Audio file formats – Understanding psychoacoustic coding


Audio compression with loss of information (lossy) is based on specialized algorithms to determine what changes simplify the representation of the sound while restoring the better sound impression.

It reduces the file size by eliminating uncollected or less essential shades to the content. Elimination ultimately creates a file in a high quality format from a compressed file.

The best-known format is MPEG-1/ 2 Audio Layer 3, with the suffix is .Mp3. This format offers a very decent sound quality for a rate of 128 kbit/s. It is this format that has been heavily used to transfer music via the Internet in the late 1990s.

Shortly,afterwards players with a rewritable memory and can directly read this format appeared.

In the 2000s, new formats were proposed. Given the progress of the algorithms, they largely outperform MP3 in terms of quality at equal speed, and can achieve higher grades.

In addition, some are less restrictive than the MP3 vis-avis rights of use (Ogg is a free format).

But MP3 is the most used because the continuous arrival of new formats, providing a relatively small advantage over previous ones, does not establish a better option than MP3, readable by all standard players.

For the same compression format, there is no single way of encoding, because each algorithm seeks the best way to represent its original language following compression.

In particular, the codec MP3 has made very significant progress since the beginning of the use of this format.

It typically allows a gain of factor of 10. This has made it possible for their exchange on the Internet, often illegally.


Formalized in 1992, AC3 compression allows up to six independent audio channels with a sampling rate of 32, 44.1 or 48 kHz and with a transfer rate of from 32 to 640 kbit/s. Dolby Digital uses this coding scheme, which is why it is often referred to by that name.


MP3 is short for MPEG-1/ 2 Audio Layer 3. It is dedicated to applications requiring low flow rates (128 kbit/s) with very rapid accession of the Internet world to this compression format.

The compression ratio is usually 1 to 10 (1:10) ( 1:4 to 1:12). Very fast encoding. Significant royalties are payable to exploit the MP3 license.

Compression Type : constant or variable ( VBR) .


Mp3PRO, a collaboration between Thomson Multimedia and Fraunhofer Institute combines the MP3 algorithm and system to improve the quality of compressed files called (in ) Spectral Bandwidth Replication for SBR.

This format was released in late 2001, a MP3pro offers 64 kbit/s quality equivalent to that of an MP3 128 kbit/s.

The file suffix is created . Mp3

Ogg Vorbis

Vorbis is different from MP3, WMA and AAC on algorithm. It segments the audio sources in successive packets, the compression algorithm acts initially on each packet independently.

This allows it to have very few weaknesses on certain frequencies and keep the same quality regardless of the type of music.

VQF or TwinVQ

The TwinVQ (Transform -domain Weighted Interleave Vector Quantization) format was developed by NTT Cyber Space Laboratories and supported by Yamaha.

In the same spirit as MP3, it compresses more with better quality. Regrettably coding duration is a bit too long, about 10 times slower than MP3.

Moreover, it was introduced much later, and is distributed under a very restrictive license, it has had few followers and is more or less abandoned.

The file suffix is .Vqf , .VQL or .Vqe.


The WMA (Windows Media Audio) format , created by Microsoft based on recommendations MPEG-4 in 1999 , is used by the Windows Media Player software.

This format is linked to a precise management of copyright ( Digital Rights Management, DRM) which allows to define such a limited lifespan for files or prohibit the possibility of burning.

There are several versions of the codec ( wma7.1 , WMA9 , WMA Pro).

The file suffix is created . Wma


The AU format is fairly common with Unix and Linux. The sampling frequency is between 1 kHz and 200 kHz . But applications are mostly read audio rendering three sampling frequencies: 8012.821 ( entered codec) , 22050 and 44100 hertz.

The file suffix is created . At

Resolutions 8, 16 , 20, 24 and 32-bit (floating) are accepted.


ASF Advanced Streaming Format, is a container format used for Microsoft audio and video streaming.


(en) AA Audible, is a format used by Apple for audio -books.


AAC (Advanced Audio Coding), is an extension of MPEG -2 and was upgraded to MPEG -4, MPEG-4 Version 2 and MPEG-4 Version 3. It was recognized in late April 1997.

The file suffix is created . Aac , . Mp4, . M4a

Apple and AAC

Apple has chosen AAC as the preferred codec, it is found in its iPod and iTunes software. For music sales online iTunes Music Store, the AAC standard does not offer a system digital rights management ( DRM ), Apple has developed its own system, called FairPlay.

It is playable on Mac OS and Windows, only with the iTunes software.

The fact that AAC is the only compression format more efficient than MP3, which is supported by iPod has greatly contributed to its popularity losses.

However inadequately it for itself as the successor of MP3, other formats being widely cited set equal in performance.

The ATRAC (Adaptive Transform Acoustic Coding) is a psychoacoustic audio compression technique (there is a lossless option) developed by Sony in 1992 for his MiniDisc.

This format has undergone several changes: ATRAC3, ATRAC3plus (colloquially written ATRAC3 +) and ATRAC Advanced Lossless successive respectively in 1999, 2002 and 2006.

Audio file formats – Understanding psychoacoustic coding

About the Author