods are implemented in a testbench program, which is used to study the usability and limitations of the approach.

screenshots from MP2-mod project The Agents

  • Post a comment

    Threaded commenting powered by interconnect/it code.