OGV to AU Converter

OGV to AU: Ogg Video Into Sun's Audio Format

This conversion takes the audio track out of an Ogg video file and writes it as .au — the audio format Sun Microsystems introduced on NeXT and Solaris workstations, and the format Java's AudioClip reached for by default. The picture is discarded; only the sound is kept.

Two things about this page are unusual enough to state before anything else. First, there is no codec dropdown: the Audio Codec group renders with nothing selectable in it, so the encoding is decided server-side. Second, what it decides on is 8-bit G.711 mu-law, the telephone-grade companded encoding that AU has carried since 1992 — not linear PCM. That single fact explains the file size, the sound, and most of the questions below.

AU Format at a Glance

Ogg, Theora and Vorbis: What You Are Extracting From

Vorbis is lossy, and G.711 mu-law is lossy again on top of it. Nothing in this conversion is an archival copy of anything.

How to Convert OGV to AU

1. Upload Your OGV File: Drag the .ogv onto the page or click "Add Files". Several files can be queued and they all run with the same settings. The file must actually contain an audio stream — see the last step for what happens when it does not.
2. Set the Audio Sample Rate: The dropdown lists ORIGINAL first, then 8000, 12000, 16000, 24000, 44100 and 48000 Hz. Leave it on ORIGINAL to keep the Vorbis track's own rate. Choose 8000 Hz only when you are deliberately reproducing AU's classic telephone profile — it throws away everything above roughly 4 kHz, which for music is most of what you were listening to.
3. Set the Audio Channel and Trim (Optional): Audio Channel offers ORIGINAL, MONO and STEREO; Mono exactly halves the output because mu-law writes one byte per sample per channel. Trim starts on Unchanged; switching to the time range gives a Start time that opens at 0 and a Duration that opens at 10, either in seconds or as HH:MM:SS.sss.
4. Convert and Download: Click Convert and save the .au. Files are uploaded over an encrypted connection, processed on our servers, and deleted automatically after a few hours — no sign-up, no watermark. If the .ogv carried no audio track at all, the job fails rather than handing back a silent file: there is nothing to encode, so no output is produced.

Size Is Arithmetic, Not a Setting

There is no File Compression group on this page, and that is correct rather than missing — mu-law has no quality parameter. It always writes exactly one byte per sample per channel, so the output size is fixed by the sample rate, the channel count and the duration, and you can predict it before uploading. We measured a five-second test conversion at both ends of the range:

The 32 extra bytes in each measured figure are the header. Reading it back confirms what the encoder wrote: magic .snd, data offset 32, encoding field 1 for 8-bit mu-law, sample rate 48000, two channels.

Note what that means in the other direction. Mu-law at 48 kHz stereo runs at 768 kbit/s — many times the bitrate of the Vorbis track it came from — while sounding worse, because 8-bit companding is a coarser quantiser than Vorbis used. If your goal is a small file, AU is the wrong format entirely; OGV to MP3 is a fraction of the size and better quality.

Frequently Asked Questions

Why is there no codec dropdown on this page?

Because none of the codec entries are offered for an AU target — the Audio Codec group renders as a heading with nothing under it, and the encoding is chosen server-side. The AU container itself can carry several encodings, and the converter's own table for .au lists mu-law, A-law and 16-bit big-endian PCM as the possibilities, but only the default is reachable from this page. That default is 8-bit G.711 mu-law, which is also what the format's original 1992 profile used.

Is the AU file lossless?

No. G.711 mu-law compands roughly fourteen bits of dynamic range into eight bits per sample, which is a lossy step by definition, and it is applied to audio that Vorbis had already compressed lossily inside the .ogv. So the output is two lossy generations deep. AU can hold linear PCM, and a linear-PCM AU would be lossless relative to the decoded Vorbis, but that encoding is not selectable here.

Why is my AU bigger than the original OGV?

Because mu-law does not compress in any meaningful sense — it writes a fixed byte per sample per channel. At 48 kHz stereo that is 96,000 bytes every second, or 768 kbit/s, while the Vorbis track inside the .ogv may have been running at 96 or 128 kbit/s and the whole file including video may have been smaller than the audio alone comes out here. Dropping Audio Channel to Mono halves it and dropping the sample rate cuts it proportionally, but the format has no compression setting to turn up.

What happens if my OGV has no audio track?

The conversion fails and no file is produced. We confirmed this end to end: an Ogg file containing only a Theora video stream leaves the encoder with nothing to write, so the job errors out rather than handing back a silent .au. Nothing in the pipeline synthesises a silent track to fill the gap. If you are unsure, open the file in VLC and check whether the audio menu offers a track at all before you upload.

Should I set the sample rate to 8000 Hz to match the classic AU profile?

Only if the software receiving the file expects that profile. Dropping a 48 kHz Vorbis track to 8 kHz applies a low-pass at roughly 4 kHz and permanently removes everything above it — fine for a voice recording being fed to a telephony system, ruinous for music. The size saving is real (our test fell from 480,032 bytes to 40,032 for the same five seconds) but you are paying for it with bandwidth you cannot get back. Leave the rate on ORIGINAL unless you have a reason.

Why can't I preview my OGV in a browser to find the part I want?

Because current browsers no longer decode Theora: Chrome disabled it by default in version 120 and Chromium removed it in 123, Firefox disabled it by default in version 126, and Safari never shipped it. VLC and mpv still play .ogv files perfectly, so open the file there, note the timestamps you care about, and type them into the Trim fields before converting.

Can I trim the clip instead of converting all of it?

Yes, and on this page it is the most effective way to keep the file manageable, because size scales exactly linearly with duration. Switch Trim from Unchanged to the time range and enter a Start time and a Duration in seconds or as HH:MM:SS.sss. The controls open at 0 and 10 respectively. Only that window is decoded and written.

Does the AU keep any tags, chapter marks or the Ogg comments?

No. The AU header stops after the channel count — magic, data offset, data size, encoding, sample rate, channels, and then the samples begin. There is an optional annotation area between the header and the audio, but nothing maps the Vorbis comment block into it, so titles, artist tags and any chapter information in the Ogg are dropped. Keep the original .ogv if that metadata matters.

Should I be using AU at all?

Only when something specific asks for it — an old Solaris or NeXT tool, a Java program built around AudioClip, or a telephony pipeline that speaks G.711 natively. For anything you intend to listen to or share, OGV to MP3 is far smaller and better sounding, and OGV to WAV gives you uncompressed little-endian PCM that every editor on every platform opens without argument. AU's virtue is a header you can parse in a few lines, not its audio quality.

How are my files handled while they are on your servers?

The .ogv is uploaded over an encrypted connection, decoded on our servers, and both the upload and the resulting .au are deleted automatically after a few hours. Files are never shared or made public, no watermark is added, and no account is required. Because the work runs server-side, the practical limit on a long recording is upload size and time rather than anything about your own machine.