I use the WebRTC and Web Audio API to capture the users audio from their microphone, compress it in chunks, and then upload it to the hosts dropbox account. This all happens in the browser. The server simply serves the app and sends some websocket messages back and forth so the host can start/stop the recording.
That depends on how bad your internet connection varies. Skype compression artifacts can be brutal. But yes, you are correct in that the microphone makes a big difference as well as the room you record in.
That's the primary problem with using Skype for podcasting: you have no control over codec selection. When Skype detects network disruptions, it automatically changes codec to reduce bandwidth usage. Granted, if your network cant transport at the higher quality codec, you're going to get drops in audio anyway. Personally, I'd prefer the control though. It's easier to wait out a brief disruption in bandwidth than it is to coax Skype back up to a better quality codec.
