MJPEG to OPUS Converter

MJPEG to Opus: Read This Before You Upload

Opus is an excellent target for extracted audio — it is defined by RFC 6716 (IETF, September 2012, Standards Track), it is royalty-free, and it scales from 6 kbit/s speech to 510 kbit/s stereo music. The catch is on the input side: a genuine .mjpeg file is a picture-only elementary stream — a run of standalone JPEG frames with no audio track, no timestamps, and no duration field. If that is what you have, there is literally nothing for the Opus encoder to read. Many files named .mjpeg are not raw streams at all, though, and those extract fine. This page shows you how to tell which one you are holding, and gets you a real Opus file either way.

How to Convert MJPEG to Opus

1. Check Whether Your File Carries Sound: Open it in VLC or QuickTime and listen. If you hear audio, the file is a container (AVI, QuickTime, or Matroska) that merely uses the .mjpeg name, and this page will work on it. If it plays picture-only, it is a true elementary stream and no audio exists inside it.
2. Upload Your MJPEG File: Drag it onto the drop zone or use "Add Files". The server identifies the real container by inspecting the bytes, not by trusting the extension, so a renamed AVI or MOV is still recognised correctly.
3. Set Audio Sample Rate to 48000 HZ: Open Advanced Options and find Audio Sample Rate. Opus only encodes at 8, 12, 16, 24 and 48 kHz, so leave it on ORIGINAL when the source is already one of those, and choose 48000 HZ otherwise. The 44100 HZ entry is offered by the dropdown but Opus cannot accept it.
4. Choose a Rate Mode, Convert and Download: Under File Compression the default is Quality Preset (Very High is the recommended entry); switch to Constant Bitrate if you want an exact number, or Specific file size to aim at a byte budget. Set Audio Channel to MONO for voice, then convert. No sign-up, no watermark.

Walk-through: Is Your .mjpeg a Stream or a Container?

This is the whole decision, and it takes one playback attempt to settle. Motion JPEG is a codec, not a container — the same frames ride inside an AVI, a QuickTime .mov, a Matroska file, or an RTP stream, and only the container versions can carry a companion audio track. Old digital cameras, dashcams and IP cameras almost always wrote a container. What produces a bare .mjpeg is usually a later step: a demux, an ffmpeg -f mjpeg export, or a direct save of an HTTP camera stream. That step keeps the picture and discards the sound.

If you still have the original camera file, convert that instead: AVI to Opus, MOV to Opus or MKV to Opus. If what you want is the picture rather than the sound, Opus is the wrong target entirely — use MJPEG to MP4 for a playable video, or MJPEG to JPG for stills.

Opus Settings That Actually Change the Output

Common Errors and How to Fix Them

• The job fails and no file comes back at all — the source is a true elementary stream. The audio path disables the video stream before encoding, so a picture-only input leaves the encoder with nothing to write, and it stops rather than inventing silence. Go back to the container the stream was demuxed from.
• You picked 44100 HZ and got a 48 kHz file — that is expected, not a fault. Opus is defined only at 8, 12, 16, 24 and 48 kHz. The 44.1 kHz entry is shared UI furniture from other codecs; on an Opus target the request is snapped up to 48000 HZ rather than failing the job.
• Your MKV had two languages and only one came out — extraction takes a single audio track, the one flagged "default" in the container. There is no track picker, so re-mux the file with your preferred track set as default before uploading.
• The Opus file plays in VLC but not somewhere else — Opus playback is solid in Chrome 33+, Firefox 15+ and Edge 14+, but caniuse records Safari as only partially supporting it on the desktop, with full support arriving on iOS in 18.4. Plenty of older DAWs and hardware players never added it at all. Extract to MKV to WAV or MKV to MP3 for those destinations.
• The output is much shorter than the video — check the Trim group; if it is set rather than left on Unchanged, the default Duration is 10 seconds.

When This Doesn't Work

If the original recording is genuinely silent — many security cameras, microscope rigs and machine-vision capture cards record no microphone at all — no converter can recover audio, because none was ever encoded. The same applies to a stream pulled from an MJPEG-over-HTTP camera endpoint: that transport carries frames only. If the footage came from a device that did record sound, the audio is sitting in the file the device wrote; find that original rather than the .mjpeg you exported from it. And if you only need part of a long recording, trim the source with Video Cutter first — the audio-only cutter rejects video files.

Frequently Asked Questions

Why does MJPEG to Opus produce no file instead of a silent one?

Because silence would be a fabrication. The extraction path disables the video stream and asks the encoder to write audio; with a picture-only input there are no audio samples, so the muxer is told the output would contain no streams and the job stops. We fed a 100-frame raw .mjpeg through that exact path and ffmpeg reported "Output file does not contain any stream" and wrote nothing at all — no zero-length file, no silent track. An error is the honest result, and it tells you to go looking for the real source file.

My camera clip definitely had sound — where did it go?

Into the container you no longer have. Motion JPEG is only a way of coding frames; the sound lives in a sibling track inside an AVI, MOV or MKV wrapper. Whatever produced the bare .mjpeg — a demuxer, a frame-server export, a saved camera stream — kept the video track and dropped the rest. Check your camera's memory card or the original download folder for a same-named .avi or .mov.

Will this page work if I just rename my AVI to .mjpeg?

Yes, and that is a legitimate workaround when a page only accepts one extension. The server probes the actual bytes, so an AVI or QuickTime file renamed to .mjpeg is still recognised as that container and its audio track extracts normally. Renaming does not change what is inside a file — it only changes which page will accept the upload.

Is Opus a better target than MP3 for extracted camera audio?

For anything recorded after about 2013, yes. At the low bitrates typical of surveillance and dashcam audio — 24 to 64 kbps — Opus is substantially clearer than MP3, because its speech mode was designed for exactly that range while MP3's was not. MP3 still wins on universal playback: car head units, older phones and legacy editing software all read MP3 without a plugin. If the file has to open anywhere without questions, use MP3.

Why does the sample rate dropdown offer 44100 Hz if Opus can't use it?

The Audio Sample Rate control is shared across every audio target on the site, and 44.1 kHz is valid for most of them. Opus is the exception: RFC 6716 defines the codec at 8, 12, 16, 24 and 48 kHz only, and the reference encoder rejects anything else outright. Internally Opus resamples everything to 48 kHz anyway, so choosing 48000 HZ costs you nothing.

Does converting to Opus lose quality compared to the audio inside the container?

Yes, once — this is a re-encode, not a copy. The source audio is decoded and re-compressed with Opus, so a small amount of detail is discarded even at high bitrates. In practice it rarely matters here: camera and dashcam audio is usually already low-bitrate AAC, PCM at 8 kHz, or ADPCM, and 64 kbps Opus preserves everything those sources actually captured. If you need a bit-faithful copy of the original samples, extract to an uncompressed target such as MKV to WAV instead.

Can I extract just a 30-second section rather than the whole recording?

Yes. Expand Trim in Advanced Options, switch it off Unchanged, then enter a Start time and a Duration. Both accept plain seconds or HH:MM:SS.sss. Trimming happens before encoding, so a 30-second extract from a two-hour file finishes in a fraction of the time and produces a correspondingly small .opus. Note that trimming needs real timestamps, so it works on container sources — a raw elementary stream has none.

What happens to my file after the conversion?

It is uploaded over an encrypted connection, converted on our servers, and deleted automatically after a few hours. Files are never shared, never made public, and never used for anything but your conversion. There is no sign-up, no watermark, and no account required to download the result.