[asterisk-dev] iLBC packet loss concealment (was: code-cleanup concerns)

[Matt Ranney]

On Apr 16, 2006, at 6:33 AM, Steve Underwood wrote:

> Skype is certainly better than the PSTN when going Skype to Skype,  
> but why do people think it sounds better than the PSTN when going  
> to the PSTN? Is it just psycological, or are they using an  
> environment (e.g. headset) that just sounds better than their usual  
> phone? Having they been using a Cisco 7940 with G.729, where a real  
> PSTN to PSTN connection would sounds rather better? I find it an  
> interesting comment that Matt made.

Part of it I'm pretty sure is that headsets just sound better than  
handsets in most cases.  Part of it could also be that Skype/ilbc  
wideband sounds better than g.729.

Those two factors aside, I don't think that Skype works better in my  
environment because of codecs and sample rates.  My users certainly  
claim that Skype SOUNDS better, but they also claim that it WORKS  
better, meaning that it has less dropouts and garbled periods.  I  
think Skype works better because of better jitter buffering, PLC, or  
whatever other magic they employ.  Even in cases where we don't allow  
UDP through the firewall, users still claim that Skype over TCP works  
better than asterisk over UDP.

I've tried the async timing patch and all manner of jitter buffer  
parameter tweaking in asterisk.  Skype over TCP is still somehow  
better than IAX in dealing with arbitrary WAN links.