Au file format

The Au file format is a simple audio file format introduced by Sun Microsystems, primarily utilized on NeXT systems and early web pages, and characterized by its straightforward header structure and support for various audio encodings.

History and Usage

The Au format was introduced by Sun Microsystems in the late 1980s for use on their SPARC workstations. It later became the standard audio format for the NeXTSTEP operating system and, consequently, early versions of Mac OS X. During the early days of the World Wide Web, the Au format was widely used for embedding short audio clips in web pages. This popularity was largely due to its support for μ-law (mu-law) encoding, which provided acceptable voice quality at a low bitrate (8-bit, 8000 Hz mono), making it suitable for the limited bandwidth of early internet connections. Over time, as bandwidth increased and more advanced compression algorithms were developed, the Au format was largely superseded by formats such as WAV, AIFF, MP3, and Ogg Vorbis. Despite its decline in mainstream use, it remains supported by numerous digital audio workstations, audio editors, and command-line utilities.

File Structure

An Au file consists of three main components: a file header, the audio data, and an optional annotation chunk.

Header

The header is typically 24 bytes long and contains the following fields, all stored as 32-bit unsigned integers:

• Magic Number: The file begins with the magic number .snd (hexadecimal 0x2e736e64), which identifies the file as an Au audio file.
• Data Offset: This indicates the byte offset where the actual audio data begins. It is equal to the size of the header plus the size of any optional annotation. The minimum value is usually 24.
• Data Size: This specifies the size of the audio data in bytes. If the size is unknown (such as when streaming), this field is set to 0xFFFFFFFF.
• Encoding Format: An integer that defines the audio encoding used in the data chunk (e.g., 1 for 8-bit ISDN μ-law, 3 for 16-bit linear PCM).
• Sample Rate: The number of samples per second (e.g., 8000, 22050, 44100).
• Channels: The number of interleaved audio channels (e.g., 1 for monaural, 2 for stereo).

Audio Data

The audio data immediately follows the header (and the optional annotation). The byte order of the audio data is traditionally big-endian, reflecting the big-endian architecture of the original Sun SPARC workstations. While some modern implementations may support little-endian variants, big-endian remains the standard for strict compliance.

Annotation

Between the header and the audio data, there may be an optional, null-terminated text string. This annotation can contain arbitrary information, such as the title of the recording, the author, or recording conditions. The length of this string must be a multiple of four bytes, padded with null characters if necessary, to maintain 32-bit alignment for the audio data.

Encoding Formats

The Au format supports a wide variety of audio encoding schemes, identified by the encoding format field in the header. Some of the most common encodings include:

• μ-law (G.711): 8-bit logarithmic compression, historically the most common format for web audio and telephony.
• Linear PCM: Uncompressed pulse-code modulation in 8-bit, 16-bit, 24-bit, and 32-bit integer formats.
• IEEE Float: 32-bit and 64-bit floating-point PCM, used for high-precision audio processing.
• A-law (G.711): 8-bit logarithmic compression, similar to μ-law but primarily used in European telecommunications.
• ADPCM: Adaptive differential pulse-code modulation, including CCITT G.721 and G.723 variants, which offer better compression ratios than standard PCM.

Technical Specifications and MIME Types

Because the Au format was heavily used in early web browsers, specific Multipurpose Internet Mail Extensions (MIME) types were assigned to it. The most historically significant MIME type is audio/basic, which was strictly defined by RFC 2046 to represent 8-bit μ-law encoded audio at an 8000 Hz sample rate in a single channel. For other encodings and variations within the Au container, the MIME type audio/x-au or audio/au is typically used.

Legacy and Modern Relevance

Today, the Au file format is considered a legacy format. Modern web standards and operating systems favor more versatile and efficiently compressed formats like MP3, AAC, and FLAC. However, the Au format's structural simplicity makes it highly reliable for raw audio data exchange in academic research, telecommunications, and low-level audio programming. It is fully supported by modern open-source audio processing tools such as SoX (Sound eXchange), FFmpeg, and Audacity, ensuring that legacy Au files can still be easily played, converted, and analyzed.