Convert M4V to WebP

These are two different jobs and they want two different formats. Artwork for a Plex, Jellyfin or Emby library is matched to a video by file name and handed to a server that expects JPG or PNG, and the JPG route on this site is written for exactly that. Nothing here improves on it.

This page is for the other case: a still that will be downloaded by a browser as one of twenty things on a page while somebody decides whether to stay. Its weight in kilobytes is a cost paid by every visitor, and the format only has to satisfy the CMS, the build step and the browser. That is where WebP wins and where the library case is irrelevant.

The timestamp accepts up to 86,400 seconds, which is a day, so nothing about the length of an M4V puts a moment out of reach. Type the number of seconds and the decoder seeks to it; ask for a time past the end and the conversion tells you how long the video actually is rather than producing something arbitrary.

That is worth saying because the GIF route on this site cannot do it — animated output is always sampled from the beginning of the file. Still extraction has no such limit, which makes it the right tool for long-form material and means a catalogue of hour-long lectures can be illustrated properly without trimming anything first.

WebP grew out of the intra-frame coding in VP8, the codec WebM was built around, so a still taken from a WebM stays inside one lineage. An M4V holds H.264 or HEVC, both from the MPEG side of the fence, and WebP knows nothing about either.

In practice this costs less than it sounds. The frame is decoded fully to pixels in between, so the encoder is working from an image rather than from a compressed representation it has to reinterpret. What it does mean is that no argument about family resemblance applies here: the frame is judged as a picture, and how good a picture it is was decided by the video encoder long before this conversion saw it.

The frame arrives at the video’s own pixel dimensions with nothing scaled, and it is encoded at a fixed quality of 82. There is no quality slider and no resize on this path — the only option is the timestamp.

Both defaults are the right ones for a page. Eighty-two is where WebP stops being visibly lossy on photographic content, and full resolution is what a responsive image pipeline wants as its source, because whatever generates your `srcset` knows the widths the layout needs and does better work from the largest honest version. Take the big file here and let the build step make the small ones.

Video encoders spend their bits where movement is and economise everywhere else, on the assumption that no frame will be stared at. Freeze one and the economy shows as softness. A held shot, a slide, a speaker between gestures — those frames were expensive for the encoder and come out clean; a pan, a cut or a camera move does not.

For long-form Apple material there is a second consideration. The opening minute is titles, a slate or an institutional logo, and the closing minutes are credits, so the useful range is the middle. Ten per cent of the running time is the rule of thumb media servers use for the same reason, and trying two or three timestamps a second apart costs a couple of seconds each because nothing is uploaded between attempts.

Against a JPG of the same frame at comparable quality, WebP is smaller by a margin that is worth having on a grid — around thirty per cent is the figure usually quoted, and it compounds across a catalogue where every entry has a picture. No fallback markup is needed, because current browsers all decode it.

The saving disappears the moment the file leaves the browser. Some email clients will not render WebP, a number of older desktop applications and print workflows reject it, and upload forms on other people’s sites sometimes refuse the extension outright. Inside a page it is the pragmatic default; outside one, take the JPG and stop thinking about it.

WebP is one of the few formats that does stills, animation and an alpha channel in one specification, and it is a reasonable thing to reach for when you want a short silent loop that behaves like a GIF but looks like video. None of that applies here. This conversion writes a single opaque frame.

If motion is what you want from the M4V, the GIF route is what this site offers and its own page explains why the result will be large. If transparency is what you want, there is none to be had — the M4V holds H.264 or HEVC and neither carries an alpha channel in Apple’s delivery pipeline, so the still is fully opaque and nothing about the target format changes that.

Anything bought or rented from the iTunes Store or the Apple TV app is encrypted with FairPlay. There is no key in the browser, so no frame can be decoded, and the conversion stops with a message reporting that nothing in the file could be converted.

If a folder of files works except for a few, the few were purchases rather than damaged. Everything produced by QuickTime, iMovie, Final Cut or Compressor, recorded by a camera, or downloaded from a course or conference platform is unprotected and behaves normally.

The M4V is decoded in the tab with the codecs the machine already has, one frame is encoded, and nothing carrying the video is transmitted. The free tier takes files up to 100 MB, and the time a conversion costs is the time your machine needs to seek rather than anything to do with a network.

The proportions are what make the local approach obvious on this pair. Sending a gigabyte of lecture to a server so that it can return forty kilobytes of thumbnail is a great deal of traffic for the result, and repeating it once per video across a course catalogue is an afternoon spent watching progress bars that did not need to exist.