9.1 Introduction

Media streaming is the delivery of content as a continuous boundless stream of bits. Media streams could be used for transmission of audio, video, and data (e.g., captioning text, stock tickers). It is possible to categorize issues related to media streaming in the layered structure depicted in Figure 9.1. The two transport options are MPEG-based and IP-based. Above the transport layer is the media coding layer, which addresses issues of decoding and encoding content, as well as issues related to media format. Above media coding is the synchronization layer, concerned with coordinating the various media types (e.g., audio, video, and text) to create a high quality viewer experience. Finally, the application integration layer provides integration API (e.g., JMF) and markup formats (e.g., SMIL).

Figure 9.1. Conceptual media-streaming layers .

Streaming has the aspects of capture, transmission, and rendering. Content capture is the process of converting a physical signal, such as sound or image, into a sequence of bits that is an approximate electronic representation of that signal. Content transmission is the process of repeatedly modifying the format of the bits (e.g., through compression, encapsulation and modulation) to deliver the information along a path that ultimately ends at a content rendering device; this path may include multiple multiplexing, re-multiplexing, and de-multiplexing stages, as well as storage in an archive or delayed retransmission. Typically, compression changes the representation of the signal from the input space into the frequency space, along the time or space dimensions, resulting in a more compact representation. Content rendering is the process of converting the representation delivered back into a physical signal, such as sound or image, which approximates the original signal captured.