MP3 & Compression
1993 onwardsA CD track holds far more data than a 1990s internet connection could move in any reasonable time. Compression solved that not by storing the sound more cleverly, but by deciding which parts of it a listener would never notice were gone.
Origin
MPEG-1, the standard that includes Layer 3 audio coding, was published by ISO in 1993. The three audio layers were developed collaboratively, with Fraunhofer IIS and the University of Erlangen among the contributing institutions.
The familiar name came later and more casually than the technology: the ".mp3" file extension was settled by an internal email poll at Fraunhofer IIS on 14 July 1995.
How it works
The technique is called perceptual coding, and it is built on how hearing actually behaves rather than on the mathematics of the waveform.
A loud sound masks a quieter one at a nearby frequency — you genuinely cannot hear the quieter one while the louder is present. The encoder finds those masked components and simply does not store them. It also spends fewer bits on frequencies where the ear is less sensitive.
This is lossy: the discarded information is gone permanently. Encoding an MP3 back to a CD does not restore it, and re-encoding an MP3 repeatedly compounds the damage — which is why bitrate arguments were so heated and why archives keep lossless masters.
What it changed
A song became a file: something you could email, put on a hard disc, carry a thousand of, and copy perfectly and instantly at no cost. Distribution stopped being a physical problem, and with it went the industry's main point of control.
The consequences arrived faster than anyone's business model. File sharing, the collapse of the album as a purchase unit, the legal download store, and eventually streaming are all downstream of making the file small.
Then and now
Newer codecs — AAC and others — do the same job more efficiently, and bandwidth is now cheap enough that lossless streaming is ordinary. The specific format matters less than it did.
The idea, though, is now everywhere. Every streamed song you hear is compressed by some descendant of this reasoning about what the ear will not miss.
Names that mattered
Listed because of what they did, not as recommendations. Velorex Music did not make any of these.
Did you know?
- The name was decided by an internal poll, not a committee — on 14 July 1995.
- MPEG-1 defined three audio layers of increasing complexity; Layer 3 is the most efficient and the hardest to encode.
- Perceptual coding works because of masking: a loud tone genuinely renders a nearby quiet one inaudible.
- Lossy compression is one-way. Nothing downstream can recover what the encoder discarded.
Remember this?
For a certain generation the memory is a progress bar: a single song arriving over a dial-up connection across several minutes, the filename usually wrong, the artist often misattributed, and the last few seconds sometimes missing.
Sources & further reading
- Fraunhofer IIS — 30 Years of .mp3
- Library of Congress — MP3 (MPEG Layer III Audio Encoding) format description
- Fraunhofer — MP3 and AAC explained (AES)
Written from these sources, not copied from them. External links open in a new tab.