There are an astonishing number of audio file formats in use. I hope that the following list of common audio formats includes most of the major ones in use as of 2005: Possibly some of the rates below are wrong by a factor of two as it is often unclear whether published rates (which are not especially easy to find) are for mono or stereo. Some are likely wrong anyway. To the extent that I could do so with often fuzzy data, I have attempted to quote data rates for typical material -- which implies stereo (or more for AC3 and DTS).
Please keep in mind that this is only a subset of the audio formats in use.
AAC -- Advanced Audio Coding. AAC is an enhanced MP3 using somewhat modified compression algorithms to yield better fidelity at the same bit rate. It is not backward compatible with MP3. Data rates are similar to, but slightly better than MP3. Say 90-200 Kbps.
AC3 -- a common file format for music on DVDs. Also used for HDTV audio. Supports surround sound -- five channel plus low frequency effects. This is a lossy compression format run at various speeds. Common bandwidths for five channel sound are 384-448 Kbps, but 640 Kbps is often used in theaters. AC3 can also be used with fewer channels. 2channel stereo typcially is recorded at 192Kbps. AC3 is also know as Digital Dolby
AIF -- Audio Interchange File. The Apple equivalent of the PC's WAV files. Uncompressed or only slightly compressed data with variable format and recording rates. CD quality stereo requires around 700KBits per second
ASF -- Advanced System Format. A file format used as a wrapper for Microsoft WMA files.
ATRAC -- A proprietary perceptually compressed format used by Sony for compact CDs and other consumer devices.
CD ("RAW")-- An uncompressed format used for music CDs. 1.4MBit per second. Typically is not saved to or played from PC files -- thereby explaining the lack of an extension.
Digital Dolby -- AC3 with a new name. (There was something wrong with the old name?)
DTS - Digital Theater Systems. A multichannel surround system first used in theaters for the movie Jurassic Park. DTS is reprted to provide better quality than Digital Dolby at the cost of higher bit rates. 1500KBps.
FLAC -- Free Lossless Audio Codec -- An open source lossless compression technology using linear prediction. Data rates are variable depending on the data. Typical might be around 50%. i.e. 700Kbps. Then again, it might not.
Midi -- A technology that generates music by emulating an assortment of instruments being played according to more or less standard music notation. Midi is quite compact, but can't easily emulate arbitrary sounds like singing. Data rates for instrumental music may be as low as 2.2KBits per second.
Monkey's Audio -- A lossless compressor that compresses between 2:1 and 4:1 depending on the material. 350-700Kbps.
MPC -- MusePack audio. A lossy compression approach similar to MP3 and Ogg Vorbis. It claims to be a non-proprietary optimization of MP1/MP2. Data rates are similar to MP3. MPC is not as widely used as MP3 or Ogg Vorbis. It is claimed to deliver superior sound at rates above 160Kbits at the cost of relatively poorer audio at lower data rates.
MPEG, MP3 -- A family of audio technologies that use Perceptual Compression to reduce sound to reduce file size. There are (I think) two specifications MPEG and MPEG2 and -- within MPEG1,several formats--MP1, MP2, MP3, and AAC . Compression is variable, but a factor of 8 to 12 relative to CD is generally quoted. Rates of 95 to 170 KBit per second are typical. 128KBit per second is often used for "near CD Quality"
Ogg Vorbis -- Intended as an open source alternative to MP3 which uses proprietary algorithms. Like MPEG Perceptual compression is used. Bit rate is variable, but compression is reputed to be similar to MP3 with slightly better fidelity at the same compression ratio.
PAC -- Perceptual Audio Coding -- An encoder designed by Lucent technology specifically for broadcast applications. Uses perceptual compression with data rates selectable in 16Kbps increments. Used in the XM Radio Service. CD quality audio is said to be available from 96Kbps PAC.
RA and RM -- Real Audio -- A highly compressed lossy format widely used for streaming audio. The format is proprietary(?). Minimum data rates appear to be about 14.4Kbits for voice, 28.8Kbits for music.
SHN -- An implementation of a lossless compression algorithm that is free for non-commercial use. Compression varies with the material and seems to be around 50-60%. Call it 600Kbps.
VOX -- A highly compressed lossy format optimized for voice data. Compression ratios of 53:1 relative to CD are claimed. That would work out to around 26.4Kbits per second.
Wave -- An uncompressed or only slightly compressed storage format used to store data on (windows) computers -- usually with a .WAV extension. It has variable format and recording rates within a framework that allows the format to be determined. CD quality stereo requires around 700KBits per second. Microsoft uses a variety of different (incompatible of course) techniques for representing audio within .WAV files.
WavPack -- a compression program that operates on WAV files using either lossless or lossy compression to generate .WVC compressed files. Can generate stand alone .EXE files for audio playback on PCs. Said to be less effective than MP3. Around 256Kbps.
WMA -- A collection of Microsoft developed compressed audio formats offering various combinations of compression and quality. Bit rates vary from around 600Kbps for lossless down to around 64KBps. See https://www.loc.gov/preservation/digital/formats/fdd/fdd000027.shtml which may or may not be a complete summary of WMA formats.
Return To Index Copyright 1994-2017 by Donald Kenney.