Real- Time High Fidelity Audio for Networked Music Performance
Audio transmission over the Internet is a vital component of modern communications, and digital formats have mostly supplanted physical media for the creation and delivery of cultural content by musicians. From the perspective of artists, access to computers and the Internet democratize the creation and distribution of media, allowing content creators much greater freedom. Just as the MP3 revolution and the growth of Bit Torrent revolutionized the distribution of recorded music, increasingly fast Internet connection speeds lead us toward a future wherein musicians in different geographic areas will be able to record or perform music together in real time, bypassing the need for immediate physical proximity. This concept, known generally as networked music performance, is an increasingly viable means of creative expression for the general public, whereas previous implementations required specialized laboratory environments. As the speed of data connections increases year after year and newer encoding techniques emerge, we approach a tipping point at which the general Internet-using public may collaborate at audio fidelity levels that effectively equal the quality of the original audio source. In the field of communications, Voice over Internet Protocol (VoIP) platforms continue to gain market share as a means of providing efficient and reliable real-time voice services by utilizing the Internet Protocol (I P) over a packet- switched network, and this has led to the development of increasingly mature encoding formats for delivering high-quality sound. As a result, live audio data encoded in real time for VoIP approaches audio fidelity levels acceptable for music without the onerous bandwidth requirements of transmitting uncompressed signals over a network. Until recently, the vast majority of encoding formats available to the general public failed to meet adequate performance requirements for delivery of high fidelity audio data with very low latency, but the emergence of the open source Opus audio format provides a viable mechanism to transmit live music at an acceptable quality of service. In this study, I utilize VoIP network protocols and the Opus audio format to allow two (or potentially more) musicians to perform together at the highest possible level of audio fidelity. My experiments employ a variety of configurations and modifications to Opus and network transmission media with the goal of determining how best to deliver the highest quality audio possible given the speed of a network connection over time. The purpose of my study is to prove the viability of Opus and current technologies for networked music performance by establishing realistic constraints for transmitting audio via current Internet connection speeds, creating a program that utilizes network feedback to adjust the fidelity of encoding to ensure on-time delivery and qualitatively analyzing the fidelity of audio transmitted to the destination.