The encyclopedia · Engineering & Operations · Technical decision · 1991–1995
MP3 shrank music files by encoding only what ears can hear.
Fraunhofer's perceptual audio codec hid quantization noise under the ear's masking threshold; by 1995 it had a name, .mp3.
Fraunhofer Institute for Integrated Circuits
the move
Uncompressed digital audio was too heavy for the storage and network links of the early 1990s. Fraunhofer's researchers, led by Karlheinz Brandenburg, built the MPEG-1 Layer-3 codec around what psychoacoustics says the ear cannot hear; Brandenburg's paper notes Layer-3 was defined in 1991 and finalized with ISO/IEC 11172 in late 1992.
The codec's mechanism is perceptual encoding: an analysis filter bank splits the signal, a psychoacoustic model computes the masking threshold, and quantization places error where it is inaudible. Brandenburg's 1999 paper, MP3 and AAC Explained, describes how the perceptual model hides the coding noise.
On July 14, 1995 the Fraunhofer team chose .mp3 as the file extension for the technology, the moment, per the institute's own account, the format got its name. Small files plus the internet turned MP3 into the default way music moved.
why it works
- Masking lets the encoder throw away bits no listener could notice.
- Small files made music fit on hard drives and phone lines.
- The codec was standardized inside MPEG-1, so it was open to all implementers.
- The timing matched the consumer internet's takeoff.
what transfers
Measure the constraint your customer actually has, the ear, the eye, the hand, and engineer to it: data the senses cannot perceive is overhead, not fidelity.
what came after
MP3 became the defining file format of the digital music era, powering file-sharing and then legal download services; Fraunhofer's licensing made it one of the most commercially successful audio technologies ever. The perceptual-coding approach it proved carried into AAC and every modern codec.
references
spotted an error? The archive wants to know.