Audio Fundamentals
This section is intended to introduce you to the fundamentals of audio terminology and how audio codecs operate. This section introduces the concepts of bitrate, samplerate, channels, and format in regards to audio encoding and configuration. These are common terms used in conjunction with audio encoders, servers and streamers. Before embarking on any excursion into the realm of audio streaming, familiarize yourself with the proper terminology and concepts.
Again, this chapter covers the basics of MP3 audio and how it operates. It also gives beginning users an insight into audio encoding and its fundamental processes.
 | Let me also make it perfectly clear that much of this chapter was borrowed from the O'Reilly and Associates publication (currently out of print); MP3: The Definitive Guide. This chapter was posted on the O'Reilly web page as the book's sample chapter and as such I felt it important enough to include a comprehensive description of the basic of MP3 encoding. Should anyone feel this chapter oversteps basic copyright, after having given proper acknowledgement of my source, I will be happy to remove this section or rewrite it in my own terms. |
MP3 Basics
The past few years have seen much discussion about the technical merits and disadvantages of MP3 audio encoding. To be sure, it has revolutionized the way people listen to music and audio files in general. Where would we be now without the compression and fidelity level accomplished by the MP3 formats? There would be no MP3 or audio players, no free exchange of legal audio files on the Internet, no streaming audio, and the list goes on. How the MP3 standard compresses and decompresses audio, as well as encodes and decodes binary data into compact bitstreams is what makes possible the conversion of formerly large audio files into compact, interchangeable data packets. MP3 files are typically around one-tenth the size of their corresponding uncompressed audio source, much like those found on typical music compact disks.
To be fair, the MP3 format has seen its fair share of nay-sayers and critics. Due to its proprietary nature there are many who prefer a more free encoding model. The imposition of licensing fees by the owners of the MP3 codecs has driven many to seek out alternative audio compression methods. This has drawn out the Ogg Vorbis audio format, a completely open compression methodology. This book discusses the use of Ogg Vorbis in connection with icecast.
MP3 encoding operates on a very basic premise. According to most audio models built into the encoder, audio data is either saved or discarded, depending on what is deemed necessary to the stored model data. The model data is based on mathematical models that represent human hearing patterns. If the data fits, it is kept. If not, it is discarded.
| This study of human hearing is called psychoacoustics and is mentioned throughout this chapter. |
The person doing the encoding specifies how many bits of storage are allotted to each second of music or audio data. This sets the "tolerance" level for the amount of space consumed by the encoding process. As the tolerance level increases, more data is saved and the sound improves. The lower the tolerance level, more data is discarded resulting in poorer quality sound. This compression process is commonly referred to as "lossiness". A second compression run is also made and the data further reduced. This is similar to how the zip compression tool operates on other data files.
The final MP3 file is composed of very short frames. These are often equated to the video picture frames of a filmstrip. Simply transpose the static image into a binary audio data format. Each frame of audio data is preceded by a header, containing information about the stored data. Depending on the encoding, the frames may interact with one another. For example, if one frame has leftover storage space and the next frame does not have enough, they may work together to optimize the end result.
Depending on how the file is built, an MP3 audio clip may contain additional information about itself. This includes the name of the artist(s), the track title, the album name, recording year, the genre, and so on. This information is referred to as ID3 data. When streaming static files this information is very useful for informing listeners as to exactly the type of file to which they are listening.