Sophisticated applications of Internet multimedia conferencing will become increasingly important only if their users will perceive the quality of the communications as sufficiently good. The result of extensive experiments has shown that audio is frequently perceived as one of the most important components of multimedia communications. Unfortunately, the actual architecture of the Internet is not a good environment for real-time audio communications, since very high transmission delay and transmission delay variance (known as jitter) may be experienced that impair human conversations. Hence, in the absence of network support to provide guarantees of quality to users of Internet voice software, an alternative to coping with problems caused by delay and delay jitter is to use adaptive control mechanisms. These mechanisms are based on the idea of using a voice reconstruction buffer at the receiver in order to add artificial delay to the audio stream to smooth out the jitter. In this chapter, we describe three different control mechanisms that are able to dynamically adapt the audio application to the network conditions so as to minimize the impact of delay jitter (and packet loss). We also present a set of performance results we have gathered from an extensive experimentation with an Internet audio tool we have designed and developed in order to conduct voice-based audio conversations over the Internet.