Audio format

AU converter

An .au file is the Sun audio format, one of the earliest structured audio files, created for Unix workstations. A minimal header followed by sample data, classically 8 kHz mu-law telephony audio, makes .au about as simple as an audio format can get.

Private for supported formats — processed in your browser

Convert supported files

Runs on your device

Drop AU files here

Batch files can each use a different output. Nothing uploads for local conversions.

Working inputs include camera RAW, browser-local audio/video, PDF, CBZ/CBR comics, office documents, ebooks, markup, 3D models, structured text, images, and archives.

Where AU comes from

Sun Microsystems introduced the format for its SPARC workstations in the late 1980s, sharing the underlying design with NeXT's .snd files. It became the web's first common sound format in the Mosaic era and lived on as the only format early Java applets could play.

How AU works

  • A 24-byte header carries a '.snd' magic number, data offset, encoding type, sample rate, and channel count
  • Defined encodings include 8-bit mu-law, linear PCM up to 32 bits, A-law, and floating point
  • The classic mu-law flavor applies logarithmic companding, while the PCM encodings store samples exactly
  • Data can be written with unknown length, which suited live piping between Unix tools

When to use AU

  • Maintaining sound assets in vintage Unix and Java software
  • Converting old voicemail and IVR prompt libraries
  • Opening sound files saved from the 1990s web

Strengths and limitations

Strengths

  • So simple that nearly any audio tool can parse it
  • PCM variants preserve the recording exactly
  • Long-standing support in sox, ffmpeg, and Audacity

Limitations

  • Essentially obsolete outside legacy systems
  • The widespread mu-law flavor is telephone quality
  • No tagging or artwork of any kind

Compatibility

Audio tools such as Audacity, sox, and ffmpeg open .au readily, and Java's sound APIs still accept it. Consumer players, phones, and browsers dropped meaningful support long ago, so conversion to WAV or MP3 is the practical route.

An optional annotation field in the header can hold a short text string; no formal tag structure exists.