MPEG-1 or MPEG-2 Audio Layer III is also known as MP3. This is format of audio coding for audio digitally. It uses form of data lossy compression. This is audio format commonly for consumer’s audio storage or streaming. It has de facto standards of audio digital compression for transferring and music playback on many audio digital players. Use of compression lossy has been designed for reducing greatly data amounts needed for representation of audio recordings which sounds still like reproduction faithful of uncompressed original audio for many listeners. MP3 files are created by usage of 128 kbit/s setting. This result in file about 1/11 being size of CD file. This is created through original audio sources (44100 samples per second* 16 bits per sample*2 channels = 1411200 bit/s). MP3 being compressed at 128 kbit/s: 128000 bit/s [1 k = 1000 not 1024, since it is bit rate]. Ratio: 1411200/128000 = 11.025. MP3 file is constructed even at lower and higher bit rates with lower and higher quality resulting.
Designs: File Structures: MP3 files are created of MP3 frames. This consists of data block and header. Frame sequence is known as Elementary Stream. As because of “byte reservoir”, frames have not been items individually. It usually cannot be extracted on boundaries of arbitrary frames. MP3 Data Block contains audio information compressed in terms of amplitudes and frequencies. MP3 Header consists of Sync Word. This is used for identifying valid frame’s beginning. It is followed by bits indicating MPEG standard. 2 bits indicates usage of layer 3. Therefore it is called MPEG-1 Audio Layer 3 or MP3. Then values differ whilst depending on MP3 files. ISO/IEC 11172-3 defines value ranges for all sections of header together with header’s specification. Most MP3 files these days contain ID3 metadata that follows or precedes MP3 frames. A data stream contains checksum optional but checksum protects only header data and not audio data. Joint Stereo is conducted on only frame-to-frame basis. Decoding and encoding: MPEG-1 standard do not include specification precisely for MP3 encoders. But it provides examples of rate loop, psychoacoustic models and like non-normative parts of standard being original. Presently, implementations suggested are till date. Standard implementers was devising suppositions of own suitable algorithms for part removals of information from inputs of audios. Result is that, there are lots of MP3 encoders present. Each produce files of quality differing. Comparisons are present widely. So it is easy for users prospective of encoders for researching best choices. Encoders proficient at encoding at high bit rates like LAME is necessarily not as good at low bit rates. Opposite of encoding is decoding. Quality factor must be seen.