WMA to AU Converter

How to Convert WMA to AU Online

1. Upload Your WMA File: Drag and drop your .wma file onto the upload box, or click "+ Add Files" to select from your computer. Batch uploads are supported — queue several WMAs and convert them in one pass with the same settings.
2. Pick Audio Sample Rate and Audio Channel: Default is Original (the source WMA's rate and channel layout pass through). For classic Sun/NeXT or Java AudioClip compatibility, set Audio Sample Rate to 8000 Hz and Audio Channel to Mono — this matches the historical 8-bit μ-law at 8 kHz mono baseline. For higher-fidelity AU output keep 44100 Hz Stereo.
3. Trim (Optional): Use the Trim control to pick a start time and duration in HH:MM:SS.mmm. Useful for extracting a single voice phrase, alert beep, or short sample from a longer WMA recording before exporting to AU.
4. Convert and Download: Click "Convert". Files are processed server-side, returned as .au with a big-endian header (magic bytes .snd / 0x2e736e64), and download links stay live on our servers — no sign-up, no watermark.

Why Convert WMA to AU?

WMA (Windows Media Audio) was released by Microsoft on August 17, 1999 and stores audio inside an Advanced Systems Format (ASF) container. AU is the older, simpler audio format introduced by Sun Microsystems for NeXTSTEP and Solaris workstations; it has a fixed 24-byte big-endian header followed by raw PCM, μ-law, or A-law samples. Converting WMA to AU strips the ASF container and re-encodes the audio into a format that legacy Unix tools, embedded systems, and old Java code can read without any third-party codec.

• Java AudioClip and Applet.play() compatibility — Java's original sun.audio API only accepted 8-bit μ-law AU at 8000 Hz mono. Older Java applets, AudioClip-based desktop apps, and lecture-archive code from the late 1990s still expect that exact format for system sounds and short voice cues.
• Solaris, SunOS, and NeXTSTEP archives — Sound libraries on SPARC workstations and NeXT cubes shipped as .au. Re-mastering a WMA voice memo into AU is the cleanest way to drop it back into a legacy /usr/share/audio tree or NeXTSTEP .snd resource.
• Telephony and IVR systems — Asterisk, FreeSWITCH, and many legacy IVR stacks accept 8 kHz mono μ-law AU as a native prompt format. The μ-law companding logarithmically packs ~14 bits of dynamic range into 8 bits, which is exactly the G.711 toll-quality envelope used over POTS lines.
• Embedded and educational hardware — μ-law AU is trivially decodable on microcontrollers and DSP boards because the header is fixed-size and samples are raw 8-bit. No FFmpeg, no Media Foundation, no licensing.
• Cross-platform without WMA codec licenses — WMA playback depends on Microsoft's Media Foundation on Windows or projects like FFmpeg/Rockbox elsewhere. AU is plain enough that every Unix sox, play, aplay, and even od can inspect or play it.
• Trimming a long WMA dictation into short clips — Combine the Trim control with mono 8 kHz output to slice a Windows Voice Recorder .wma into IVR-ready prompts under 100 KB each.

WMA vs AU — Format Comparison

AU Encoding & Sample-Rate Quick Guide

The AU header's encoding field is a 32-bit code that tells decoders how to interpret the sample bytes. The big ones:

For sample rate, AU commonly ships at 8000, 11025, 16000, 22050, 32000, 44100, or 48000 Hz. Pick 8000 Hz mono μ-law for IVR/Java-applet parity; pick 44100 Hz stereo 16-bit PCM if you want a lossless, Unix-friendly archive of a music WMA.

Frequently Asked Questions

Will my AU file play in old Java applets and Applet.play()?

Yes, if you set Audio Sample Rate to 8000 Hz and Audio Channel to Mono — that matches the 8-bit μ-law / 8 kHz / mono format the original sun.audio API in pre-1.2 JDKs required. Java 1.2 and later (Java Sound API) accept a wider range of AU variants including 16-bit PCM at higher rates, but if you're targeting genuinely old applet code stick to the 8 kHz μ-law mono profile.

What's the difference between μ-law and A-law inside AU?

Both are 8-bit logarithmic companding schemes from the ITU-T G.711 telephony standard. μ-law (encoding code 1) is used in North America and Japan; A-law (encoding code 27) is used across Europe and most of the rest of the world. They each pack roughly 14 bits of dynamic range into 8 bits per sample, giving you 64 kbps streams at 8 kHz mono. Audibly very similar — pick μ-law for US/Asia telephony interoperability, A-law for European.

Why is my AU file bigger than the WMA source?

WMA is a lossy compressed format (typically 64–192 kbps for voice and music). The default AU output is uncompressed PCM, so a 3-minute 128 kbps WMA (~2.8 MB) decompressed to 16-bit 44.1 kHz stereo PCM lands around 30 MB. To keep AU small, choose 8000 Hz mono — μ-law at that rate is ~64 kbps, comparable to a low-bitrate WMA.

Will converting WMA to AU and back lose quality?

Yes. WMA is already lossy, and re-encoding through μ-law AU adds a second lossy step. If you're round-tripping for archival, render the AU as 16-bit (or 24-bit) linear PCM at 44.1 kHz or higher — that's mathematically lossless from whatever the WMA decoder produced, so no further degradation is added on the AU side.

Can SoX, FFmpeg, and QuickTime open the output?

Yes. AU is one of the oldest formats SoX supports (sox input.au output.wav works out of the box), FFmpeg reads and writes all the common encodings (μ-law, A-law, 8/16/24/32-bit PCM, IEEE float), and macOS QuickTime / Music handles 16-bit PCM AU. On Linux, aplay and paplay both read .au natively.

Is AU still used anywhere in 2026, or is it purely historical?

It survives in a few niches: Java code that still relies on AudioClip for system beeps, telephony lab kits (Asterisk and FreeSWITCH still accept G.711 μ-law .au), Praat phonetics research at the University of Amsterdam (which documents AU as a first-class input), and embedded firmware that wants a header-light raw-PCM container. For new general-purpose audio, WAV or FLAC is the better choice — see WMA to WAV.

Will WMA Pro multichannel survive the conversion?

The audio will be decoded but downmixed to the channel layout you pick (Mono, Stereo, or Original — which preserves up to whatever AU's PCM encoding supports for that channel count). AU has no widely supported 5.1/7.1 surround encoding the way WMA Pro does, so for multichannel masters you're usually better off targeting WAV or FLAC.

Why does my AU file start with .snd when I open it in a hex editor?

That's the magic number. The AU header's first 4 bytes are 0x2e736e64, which is the ASCII string .snd. Everything after it (header size, data size, encoding code, sample rate, channel count) is stored in big-endian byte order, which is why AU is sometimes called the "network byte order" audio format. If your decoder reads little-endian it will see garbage — that's the most common cause of "AU file plays as static."

Can I go the other direction — AU back to WMA or to MP3?

Yes. See AU to MP3 for the modern lossy round trip, or AU to WAV for an uncompressed Windows-friendly archive. For the inverse of this exact page, use WAV to AU if you've already converted your WMA to WAV elsewhere.