[MUD-Dev] online voice communication

[Koster]

From: Amanda Walker
> On 11/25/02 1:07 AM, Ted L. Chen <tedlchen at yahoo.com> wrote:
 
>> The raw phonetic translation seems to be quite lossy and
>> susceptible to sample noise.  It gets worse if the speaker
>> doesn't talk in their normal tone or changes it halfway, like a
>> squeaky "Lookout!!" gets translated into 'loo-phiz' since the
>> 'out' hits the microphone gain.
 
> This is a problem; so is delay.  Speech -> phoneme -> speech will
> produce much more of a delay than a simple FFT -> twiddle -> IFFT
> (which is what Xbox Live and most other voice masking systems
> use).  Hearing your own speech with a delay is really, really
> disconcerting, and conversing with a delay is as well, since it
> interferes with the normal timeouts for conversational cues.  Now,
> it's possible to learn to live with this, but it takes practice
> and is not particularly "natural".

The pace of the typical online conversation, and the difficulty of
"filtering" it to selectively hear only people you want to hear, is
also a major problem. Check out a crowded area on Cybertown
sometime...

-Raph
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://www.kanga.nu/lists/listinfo/mud-dev