Dolby Axon enables surround sound chat for online gamers

Dolby has developed new system for speech transmission during online games, and is calling it Dolby Axon. By providing both direction and distance cues, Axon enables speakers to be "placed" in virtual sound space. Since Axon works with particularly low bit rates, and positioning the speakers consumes little computing power, servers should be able to process and transmit the words of thousands of players simultaneously. This makes the system especially interesting for "massively multiplayer online role-playing games" (MMORPGs). No special hardware is required. Players need only an ordinary microphone or headset, and the audio can be output through a loudspeaker system having up to 7.1 channels, or headphones that simulate surround sound using Dolby Headphone.

Good quality with high compression

Before transmitting the speech from the client to the server, Dolby Axon first removes all hiss and ambient noise interference, which helps to maintain good compression ratios, and adjusts sound levels. The upstream feed to the server manages with no more than 16 kbit/s. When a player is not speaking no data is transmitted. The server collects all the client speech streams and, depending on the position of the individual characters, computes a downstream signal for each player that contains a mix of all the conversations audible to him. Although the downstream signal includes surround sound information as well, it requires only 24 kbit/s, or even 16 kbit/s at reduced quality.

However, in order to keep the computing load on the server within limits, Dolby Axon uses only rudimentary parameters for direction finding. The distance between speaker and listener is simply simulated in volume level, so more distant speakers are quieter. Dolby Axon does no filtering of any kind, neither attenuating high frequencies more strongly nor increasing the reverberation component for distant speakers. The system can take account of occlusion, i.e. if there's a wall between speaker and listener, there is additional attenuation. A game developer can specify the audible range of each individual player, so a listener might hear a babble of voices from all the players within a range of 30 metres, or just a few speakers standing right beside them.

Server records live conversations

But Dolby Axon doesn't just look after the high quality of its speech transmission, it also incorporates a "security system". In order to be able to keep track of insults and other verbal aggressions during chats, the server buffers all the speech streams for five minutes. If a player complains about someone else, the server administrator can use a clear graphical interface to replay the conversations that took place in the player's virtual environment.

This function is certainly likely to interest Dwight Toavs, a professor at the National Defense University in the USA, who recently complained how difficult it was to monitor conversations during MMORPGs like World of Warcraft, which meant that terrorists could use online games to plan their attacks. If Dolby Axon buffers conversations for five minutes, security authorities should find it easier to tap into the stream.

What's the sound like?

The Dolby Axon system was demonstrated in Paris with a handful of people playing Unreal Tournament 3. Given the low bit rates, the general intelligibility of the speech was exceptionally good, although compression artefacts were perceptible. Efficient noise suppression made a particularly useful contribution to comprehensibility. When fellow players moved around a player, their speech came out of the loudspeakers from the correct direction. This system is a distinct advance on previous chat systems, but Dolby Axon is still far from reproducing conversations in virtual space as naturally as in reality. Using only a drop in volume to indicate the speech of the more remote players still sounds very artificial.

Compared with other environmental audio systems, such as EAX by Creative Labs, Dolby uses only simple reverb parameters ("small room", "cave" etc.), which are not individually adapted to each voice, but can only be set globally. That makes Dolby Axon approximately comparable with earlier EAX versions, up to 2.0, but it doesn't achieve the stereophonic sound quality of newer versions such as EAX 5.0.

Only for PCs at first

Dolby Axon is initially to be implemented in PC games running under Windows XP or Vista. it uses its own software-based sound engine under DirectX, with the CPU doing all the stereo sound calculations, and does not need a special sound card with hardware-accelerated 3D sound. Dolby Axon can run on any system, even one that only has an on-board sound chip. But it can't unfortunately be combined with other sound engines such as Creative EAX, nor does it support OpenAL.

Next, Dolby will adapt the system for Playstation 3, in order to simplify its implementation for cross-platform games. Dolby say the Xbox 360 won't come until later, as Microsoft's console has already implemented a widely accepted speech transmission system, and other systems like Wii are not being looked at. Dolby would not say which PC games were now being developed with Dolby Axon, or when they would come out. At the moment, says Dolby, it's simply giving developers an SDK and advising them on implementation.

(trk)