TS to AU Converter

How to Convert TS to AU Online

1. Upload Your TS File: Drag and drop your .ts transport-stream recording, or click "+ Add Files" to browse. Batch uploads are supported, and Files are uploaded over an encrypted connection, processed on our servers, and deleted automatically after a few hours — no sign-up, no watermark, never shared.
2. Pick Audio Channel and Sample Rate: Defaults are "ORIGINAL" for both, preserving whatever the TS stream contains (typically 48000 Hz stereo for broadcast captures). Drop the sample rate to 8000 Hz mono if you're targeting Unix /dev/audio or μ-law telephony pipelines that expect AU's original headerless lineage.
3. Trim (Optional): Use the Trim control to clip a start offset and duration in seconds — handy for skipping the leading PAT/PMT junk that often precedes the first I-frame in over-the-air .ts captures.
4. Convert and Download: Click "Convert" and grab the resulting .au file. The output uses 8-bit µ-law by default (encoding type 1 in AU's header) — the classic Sun /dev/audio encoding; A-law or 16-bit big-endian PCM (encoding 3) can be selected in the advanced options — no watermark, no sign-up, no file count cap.

Why Convert TS to AU?

MPEG Transport Stream (.ts) is the container the broadcast world uses — DVB, ATSC, IPTV, and HLS segments all ride on it — but it was never designed for casual audio playback. Each TS file packs interleaved 188-byte packets carrying video, audio, PAT/PMT tables, and PCR timestamps, which Unix audio tooling, Java sound APIs, and legacy scientific software can't decode directly. AU (also called SND) is the opposite end of the spectrum: a 24-byte header followed by raw PCM samples, introduced by Sun Microsystems for SunOS and adopted by NeXTSTEP and early Java. Going from TS to AU extracts the audio essence and wraps it in the simplest possible container.

• Java desktop applications — javax.sound.sampled.AudioSystem ships with AU support out of the box (alongside WAV and AIFF), so an .au clip plays from a JAR without any third-party library.
• Unix and Solaris legacy systems — older SPARC, NeXT, and Solaris workstations expect AU at /dev/audio; system sounds and alert chimes on those platforms are AU files by convention.
• Telephony and voice processing — μ-law-encoded AU at 8000 Hz mono mirrors the G.711 codec used in PSTN trunks, making it the natural archive format for IVR prompts and voicemail captures.
• Scientific computing pipelines — MATLAB's audioread, SciPy's scipy.io.wavfile cousins, and many signal-processing tools read AU directly because its big-endian PCM layout is trivial to parse.
• Lecture and broadcast archives — if you've captured an ATSC over-the-air recording or an HLS rebroadcast to .ts, AU strips out the video and gives you an audio-only master you can re-encode later (select 16-bit PCM for a full-fidelity copy — the default µ-law is telephony-grade).
• Sound effects for Java applets and games — the format is still the default in older Java tutorials and lightweight games where adding LAME or libvorbis as a dependency isn't worth it.

TS vs AU — Format Comparison

AU Encoding Quick Guide

Frequently Asked Questions

Why convert TS to AU instead of TS to WAV or TS to MP3?

Use AU when the target system is built around Unix, NeXT, or vintage Java tooling — javax.sound.sampled and Solaris /dev/audio consume AU natively. For everything else, TS to WAV is a better general-purpose lossless choice, and TS to MP3 wins for size when lossless isn't required. AU and WAV are technically near-identical in audio fidelity (both can hold uncompressed PCM); the difference is byte order (AU is big-endian, WAV little-endian) and ecosystem support.

What sample rate and channels does my AU file end up with?

By default we preserve the source's rate and channels — most ATSC and DVB transport streams carry 48000 Hz stereo audio, so your AU comes out 48 kHz / 2 ch, encoded as 8-bit µ-law unless you select A-law or 16-bit big-endian PCM in the advanced options. If you set the Audio Sample Rate dropdown to a specific value (8000, 11025, 22050, 44100, 48000 Hz) or change Audio Channel to mono, those values are baked into the AU header fields and applied during conversion.

Why does my TS file have multiple audio tracks and only one ends up in the AU?

Transport streams often carry multiple language tracks (English + Spanish on US broadcasts, English + audio description, etc.) identified by different PIDs in the PMT. xconvert extracts the primary audio program. If you need a specific language, the cleanest workflow is to demux the desired track in a tool like ffmpeg first, then convert that intermediate file here.

Will the AU file play in modern media players?

VLC, Audacity, foobar2000, and FFmpeg-based players handle AU without issue. Built-in players are spottier: macOS QuickTime reads it, Windows Media Player generally does not, and most mobile apps ignore it because the format is rare outside legacy Unix and Java contexts. If broad compatibility matters more than Sun-era authenticity, convert to WAV or MP3 instead.

How big will the output AU file be compared to my source TS?

For a 60-minute TS broadcast capture at ~6 Mbps total, the audio track alone (typically AC-3 at 384 kbps) is around 170 MB. Converted with the default 8-bit µ-law at 48 kHz stereo, it expands to roughly 345 MB (48,000 samples/sec × 2 channels × 1 byte × 3,600 s); selecting 16-bit PCM doubles that to roughly 690 MB — raw sample data is always larger than the original encoded stream. Use the Specific File Size or Custom Bitrate options if you need to compress, but note that AU's strength is simplicity, not small files; for compressed output, formats like MP3 or AAC are more appropriate.

Can I convert just a portion of a long TS recording?

Yes — use the Trim control to set a start offset (e.g. 60 seconds to skip a pre-roll commercial) and a duration (e.g. 600 seconds for a 10-minute clip). Only the selected window is decoded and written to the AU. This is faster than converting the whole file and trimming afterward, especially for multi-gigabyte over-the-air captures.

Why does my AU file sound corrupted in one tool but fine in another?

Almost always a byte-order mismatch. AU is big-endian by spec, and most modern decoders handle this correctly, but a few legacy Windows utilities assume little-endian PCM and play the file as noise. The fix is either to use a spec-compliant decoder (FFmpeg, Audacity, VLC) or convert to AU to WAV for tools that only speak little-endian.

Is AU still used in 2026, or should I pick a more modern format?

AU is no longer mainstream — most new audio work targets WAV, FLAC, or AAC. It's still relevant in three niches: Java desktop apps that ship .au resources, archival projects preserving original SunOS/NeXT media, and academic signal-processing courses that use AU because its header is the simplest worth teaching. Outside those cases, consider AU to MP3 for distribution or TS to AIFF as a similarly big-endian but better-supported alternative.