Convert WebM to OPUS

A WebM made by a browser — a screen recording, a voice message, a MediaRecorder capture — carries Opus audio almost without exception, because that is what WebM was designed around and what every browser can encode. If you have looked inside the file with a media inspector and found Opus there, you were not mistaken.

So this conversion is not solving a codec problem. It is producing a file called `.opus` that holds the codec the source already held, in an Ogg container instead of a Matroska one. Everything worth knowing on this page follows from that, starting with the fact that a job this simple still costs something.

The underlying library can lift compressed frames from one container into another without a decoder ever running. This page does not take that path, because it asks the encoder for a quality band, and requesting a bitrate is precisely what disqualifies the direct copy.

The result is a genuine second generation of lossy compression on audio that did not need one. For a voice note, a meeting recording or an interview it is inaudible at any of the three settings. For music — a rehearsal capture, a loop destined for a game, anything that will be layered or processed later — it is measurable, and the next section is the answer.

FFmpeg does it in one line: `ffmpeg -i input.webm -c:a copy output.opus` writes the same Opus packets into an Ogg container in about the time it takes to copy the file, losing nothing at all. That is the correct tool when fidelity matters and a terminal is available.

MKVToolNix approaches it from the other side. A WebM is a Matroska file, so `mkvextract` can pull the Opus track straight out of it, which is worth knowing if you already have the package installed for other reasons. Both routes beat this page on quality and speed and lose to it on convenience — no install, nothing to learn, and it runs on a phone.

Both targets on this site write Opus into an Ogg container. Choosing OPUS rather than OGG changes the name of what you download and nothing inside it, and a tool that reads one reads the other once the extension stops confusing it.

That is worth stating plainly rather than implying a distinction. The reason both exist is that destinations disagree: some file pickers list `.opus` and not `.ogg`, some archives and importers expect the reverse, and Xiph’s own naming guidance is followed inconsistently across the software that reads these files. Pick whichever your destination accepts and do not look for a difference in the bytes.

Speech tooling, mostly. Voice and messaging pipelines standardised on Opus because it was built for real-time conversation, and a good deal of that software names the codec in its documentation and its file filter. Speech-recognition wrappers and audio-processing scripts often do the same, checking the extension before they check the contents.

That is a different reader from the one converting for a game engine or an open archive, who wants an Ogg because a policy names the container. It is also a different reader from the MP3 and M4A pages, who is solving a playback problem. Here nothing is broken and nobody is stuck — a specification simply named a file type, and this produces one.

Opus is unusually forgiving of being encoded twice, for the same reason it is good in the first place: it was designed to degrade gracefully rather than to fall apart at the edges of its range. Speech at conversational rates survives the round trip with nothing a listener can identify.

What does not survive as cleanly is anything with sustained high-frequency content — cymbals, room reverb, applause, a recorded instrument. The high band is worth its size there, because the first encode already spent detail and a lean second pass has nothing in reserve. If the recording is a master rather than a copy, use the FFmpeg route and skip the question entirely.

This is the practical reason to keep Opus rather than convert to something older. Hours of conversation at the small band occupy a fraction of what the same material would as MP3, and the quality difference at those rates runs in Opus’s favour rather than against it.

The other saving is on the picture. If the WebM is a screen recording rather than a voice note, the video track is discarded outright, which on any normal capture is where nearly all of the bytes were. An hour of shared screen becomes an hour of speech at a size that fits anywhere, which for a meeting archive is often the entire point of the exercise.

It is a tempting shortcut and it fails immediately. WebM is a restricted form of Matroska; an Opus file is an Ogg file. Those are different container specifications with different headers and different framing, and no amount of similarity between the audio inside them changes what a parser reads at byte zero.

What software does with the renamed file varies from a clear error to a silent failure, which is worse. If a tool has rejected your `.webm`, it is rejecting the container, and a real conversion — either this one or the FFmpeg copy — is the only thing that helps.

The decode and the encode both run inside the tab, using the Opus support every browser ships because real-time calls made it mandatory. Nothing carrying the audio is transmitted, there is no account, and up to a hundred files can be dropped at once and downloaded together as a ZIP.

On this pair that matters more than the file sizes suggest. The WebM files that need an Opus extension are voice messages, recorded calls and meeting captures — people talking, often about things that were not meant to travel. A conversion with no server in it is the only version of this job that does not create a copy somewhere else.