Anaudio coding format[1] (or sometimesaudio compression format) is a encoded format ofdigital audio, such as indigital television,digital radio and in audio and video files. Examples of audio coding formats includeMP3,AAC,Vorbis,FLAC, andOpus. A specific software or hardware implementation capable ofaudio compression and decompression to/from a specific audio coding format is called anaudio codec; an example of an audio codec isLAME, which is one of several different codecs which implements encoding and decoding audio in theMP3 audio coding format in software.
Some audio coding formats are documented by a detailedtechnical specification document known as anaudio coding specification. Some such specifications are written and approved bystandardization organizations astechnical standards, and are thus known as anaudio coding standard. The term "standard" is also sometimes used forde facto standards as well as formal standards.
Audio content encoded in a particular audio coding format is normally encapsulated within acontainer format. As such, the user normally doesn't have a rawAAC file, but instead has a .m4aaudio file, which is aMPEG-4 Part 14 container containing AAC-encoded audio. The container also containsmetadata such as title and other tags, and perhaps an index for fast seeking.[2] A notable exception isMP3 files, which are raw audio coding without a container format. De facto standards for adding metadata tags such as title and artist to MP3s, such asID3, arehacks which work by appending the tags to the MP3, and then relying on the MP3 player to recognize the chunk as malformed audio coding and therefore skip it. In video files with audio, the encoded audio content is bundled with video (in avideo coding format) inside amultimedia container format.
An audio coding format does not dictate allalgorithms used by acodec implementing the format. An important part of how lossy audio compression works is by removing data in ways humans can't hear, according to apsychoacoustic model; the implementer of an encoder has some freedom of choice in which data to remove (according to their psychoacoustic model).
Alossless audio coding format reduces the total data needed to represent a sound but can be de-coded to its original, uncompressed form. Alossy audio coding format additionally reduces thebit resolution of the sound on top of compression, which results in far less data at the cost of irretrievably lost information.
Transmitted (streamed) audio is most often compressed using lossy audio codecs as the smaller size is far more convenient for distribution. The most widely used audio coding formats areMP3 andAdvanced Audio Coding (AAC), both of which are lossy formats based onmodified discrete cosine transform (MDCT) andperceptual coding algorithms.
Lossless audio coding formats such asFLAC andApple Lossless are sometimes available, though at the cost of larger files.
Uncompressed audio formats, such aspulse-code modulation (PCM, or .wav), are also sometimes used. PCM was the standard format forCompact Disc Digital Audio (CDDA).
In 1950,Bell Labs filed the patent ondifferential pulse-code modulation (DPCM).[3]Adaptive DPCM (ADPCM) was introduced by P. Cummiskey,Nikil S. Jayant andJames L. Flanagan atBell Labs in 1973.[4][5]
Perceptual coding was first used forspeech coding compression, withlinear predictive coding (LPC).[6] Initial concepts for LPC date back to the work ofFumitada Itakura (Nagoya University) and Shuzo Saito (Nippon Telegraph and Telephone) in 1966.[7] During the 1970s,Bishnu S. Atal andManfred R. Schroeder atBell Labs developed a form of LPC calledadaptive predictive coding (APC), a perceptual coding algorithm that exploited the masking properties of the human ear, followed in the early 1980s with thecode-excited linear prediction (CELP) algorithm which achieved a significant compression ratio for its time.[6] Perceptual coding is used by modern audio compression formats such asMP3[6] andAAC.
Discrete cosine transform (DCT), developed byNasir Ahmed, T. Natarajan andK. R. Rao in 1974,[8] provided the basis for themodified discrete cosine transform (MDCT) used by modern audio compression formats such as MP3[9] and AAC. MDCT was proposed by J. P. Princen, A. W. Johnson and A. B. Bradley in 1987,[10] following earlier work by Princen and Bradley in 1986.[11] The MDCT is used by modern audio compression formats such asDolby Digital,[12][13]MP3,[9] andAdvanced Audio Coding (AAC).[14]
| Basic compression algorithm | Audio coding standard | Abbreviation | Introduction | Market share(2023)[15] | Ref | |
|---|---|---|---|---|---|---|
| Production | Streaming | |||||
| Modified discrete cosine transform (MDCT) | Dolby Digital (AC-3) | AC3 | 1991 | 36–54%[n 1] | 37–61%[n 1] | [12][18] |
| Dolby Digital Plus (E-AC-3) | EAC3 | 2004 | [19][20] | |||
| Adaptive Transform Acoustic Coding | ATRAC | 1992 | Unknown | Unknown | [12] | |
| MPEG Layer III | MP3 | 1993 | 15% | 19% | [9][21] | |
| Advanced Audio Coding (MPEG-2 /MPEG-4) | AAC | 1997 | 83% | 87% | [14][12] | |
| Windows Media Audio | WMA | 1999 | Unknown | Unknown | [12] | |
| OggVorbis | Ogg | 2000 | 6% | 4% | [22][12] | |
| Constrained Energy Lapped Transform | CELT | 2011 | — | — | [23] | |
| Opus | Opus | 2012 | 12% | 9% | [24] | |
| Dolby AC-4 | AC4 | 2014 | Unknown | Unknown | [25] | |
| LDAC | LDAC | 2015 | Unknown | Unknown | [26][27] | |
| Adaptive differential pulse-code modulation (ADPCM) | aptX / aptX-HD | aptX | 1989 | Unknown | Unknown | [28] |
| Digital Theater Systems | DTS | 1990 | 8% | 6% | [29][30] | |
| Master Quality Authenticated | MQA | 2014 | Unknown | Unknown | ||
| Sub-band coding (SBC) | MPEG-1 Audio Layer II | MP2 | 1993 | Unknown | Unknown | [31] |
| Musepack | MPC | 1997 | ||||
| SBC | SBC | 2003 | Unknown | Unknown | [32] | |