Convert M4V to WebM

Every other conversion in this set carries something over. The MP4 and MOV targets keep the same H.264 and AAC the M4V already held; the audio targets keep the sample rate. This one keeps nothing. WebM is written with VP9 video and Opus audio, and the M4V holds H.264 or HEVC with AAC, so both streams are decoded to raw and built again from scratch by different codecs.

That is the point rather than a side effect. The reason a destination asks for WebM is almost always that H.264 and AAC are patent-encumbered and VP9 and Opus are not. A conversion that preserved anything of the original codecs would not have solved the problem the reader actually has.

The most common reason people look for a WebM is "I am putting a video on a web page", and for that purpose the MP4 route is better. H.264 in an MP4 is decoded natively by every current browser, it encodes in a fraction of the time, and it is the file you would need anyway as a fallback. If nothing has told you that H.264 is unacceptable, take the MP4 and move on.

What makes WebM worth the wait is a rule you did not write. Some projects will not host patent-encumbered video at all. Some organisations avoid H.264 for licensing reasons that have nothing to do with quality. And if you are already serving two sources from one `<video>` element, listing the VP9 file first lets browsers that can take it download fewer bytes for the same picture.

Open-content platforms are the clearest case. Wikimedia Commons and a number of open archives accept WebM and Ogg and decline formats that carry patent licences, which means an Apple-produced lecture or demonstration cannot be contributed as it stands. The conversion is the price of admission rather than an improvement.

Documentation systems, static-site pipelines and a few upload forms behave the same way for their own reasons, usually a policy inherited from an open-source project. What these destinations share is that the requirement is written down somewhere. Nobody’s television asks for VP9, which is precisely why this page has a different reader from the MP4 and MKV pages.

H.264 encoders are in essentially every phone, laptop and graphics chip made in the last fifteen years. VP9 encoders are far less common in hardware, so the work more often falls to software running in the tab. The same clip that becomes an MP4 in one unit of time can take several to become a WebM, and the gap grows with resolution rather than with file size.

Plan around it. Convert the clip you need rather than the whole recording, drop the resolution if the destination is a page rather than a screening, and leave the tab open — closing it cancels the work instead of queueing it. If the browser has no VP9 encoder available at all, the conversion says so before it takes the file rather than failing partway through.

VP9 describes the same picture in fewer bytes than H.264 does, which is one of the arguments for it, but the far larger saving on a web page is not choosing a better codec — it is not shipping pixels the layout never displays. A 4K lecture recording embedded in a column 800 pixels wide is spending most of its bytes on detail the browser immediately throws away.

The control offers 2160p down to 360p with the width following automatically so nothing is stretched, and asking for more than the source has is ignored on purpose. Reducing before encoding also shortens the encode itself, which on this target is the difference between a wait and an afternoon.

The AAC in the M4V becomes Opus, published by the IETF in 2012 as RFC 6716 and written at Xiph.Org and Skype. Opus is a stronger codec than AAC at the bitrates web video uses, particularly on speech, and every browser decodes it because WebRTC made support mandatory. On a recorded talk or a narrated demonstration that is a real gain rather than a sideways move.

It is still a second generation of lossy compression, and no encoder can put back what the AAC pass removed. What the better codec buys is that the second pass costs less than it would have with a weaker one — which is why the balanced band is usually enough here even on material where it would not be enough elsewhere.

An M4V bought from the iTunes Store is encrypted with FairPlay. There is no key in the browser, so the tracks cannot be decoded, and the conversion reports that nothing in the file could be converted. No tool on this site removes protection and no site that claims to should be handed the file.

The collision is not only technical. The destinations that want WebM are open-content projects, and material bought from a store is not yours to publish there in any format. Unprotected M4V files — QuickTime, iMovie, Final Cut and Compressor exports, camera output, course downloads — convert normally and are usually the ones somebody has the right to post.

In a `<video>` element, list the WebM as the first `<source>` and an MP4 as the second. Browsers pick the first they understand, so the ones with a VP9 decoder take the smaller file and everything else falls back without any JavaScript involved. Serving WebM alone is a decision to lose whoever cannot decode it, which is a smaller group than it once was and not an empty one.

Add a poster image while you are there. A still taken from the video itself gives the reader something to look at before playback starts, and the frame-extraction pages on this site produce one from the same M4V — take it a few seconds in rather than at zero, where most recordings are still a black frame.

Decoding, encoding and muxing happen inside the tab on your own processor, using the codecs the machine already ships. Nothing carrying the video is sent anywhere and there is no account. The free tier accepts files up to 100 MB, which on this pair is rarely the binding constraint, since the material being published to the web is usually a clip rather than a feature.

Batches work the same way: up to a hundred files at once, each with its own progress, downloadable together as a ZIP. Given how much processor time VP9 asks for, a large batch is worth starting when you do not need the machine — the cost is your own hardware rather than somebody’s queue, which is also why there is no daily limit on it.