I don't know how Kvirc intergrates voice support, but I feel if there's going to be voice support that there should be 2 methods.
1. Same As SOUND. For simple one-time (brief) voice messages to the channel or query. The person sending their voice issues a /VOICE [TARGET] command, at which point the client brings up a record dialog / begins recording. When finished recording, the client issues a CTCP (eg: VOICE IP PORT, or DCC VOICE IP PORT), or it could be 2 stage like SOUND with !Nick Requests.. then those interested simply connect to the sender and receive the .wav/mp3 file, which gets placed under a \voice folder or simply deleted after being played.
2. Same As CHAT. For voice conversations, one would want to set up a DCC CHAT which either has transparent CTCP VOICE commands sent over the chat whenever a party talks, or would open a new DCC connection where streaming voice is sent back and forth simultaneously.
In all honesty, I like Yahoo Msgr's implimentation of Video and Voice chat. They both seem to hug the median between latency and loss, providing a tolerable listening/viewing experience which uses little bandwidth and doesn't sound/look too bad.
- Raccoon
