WEBA to AU Converter

How to Convert WEBA to AU Online

1. Upload Your WEBA File: Drag and drop or click "+ Add Files" to select one or more .weba audio files from your device. Batch conversion is supported — queue multiple recordings in one session.
2. Pick Audio Channel and Sample Rate: Defaults to "Original" for both, which preserves what the WEBA file already has (typically mono Opus at 48 kHz from Chrome's MediaRecorder). For classic Sun/NeXT-style AU output, set Audio Channel to "Mono" and Audio Sample Rate to "8000 Hz" — that matches the original AU specification of 8-bit μ-law at 8 kHz that Java applets and /dev/audio historically expected.
3. Trim (Optional): Leave "Unchanged" to convert the full file, or set a start time and duration to extract a clip. Useful for stripping browser-recorded silence from the start of MediaRecorder captures.
4. Convert and Download: Click "Convert". Each file is decoded from the WebM/Opus container, re-encoded as PCM, and wrapped in the 24-byte AU header. No sign-up, no watermarks, no installs.

Why Convert WEBA to AU?

WEBA is the audio-only flavour of WebM — typically Opus or Vorbis inside a Matroska container, produced by Chrome, Firefox, and Edge when the MediaRecorder API saves microphone input from a web app. AU is the older format Sun Microsystems introduced in the late 1980s for Unix workstations: a fixed 24-byte header (starting with the magic bytes .snd), then raw PCM or μ-law samples. Converting WEBA to AU re-wraps the audio into a format that older toolchains read natively, without the Opus decoder dependency.

• Java audio pipelines — Java's javax.sound.sampled.AudioSystem ships with built-in readers for AIFF, AU, and WAV but not WebM/Opus. If you're feeding clips into a Swing app, a server-side audio worker, or a JUnit-tested sound library, AU is one of three formats Java reads out of the box.
• Solaris, NeXTSTEP, and legacy Unix tools — /dev/audio on Sun workstations, sndplay on NeXT, and many older Tcl/Tk and Motif demo programs expect AU. If you're maintaining vintage Unix software or running emulators (SIMH, ARDI Executor), AU drops in cleanly.
• Telephony test files — μ-law-encoded AU at 8 kHz mono mirrors the G.711 PCM stream used in PSTN, SIP, and Asterisk testbeds. Engineers building IVR systems often prefer AU as a checked-in reference because the encoding is unambiguous and lossless within μ-law's dynamic range.
• Embedded systems and microcontroller audio — AU's header is small and trivially parseable in C; many DSP demo kits ship example code that reads .au. WAV's RIFF-chunked layout is harder to parse on a tight memory budget.
• Academic and reference corpora — Datasets from CMU, Berkeley, and Princeton's intro CS courses (e.g., StdAudio.java) distribute samples as AU because the format predates Windows-era audio and remains license-free.
• Archival of voice memos — Saving a Chrome-recorded note as AU PCM gives you a self-describing, uncompressed copy that any future toolchain can decode, without depending on the Opus codec library staying installed.

WEBA vs AU — Format Comparison

Sample Rate & Channel Quick Guide

Frequently Asked Questions

Why is my AU file so much larger than the WEBA original?

WEBA from a browser is almost always Opus-compressed at 96–128 kbps, while AU defaults to uncompressed PCM. A 1-minute 48 kHz stereo Opus track is roughly 1 MB; the same minute as 16-bit PCM AU is about 11 MB. If you need a smaller AU, pick "Mono" + "8000 Hz" — μ-law at those settings runs roughly 480 KB per minute, close to the original AU spec.

Will Java's AudioSystem.getAudioInputStream() actually read the output?

Yes, for PCM-encoded AU. Java's built-in com.sun.media.sound providers handle 8/16/24/32-bit PCM and μ-law/A-law AU files — those are the encodings AudioSystem has supported since JDK 1.0. If a downstream pipeline ever throws UnsupportedAudioFileException, check that the encoding is PCM_SIGNED or ULAW and not one of AU's exotic 28 encoding IDs (most converters, including this one, default to PCM_S16BE which Java reads natively).

Does the output start with the .snd magic bytes?

Yes. Every valid AU file begins with the four-byte signature 0x2E 73 6E 64 (ASCII .snd), followed by a 24-byte header containing data offset, data size, encoding ID, sample rate, and channel count, all in big-endian order. That's a hard requirement of the format — any tool that produces AU without those bytes is producing something else.

My WEBA came from Chrome's MediaRecorder at 48 kHz. Should I resample to 8 kHz?

Only if your downstream tool needs it. Resampling 48 kHz to 8 kHz is irreversible — you lose everything above 4 kHz of audio bandwidth (which kills sibilance and most musical content but is fine for speech). Keep "Original" for music or wide-band voice; pick 8000 Hz only when you're targeting Java 1.1 compatibility, telephony testbeds, or a strict storage budget.

Why does my AU player report a duration of zero or play forever?

Older AU readers expect the data-size field in the header to be either the exact byte count or 0xFFFFFFFF (the "size unknown" sentinel). Some converters write 0 instead, which legacy players interpret as "no data" or "stream forever". Files produced here use the correct exact byte count, so duration reports correctly in VLC, Audacity, and Java's Clip API.

Can I keep the WEBA's Opus compression inside an AU container?

No. AU's 28 documented encodings cover PCM (8/16/24/32-bit), μ-law, A-law, IEEE float, and a handful of older ADPCM variants — Opus is not one of them. The conversion must decode the Opus stream and re-encode it as PCM (or μ-law/A-law). If you specifically need lossless re-packaging, see WEBA to FLAC instead, which keeps file size down while preserving exact samples.

Is there a file-size cap?

xconvert handles audio uploads well past what browser MediaRecorder sessions typically produce (most recordings are a few MB). For long-form recordings, batch them into shorter clips using the Trim controls rather than uploading multi-hour single files, which gives more reliable conversion performance.

What's the difference between this and WEBA to WAV?

WAV uses a RIFF chunked container with little-endian samples; AU uses a flat header with big-endian samples. Audio fidelity is identical when both are 16-bit PCM at the same sample rate. Pick WAV for Windows-first workflows, DAWs (Pro Tools, Reaper, Audition), and tools that prefer RIFF. Pick AU for Java, Unix shell pipelines, and anything that grew up on Sun or NeXT hardware. For lossless compression that's smaller than either, use WEBA to FLAC.

Can I batch-convert and then compress the AU files later?

Yes. After converting WEBA to AU, you can shrink the result with the audio compressor — useful if you produced PCM 16-bit AU for archival and want a smaller μ-law copy for distribution. Or convert directly to a compressed audio format like WEBA to MP3 in one step if AU isn't required downstream.