Video to AU Converter

How to Convert Video to AU Online

1. Upload Your Video File: Drag and drop or click "Add Files" to select video files. Works with 35+ formats: MP4, MOV, AVI, MKV, WebM, FLV, WMV, MPEG, 3GP, OGV, MTS, and more. Batch is supported.
2. Pick the AU Audio Codec: Default is PCM mu-law (the classic Sun/NeXT 8-bit telephony codec — the original AU codec). Choose PCM A-law for European telephony (G.711 a-law), PCM 16-bit big-endian for full-fidelity uncompressed AU (the modern AU PCM mode), or PCM 16-bit little-endian if your downstream tool expects it. Set Audio Quality Preset to Lowest/Low/Medium/High/Highest, target a specific file size, or fine-tune with Custom Bitrate (8 kbps to 320 kbps depending on codec).
3. Set Sample Rate, Channels, Trim (Optional): Choose 8000 Hz for telephony-grade mu-law/A-law output (the de facto AU standard), 22050 Hz for vintage Unix workstation audio, or 44100/48000 Hz for full-fidelity PCM. Pick mono (smaller, voice/telephony) or stereo (music). Optionally trim using start time + duration in HH:MM:SS.sss format to extract just a clip.
4. Convert and Download: Click Convert. Files process in your browser session and download individually or as a ZIP — no sign-up, no watermark, no file count limit.

Why Convert Video to AU?

AU (.au, also .snd) is the legacy audio format created by Sun Microsystems in the late 1980s for SunOS, Solaris, and NeXTSTEP. It was the de facto Unix workstation audio format throughout the 1990s and remains the native format for Java's javax.sound.sampled API. Common reasons to extract audio from video and save as AU:

• Legacy Unix and Solaris workstations — SPARC machines, old Solaris boxes, and NeXT/OpenStep systems still in service at universities, research labs, and telecom backbones often expect .au input. mu-law 8 kHz mono is the universally readable AU profile.
• Java audio applications — javax.sound.sampled.AudioSystem natively reads and writes AU. If you're writing a Java app, JavaFX game, or Android system sound, AU is the lowest-friction format with zero external codec dependencies.
• Telephony and VoIP development — μ-law 8 kHz (North America/Japan) and A-law 8 kHz (Europe/rest of world) are the G.711 codecs used in PSTN and SIP/VoIP. Extract video soundtracks as AU mu-law for testing call recordings, IVR prompts, and softswitch integration.
• Academic and scientific archives — DSP courses, speech-processing research datasets, and historical audio corpora (TIMIT, NIST) often distribute as AU. Extract lecture or interview audio in matching format for direct ingestion.
• Bell Labs / NeXT / Solaris audio assets — Restoring or reusing audio with tooling that predates WAV adoption. AU was the format of choice on NeXTSTEP and early web (Sun's .au was the first mainstream web audio format before MP3).
• Embedded systems with minimal codec support — Some legacy embedded firmware ships with only μ-law decoders. AU mu-law mono 8 kHz is ~8 KB/sec — small, simple, and trivially decodable.

AU vs WAV vs MP3 — Format Comparison

AU Codec Quick Guide

μ-law and A-law are 2:1 logarithmic companding codecs — they preserve dynamic range better than linear 8-bit PCM at the same size. They're the original .au codecs and what most legacy software expects. PCM 16-bit BE is the modern AU PCM mode and is bit-faithful to the source video's audio (within the chosen sample rate).

Frequently Asked Questions

Should I pick μ-law, A-law, or PCM for my AU file?

μ-law 8 kHz mono is the canonical AU and what 95% of legacy software expects (SunOS, Solaris, Java's default .au, classic NeXT system sounds). Pick A-law only if your target system specifically needs European G.711 a-law (rare outside telecom). Pick PCM 16-bit BE if you need lossless audio — file size jumps ~10x but the AU is bit-faithful to the source video's decoded audio.

What sample rate should I use?

For telephony, IVR prompts, and Java system sounds: 8000 Hz mono (the AU default — and what javax.sound's default voice format expects). For vintage Unix workstation audio: 22050 Hz. For modern lossless extraction matching the source video: 44100 Hz (music videos) or 48000 Hz (web video, MP4/MOV/WebM defaults). Note that 8 kHz μ-law tops out at ~3.4 kHz audio bandwidth (telephone quality) — fine for speech, not for music.

Will Java's javax.sound read this AU file?

Yes — that's the primary use case. AudioSystem.getAudioInputStream(file) natively reads μ-law, A-law, and PCM 16-bit BE AU files without any extra dependencies. Files produced here use the standard Sun AU header (magic number .snd / 0x2E736E64) that's been the Java audio reference since JDK 1.0.

Why is my AU file so much smaller than the video?

Two reasons stack: the video track (typically 80-95% of the file) is dropped, and μ-law 8-bit at 8 kHz is one of the smallest practical audio encodings (~8 KB/sec, ~480 KB per minute mono). A 100 MB MP4 video commonly becomes a sub-1 MB AU when extracted at telephony settings. For a fuller-fidelity extraction try Video to WAV instead.

Can I extract just a section of the video as AU?

Yes. Use the trim controls to enter start time + duration in seconds (e.g., 45.5) or HH:MM:SS.sss (e.g., 00:01:30.500). Useful for grabbing a specific dialog clip, isolating a single IVR prompt from a recorded call, or extracting one segment of a lecture as a Java system sound.

What input video formats are supported?

35+ video containers: MP4, MOV, AVI, MKV, WebM, FLV, WMV, MPEG, MPG, 3GP, 3G2, M4V, M2TS, MTS, OGV, ASF, AV1, AVCHD, DV, DVR, F4V, HEVC, MJPEG, MXF, RM, RMVB, SWF, TS, VOB, WTV, Xvid, and more. The audio track is decoded then re-encoded into the AU container — input video codec doesn't matter as long as the audio stream is readable.

Why use AU instead of WAV in 2026?

For most modern use cases, WAV is the better default — better tooling support, better compatibility with audio editors. AU is specifically the right choice when (1) you're targeting Java's javax.sound and want zero codec dependencies, (2) you're integrating with legacy Solaris/SunOS/NeXTSTEP systems, (3) you're doing G.711 telephony work where μ-law/A-law is the native format, or (4) you're matching an existing dataset distributed as .au.

Does converting to μ-law lose quality?

Yes — μ-law 8 kHz is lossy by design. It limits audio bandwidth to ~3.4 kHz (telephone quality) and uses 8-bit logarithmic quantization. Speech remains intelligible and natural-sounding; music loses sparkle, cymbals, and stereo imaging (since μ-law is typically mono). For lossless extraction pick PCM 16-bit BE at 44.1/48 kHz, or use Video to FLAC for compressed lossless instead.

Will multi-channel audio (5.1, 7.1) work?

The AU container historically supports mono and stereo only — multi-channel video soundtracks are downmixed to stereo (or mono if you select MONO). For 5.1/7.1 preservation use Video to FLAC or Video to WAV instead.