Toward bandwidth-efficient and error-robust audio streaming over lossy packet networks

Bandwidth efficiency and error robustness are essential issues for different multimedia streaming applications. This paper presents strategies for high-quality audio streaming based on fragmenting perceptually coded audio frames and shuffling the data components among multiple packets for transportation. This is done to increase robustness against packet loss. We also address the delivery of audio data consisting of components with different proportional priorities. Our approach is rationalized with streaming tests using the MPEG AAC audio codec in a simulated network environment and formal listening tests to evaluate the resulting audio output. According to the results, the proposed schemes improve audio quality significantly with reasonable increase to network resource utilization compared to traditional error robustness measures.