There are two places where the audio quality is specified: 1) where the room is created or edited, Community Server Administration, and 2) where the room is configured, Room Administration. The same parameter is set/changed in both places.
What is the best setting for this parameter?
Before I answer that question we need to briefly discuss what we are talking about. We are speaking about the quality of the audio format that is transmitted to each user. We are not speaking about the quality of your sound card or the speaker quality. Those components will have a impact on the audio quality but are not addressed in this article. An audio file format is a container format for storing audio data on a computer system. Uncompressed audio formats retain the best audio quality but require very large amounts to storage space and are therefore not feasible for Internet usage. To solve this problem, audio files are compressed using specific algorithms. The algorithms will compress the audio file into something smaller and when the file is played the same algorithms will convert the file back to the original state, well almost - it is never exactly the same. We don't actually store the audio format. For web conferencing we stream the data to your computer. The issues however are the same. We use a MP3 format ( MPEG-1 Audio Layer 3) for the audio stream. MP3 uses a couple of parameters to determine the compression ratio. One of those parameters is the bit rate which we have set at 16 kbits/s (kilobits per second) and the other is the sample frequency which you can change. The higher the number the better the quality. But there is a catch. A higher value also creates a larger file or data stream. A setting of 8 kHz (number of samples per second) has a quality similar to a low/medium telephone.
With that background we can answer the question: "What is the best setting to use?".
Actually there is no right or wrong answer to that question. The question does require further consideration and it is likely you will need to make some compromises.
You might ask: "If higher is better, why not simply use the highest value possible”. A good question!
As an example we will look at using a setting of 48kHz. With that setting the bandwidth needed to stream the audio will be between 90 and 95 percent of the available bandwidth available with 56Kbps dial-up services (high speed ADSL and cable services would use 1 to 2 percent of the available bandwidth). That means, for dialup users, everything needs to work perfectly: no Internet congestion – no line errors – minimal applications running on the PC – and so on to deliver acceptable audio quality. That probability is highly unlikely. Therefore in this case the audio quality for dial-up users will be compromised and some parts of the conversation will be interrupted and even lost. It would be equivalent to taking a cassette tape, cutting out an eight of an inch here, a quarter of an inch there and so on and playing back the spliced tape.
You might ask: "Well if using high value for this setting has these issues, why not opt for the lowest value available"?. Another good question!
A value of 8kHz is equivalent to medium to poor telephone quality. Therefore, by using that setting you will penalize users with high speed ADSL and Cable services who have the available bandwidth (1 to 2 percent) to successfully process an audio data stream with a setting of 48kHz.
Another factor to consider.
Bandwidth is also needed to send other information to each user during a conference or meeting. Text chatting is usually quite small and is not an issue but web browser content may add significant bandwidth usage. The technology doesn’t push the actual image to each user, unless you are using the Whiteboard or Desktop Sharing, but it does send the URL and the page associated with that URL will be downloaded and displayed on each user's computer. A simple web page with mostly textual content would have little impact however a complex web page with many images could have a significant impact. The extreme example would be pushing a full page image. In that scenario the bandwidth of a 56Kbps dial-up service would be 100 percent utilized for a period of 10 to 15 seconds, if everything worked perfectly.
Summary
So what is the best setting for the audio quality? The answer is: "It depends!". I recommend you use the information from the above discussion to choose a reasonable choice that will work for you. After making that decision I would recommend some experimentation to validate that decision.