Extract Audio from MKV to AU — Free Online Tool
Extract audio from MKV video files and convert it to Sun AU format, encoding the audio stream as 16-bit big-endian PCM (pcm_s16be) — a lossless, uncompressed representation ideal for Unix/Linux environments and legacy audio workflows. The MKV container's audio track (whether AAC, MP3, Opus, Vorbis, or FLAC) is fully decoded and re-encoded into raw PCM, discarding all video data.
to
FFmpeg Command
Copy this command to run the same conversion locally with FFmpeg on your desktop. Download FFmpeg
Drop your MKV file here
or click to browse
Free — no uploads, no signups. Your files never leave your browser.
Settings
Note: Browser-based encoding uses approximate quality targets. For precise CRF compression, copy the FFmpeg command above and run it on your desktop.
Estimated output:
Conversion Complete!
DownloadHow It Works
During this conversion, FFmpeg reads the MKV container and identifies its audio stream, which could be encoded in any of several codecs — commonly AAC or FLAC in modern MKV files. The video stream is completely discarded (not processed at all, saving time and CPU). The audio stream is then fully decoded from its source codec and re-encoded as 16-bit big-endian PCM (pcm_s16be), which is the default and most compatible audio encoding for the AU format. The resulting .au file consists of a minimal Sun AU header followed by raw, uncompressed audio samples. Because pcm_s16be is lossless and uncompressed, there is no generational quality loss in the audio itself — however, if the source MKV audio was lossy (e.g., AAC or MP3), those compression artifacts are already baked in and preserved in the output. AU does not support chapters, subtitles, multiple audio tracks, or metadata beyond the most basic fields, so all of that information from the MKV is dropped.
What Each Flag Does
| Flag | What it does |
|---|---|
ffmpeg
|
Invokes the FFmpeg multimedia processing tool. In the browser-based version of this tool, FFmpeg runs entirely via WebAssembly (FFmpeg.wasm) inside your browser — no data leaves your device. |
-i input.mkv
|
Specifies the input Matroska (.mkv) file, which may contain any combination of video, audio, subtitle, and chapter streams. FFmpeg reads the MKV container and demuxes its streams for processing. |
-vn
|
Disables video output entirely. This tells FFmpeg to ignore all video streams from the MKV and produce an audio-only output — critical here because the AU format cannot contain video data, and skipping video decoding also speeds up processing. |
-c:a pcm_s16be
|
Specifies the audio codec for the output: 16-bit signed big-endian PCM, which is the standard and most compatible audio encoding for the Sun AU format. The MKV's audio (e.g., AAC, MP3, Opus, Vorbis, or FLAC) is fully decoded and re-encoded into this uncompressed PCM representation. |
output.au
|
Defines the output filename with the .au extension, which tells FFmpeg to write a Sun AU container. The resulting file will consist of a minimal AU header followed by raw pcm_s16be audio samples, with no video, subtitles, or chapter data. |
Common Use Cases
- Extracting audio from MKV recordings for ingestion into legacy Unix or Sun Solaris audio pipelines that expect .au files with raw PCM data
- Obtaining an uncompressed PCM audio dump from an MKV file for academic signal processing, audio analysis, or waveform visualization in tools that accept AU/SND format
- Stripping the audio track from an MKV video to supply a raw pcm_s16be audio source to older Java applets or applications that use Java's built-in AU playback support
- Converting MKV audio to AU as an intermediate lossless step before further processing in audio engineering workflows where the AU container's simplicity and raw PCM layout is required
- Archiving the audio component of an MKV file in a format-agnostic, codec-free PCM representation for long-term preservation on Unix-based archival systems
- Generating AU audio samples from MKV source files for use in retro computing or emulation projects that target Sun workstation audio subsystems
Frequently Asked Questions
It depends on the audio codec in your source MKV. If the MKV contains lossless audio (such as FLAC), the conversion to pcm_s16be AU will preserve the full audio quality, since both are lossless representations. If the MKV contains lossy audio (such as AAC, MP3, Opus, or Vorbis), those compression artifacts already exist in the decoded audio and will be present in the AU output — but no additional quality is lost in the conversion itself, because pcm_s16be is uncompressed PCM.
The AU format stores audio as raw, uncompressed 16-bit PCM samples with no compression whatsoever, while MKV typically contains audio encoded with codecs like AAC or FLAC that significantly reduce file size. A stereo audio stream at 44.1 kHz in pcm_s16be consumes roughly 10 MB per minute, compared to under 1 MB per minute for AAC at 128k. The absence of video in the output does reduce total size, but the uncompressed PCM audio will generally be far larger than the compressed audio track from the source MKV.
By default, FFmpeg selects the first (default) audio track from the MKV when no explicit stream mapping is provided. AU format supports only a single audio track, so if your MKV has multiple audio tracks (e.g., different languages or commentary), only one will appear in the output. To select a specific track, you can modify the command with a stream selector such as -map 0:a:1 to pick the second audio stream before the output filename.
The AU format supports several PCM encodings: pcm_s16be (16-bit signed big-endian, the default), pcm_s8 (8-bit signed), pcm_u8 (8-bit unsigned), pcm_alaw (G.711 A-law, a lossy telephony codec), and pcm_mulaw (G.711 mu-law, another lossy telephony codec). To use a different encoding, replace pcm_s16be in the command with your preferred codec, for example: ffmpeg -i input.mkv -vn -c:a pcm_mulaw output.au. Note that pcm_s8 and pcm_u8 provide lower dynamic range and are rarely appropriate for music or high-quality audio.
Yes. On Linux or macOS, you can use a shell loop: for f in *.mkv; do ffmpeg -i "$f" -vn -c:a pcm_s16be "${f%.mkv}.au"; done. On Windows Command Prompt, use: for %f in (*.mkv) do ffmpeg -i "%f" -vn -c:a pcm_s16be "%~nf.au". The browser-based tool processes one file at a time, so the FFmpeg command is especially useful when you need to batch process many files or work with files over 1GB.
No. The AU format is an extremely minimal container — it supports only a simple fixed-size header with basic fields like sample rate, encoding type, and channel count. All MKV-specific metadata (title, language tags, chapters, subtitle tracks, and multiple audio track labels) is completely discarded during conversion. If preserving metadata is important, consider extracting audio to a richer format like FLAC before converting to AU, or documenting the MKV metadata separately.
Technical Notes
The Sun AU format (also known as .snd on some systems) was developed by Sun Microsystems and became a native audio format on Unix workstations and early internet browsers — Java's javax.sound.sampled API has supported AU natively since the beginning, which accounts for much of its continued relevance. The format's header is intentionally minimal: it encodes the data offset, data size, encoding type, sample rate, and channel count, with an optional variable-length annotation field. Unlike MKV, which can contain virtually any combination of codecs and streams, AU is strictly a single-stream audio container. The pcm_s16be encoding chosen here stores samples as 16-bit signed integers in big-endian byte order (most significant byte first), which was native to the SPARC and Motorola 68000 architectures that Sun workstations used — on modern x86/x86-64 systems (which are little-endian), this byte order is non-native but is handled transparently by any compliant AU decoder. Because no audio compression is applied, bitrate is determined solely by sample rate and channel count: a 44.1 kHz stereo stream produces exactly 1,411,200 bits per second. The MKV source may contain audio at any sample rate, and FFmpeg will preserve the original sample rate in the AU output unless explicitly resampled with -ar. One important limitation: AU has no mechanism to store timing offsets or seek tables, making it unsuitable as a distribution format but well-suited as a raw interchange format for audio processing pipelines.