WebM is a Matroska subset built for browsers, holding VP8 or VP9 video, and the slowest thing to encode on this machine by a factor of ten. Opus is the lossy encoding voice calls and browsers settled on, better than MP3 at the same bitrate and unreadable by most hardware players. ffmpeg sits between the two, and this page says exactly what happens to a WEBM file on the way to becoming an OPUS one.
What runs when a WEBM file becomes OPUS
ffmpeg held to a single thread, on a machine we rent and watch. Your WEBM file is uploaded once, ffmpeg runs once, the OPUS comes back, and neither file is kept. WEBM to OPUS is one of the 3649 pairs that engine was probed on with a real file, which is why it has a page here and why the pairs the probe could not prove do not.
What survives from the WEBM into the OPUS
The sound, and nothing but the sound. ffmpeg passes -vn on every audio target, so the WEBM video stream is dropped at the demultiplexer and never reaches an encoder at all. That is why this pair is quick: the picture is the expensive part of a WEBM file, and this pass does not look at it.
The running time. The OPUS plays for exactly as long as the WEBM did, because nothing on this path trims, cuts or speeds anything up.
The sample rate and the channel layout of the WEBM audio track, carried into the OPUS unchanged wherever the encoder accepts them, and resampled only where it cannot.
What WEBM to OPUS costs you
The picture, all of it, and on purpose. The WEBM video stream is dropped before any encoder sees it, and the OPUS that comes back is sound alone. Nothing in the OPUS remembers there was ever an image, and no later conversion can put one back.
Cover art carried inside the WEBM. ffmpeg counts an attached picture as a video stream, so the same -vn that drops the film drops the artwork with it. Keeping it is worse: on several audio containers the mux then fails outright with an error nobody can read.
Any second audio track, and the subtitles. ffmpeg selects one stream out of the WEBM, and an OPUS file has nowhere to put a subtitle track in any case.
Exactness, one more time. The audio inside the WEBM was compressed once already, and OPUS compresses the decoded result again. Two lossy passes do not cancel each other out, they stack.
If a model is going to read the OPUS
An OPUS file is a waveform, and a language model reads text, so handing an assistant this OPUS directly gets you nothing. The OPUS becomes useful the moment somebody turns the sound into words, and that is a separate job with its own page here. WEBM to OPUS is the step before it: get the sound out of the WEBM in a container the transcriber accepts, then hand the model a transcript instead of a file it cannot open.
Running WEBM to OPUS against the real engine with a real file, ffmpeg wrote 26,907 bytes of OPUS in 0.2 seconds, machine otherwise idle. That is one WEBM file on one day and not an average, which is why the number is given together with the file that produced it.
The WEBM to OPUS verdict was reached by ffprobe, meaning the container was actually demultiplexed, recognised under the right name, and carried at least one stream. Which check was used matters, because they do not all prove the same thing, and a status code of 200 proves nothing whatsoever about whether the OPUS file has anything inside it.
The same WEBM file, sent somewhere else
Other outputs the probe measured out of a WEBM file, so the cost of choosing OPUS can be read against something. Sizes do not compare across engines, because each family was probed with its own WEBM witness file.
WEBM to GIF: 177,876 bytes in 0.1 seconds, verified by a format signature.
WEBM to MP3: 24,429 bytes in 0.1 seconds, verified by ffprobe.
WEBM to MP4: 46,854 bytes in 0.3 seconds, verified by ffprobe.
WEBM to WAV: 288,078 bytes in 0.1 seconds, verified by ffprobe.
WEBM to FLAC: 178,357 bytes in 0.2 seconds, verified by ffprobe.
WEBM to OGG: 9,503 bytes in 0.2 seconds, verified by ffprobe.
Other ways into OPUS, and what they measured
Among the published routes into OPUS, WEBM is the number 14 largest output of the 16 measured. The witness files differ, so this ranks the probe run and not your document.
WMA to OPUS: 25,770 bytes in 0.2 seconds.
WMV to OPUS: 25,770 bytes in 0.1 seconds.
MP4 to OPUS: 27,864 bytes in 0.2 seconds.
MP3 to OPUS: 27,107 bytes in 0.2 seconds.
3GP to OPUS: 27,864 bytes in 0.1 seconds.
What we will not pretend about WEBM to OPUS
A WEBM file over 150 MB is refused before the upload finishes rather than after it, so you do not wait for a rejection.
A WEBM to OPUS run that passes 240 seconds is killed, and the ffmpeg process is killed with it. A run left behind would sit on one of the machine's two cores until somebody noticed.
The WEBM file behind the measurement on this page was 3 seconds of 320 by 240 test pattern at 25 frames a second, with a sine wave on the audio track, built by ffmpeg for the probe itself. That is why WEBM to OPUS came out at 0.2 seconds, and it is not what your file will do. The measurement proves the pair works and says almost nothing about how long yours will take.
WEBM to OPUS costs half a credit per conversion, whatever the length of the WEBM. That is the same as an image or a document here, and it is a measurement rather than a gesture: an audio target on this engine is done in under two seconds, where a video target can take minutes.
A WEBM longer than 60 minutes is refused before encoding. ffprobe measures the WEBM first and answers with a refusal, rather than starting a job the 240 second guard would kill halfway to the OPUS.
WEBM to OPUS waits at the audio door, which is 2 slots wide and gives up after 5 seconds rather than making you wait twenty times the length of the work. Until the 30th there was one door for everything, and a single video transcode starved every audio conversion for its whole duration: an OPUS that needed less than a second was answered with a refusal after twenty.
Nothing here reads what is inside the WEBM. No transcription, no scene detection, no summary: WEBM to OPUS is a container and codec operation, and the OPUS says exactly what the WEBM said, in a different wrapper. The sound is the whole payload.
One WEBM file at a time, chosen in the browser. There is nothing else to set up and nothing else on offer.
WEBM to OPUS: what people ask
What actually converts my WEBM file to OPUS?
ffmpeg does it, held to a single thread, on a machine we rent and watch. Not a browser trick and not somebody else service: the WEBM file is uploaded once, ffmpeg runs once, the OPUS comes back, and neither file is kept afterwards.
How long does WEBM to OPUS take?
The probe measured 0.2 seconds for 26,907 bytes of OPUS, on a 3 second file. An audio target scales far better than a video one: three minutes of source through this engine takes under a second, because only the sound is encoded. Past 240 seconds any run is stopped, and an audio one does not get close.
What do I lose going from WEBM to OPUS?
The one to know about first: The picture, all of it, and on purpose. The WEBM video stream is dropped before any encoder sees it, and the OPUS that comes back is sound alone. Nothing in the OPUS remembers there was ever an image, and no later conversion can put one back.
Does the OPUS still have the video in it?
No, and that is the entire point of the pair. The WEBM picture is dropped before any encoder touches it, so what you download is sound only. If you wanted to keep the picture you wanted a video target, and it is priced differently because it costs differently.
Is WEBM to OPUS free?
There is a free allowance every month, and one WEBM to OPUS conversion costs half a credit against it however long the WEBM runs. An audio target is charged per conversion and not per minute, because the engine encodes three minutes of sound in well under a second.