Efficient speech and channel effects using a modified low bit rate speech codec

(Or, the architecture behind TinySpeech for Games™)

The Solution, continued

The channel effects can include a mix the following:

First, increased comfort noise level. Comfort noise is sent between words, and anytime the sender is not speaking. This can act like an unsquelched radio. This does not alter the CPU requirement at the sender or receiver.

Second, random noise can take the place of speech frames. The magnitude of the noise, the percent of static frames, and the length of the longest noise pulse can all be set. This will reduce the CPU requirement at both the sender and receiver, and the higher the percent of static the less CPU cycles will be used.

Third, speech frames can be sent multiple times to simulate the echo of packet loss. This reduces the CPU requirement at the sender, and does not alter it at the receiver.

Forth, the background noise level can be increased. This slightly increases the CPU requirement at the sender, and does not alter it at the receiver.

Fifth, the speech can be over modulated and clipped. This slightly increases the CPU requirement at the sender, and does not alter it at the receiver.

Conclusion

I am sure there are other effects I, or other people, will come up with before TSGs is released. I look forward to these ideas being implemented in future games to increase immersion. Graphics quality is going up every few months, and games are yearly taking advantage of this to become more lifelike and immersive. So why is multi-player chat lacking this same immersion?


Prologue and Introduction
Background
The Problem
The Solution
The Solution continued and Conclusion