The MP3 Format
MP3 is an acronym for MPEG ( Moving Picture Experts Group) 1, Layer 3. It was developed by Fraunhofer AG as a means to transmit streaming audio from one place to another without requiring the connection speed ( bandwidth) other formats required. It does this by compressing the audio information down to just those sounds that the human ear can hear. At the time, other audio formats recorded every sound from the original work, whether we could hear it or not. There are many different formats of MPEG streaming audio and video. MPEG 1, Layer 3 is only one format that specifically addresses audio.
MP3 wasn't originally designed as a file format. As time went on it became apparent that the advantages to the MP3 compressed format could be taken advantage of in a file format so the technology advanced to incorporate streaming the audio to a file.
On the average, MP3 audio takes up 1/10th the space or speed of WAV format audio with little discernible loss of quality.
MP3 audio is divided up into frames. Each frame represents one small period of time, on the average about 26 milliseconds (26 one-thousandths of a second). When an MP3 stream or file is created the creator specifies three parameters, bitrate, frequency and mode.
Bitrate is the speed at which the audio can be sent at. This will impact how much audio information can be stored in a particular frame. If the bitrate is fast (128kbps or greater) then more information can be stored in each frame, thus yielding higher quality audio. Slower bitrates have to throw away more music information to account for less width of the stream.
Frequency is how often the music information is sampled to create the resultant audio. 44,100 Hz is used for high quality audio. The higher the frequency, the more audio information that can be stored.
Mode is whether the audio is being sent in one channel (mono), two channels (stereo) or two channels with only the differences between the two channels being sent (joint stereo). Joint stereo was developed to reduce the amount of bandwidth required for a stream because in stereo often the music is very closely related on both the channels. So joint stereo takes up less bandwidth than stereo, but more than mono.
VBR (Variable Bit Rate) encoding is something that has recently arrived on the scene. VBR really cuts down on the amount of bandwidth required. Each frame is recorded with a different bitrate depending on how complicated the music is at that particular point in time. If there's silence on the audio track, the bitrate can be reduced to almost nothing for that frame. If there is a lot of music (like a symphony orchestra at the peak of the piece), then the bitrate can be increased way above the average to give high fidelity during that frame's small period of time. The computer program that converts the analog audio to digital will be set up to yield a certain average bitrate which is used to compute the actual bitrate of each and every frame. VBR requires lots of processing overhead on the listeners computer to decode but today's PCs handle it just fine, in most cases. VBR isn't widely accepted just yet but it is gaining popularity.
The developer of MP3 created it so that audio frames could be picked out of all kinds of garbage data. This is in case the stream was scrambled for some reason. See Technical for more information.
|