Audio perception is highly susceptible to disturbance in temporal quality. In the event of congestion, audio datagrams suffer packet loss, delay and jitter. Compressed audio appears as one of the most delicate traffic-type to handle on the Internet. The recent advances in auditory perception promise opportunities where a perceptually clever adaptive audio system can respond to impending breakdown and support near flawless sound. However, the major problem is receiving fast feedback from the current Internet. Recently proposed TCP Interactive (iTCP) seems to offer some interesting opportunity to perceptual audio. This is an operational state equivalent to the conventional TCP except that applications can optionally subscribe and receive selected local end-point protocol events in real-time.
In this thesis, we implemented a novel symbiotic perceptual audio streaming mechanism, that receives fast feedback from iTCP about congestion and then responds appropriately. This mechanism combines the quantization technique to accurately represent the audio signals without distortion. The adaptive system reduces the encoding target bit rate when congestion is detected which invariably reduces the delay and jitter faced by audio traffic in the network during congestion. We have tested it over the Internet. In this thesis, we highlight the performance of this system and report on the dramatic improvements in time-bounded streaming audio we observed.