Agreed on the packet loss, but it is the same for TS3. TS3 is just not open enough to tell you in the face, but it is definitely suffering in the same way from packet loss. The fact that the user can adjust the codec settings on the client side is a plus for mumble in my opinion, because it at least gives one a chance to tailor the setting to their needs. In TS3 you just have to stick with whatever audio quality you get with only minor tweaks possible for the audio input and output. Best example is the noise suppression which is tunable from off to between -15 and -60dB in 1 dB steps in mumble, in TS3 it is just on or off and in my case just does not suppress the noise enough.

In the end it still boils down to take your time and be really careful when running the audio wizard and follow the instructions to the letter, I cannot stress that enough. And you are given 3 presets for your outgoing codec quality settings, so you don´t really have to know anything about the underlying codec, it is sufficient if you know your upstream bandwidth. But the most important step is the audio levels tuning, being sloppy there will only get people complaining about high pitched surges comming from you, and the 2nd most important step is for the voice activity detection.

TS2 is a different story, but I was aiming for an alternative to TS3 mainly. I would really appreciate if when the time comes we do not switch to TS3 but to mumble, great part of the reasoning being the cross platform compatiblity. The Teamspeak team do a "great job" at implementing deprecated sound interfaces under linux which gives me awful headaches. Looks to me, they lack behind about 3-5 years in that area and their release cycle leaves me not overly confident. By the way, am I the only one to notice that the TS2 client we are currently using is still a release candidate? A perfect example that they don´t care about getting one thing right before they begin another one.