Convert WebM to TIFF

This is the one still format in this set that people arrive at under instruction. A print house asked for TIFF. A document management system rejects everything else. A retention policy names it. A scanner’s software writes it and the workflow around that software assumes it. Nobody weighing up formats for a web page ends up here.

That changes what the page owes you. There is no argument to make about which format is better, because the choice has been made elsewhere and is usually not negotiable. What is worth knowing is what kind of TIFF this produces, how big it will be, and whether a frame taken out of a compressed video is good enough for whatever it is being submitted to — which, depending on the size it will be reproduced at, it may not be.

TIFF is less a single format than a filing system for images, and the readers that claim to support it agree on less than you would hope. LZW compression is patent-free and widely understood; Deflate is not universally supported; JPEG-in-TIFF is a lossy image wearing a TIFF wrapper and defeats the point of asking for TIFF in the first place.

So what is written here is the plain uncompressed baseline: pixels, in order, with a header describing them. It is the one variant every printer driver, every scanner application and every archive ingest process accepts. The cost is size, and the size is not small — but somebody converting to TIFF is nearly always feeding a specific piece of software, and the safest possible file is worth more to them than a smaller one that might be rejected.

Four bytes per pixel, uncompressed, regardless of content. A 1080p frame is 1920 by 1080, which comes to about 8 MB. A 4K frame is four times the area and lands near 33 MB. A flat grey screen and a detailed photograph produce identical file sizes, because nothing is being compressed and there is nothing for content to affect.

Compare that with the same frame as a PNG, which is lossless too and typically a few hundred kilobytes for interface content, and the gap looks absurd. It is the trade TIFF exists to make: no decoder, no compression algorithm, no interpretation between the file and the pixels. For an archive that expects to still open the file in thirty years, that simplicity is the feature.

Print is measured in dots per inch, and 300 is the conventional figure for anything held in a hand. A 1080p frame is 1920 pixels across, which at 300 dpi is 6.4 inches — about 16 centimetres, a little under A5 width. A 4K frame gives 12.8 inches, which covers most of an A4 page.

That is the honest ceiling, and it is where a lot of these conversions run into trouble. A video frame asked to fill a poster or a double-page spread is being enlarged several times over, and enlargement magnifies the video codec’s artefacts along with everything else. If the reproduction size is large, the answer is a photograph or a higher-resolution recording rather than a different file format.

A still lifted from compressed video carries what the codec did to it. Bits were spent on motion rather than on any single frame, colour is stored at half resolution in each direction, and a frame captured during movement was never encoded sharply to begin with. TIFF preserves every bit of that with complete fidelity.

That is the correct behaviour and it is why archives want it — the file is a record of what was decoded, not an improvement on it. It also means the format cannot rescue a poor frame. If the still looks soft, the fix is a different timestamp: find a moment where nothing is moving, where the encoder had the bits to describe the picture properly, and take it from there.

Type the number of seconds from the start of the recording. The default of zero is the very first frame, which for anything captured from a screen or a camera is usually before the subject appears.

For records work there is a second consideration that does not apply to a thumbnail: the frame has to show the thing it is being submitted to prove. A moment half a second earlier may be sharper and useless. Pick the frame that is evidentially right first, then, if it is soft, look at the frames immediately around it for one that is both right and clean.

TIFF can hold many images in one file — it is one of the reasons scanning software adopted it, and why multi-page TIFFs are common in document archives. This conversion writes one frame per file, because a video has no natural page structure and guessing at one would be worse than not offering it.

If the destination expects a multi-page TIFF, convert the frames you need individually and combine them afterwards with a tool built for the job; ImageMagick does it in one command. That keeps the choice of which frames, and in what order, with the person who knows why the record is being made.

This pair has the worst upload arithmetic on the site. The input is a recording that might run to hundreds of megabytes, and the output is a 33 MB uncompressed image. Sending both across a network to a converter and back is minutes of transfer for a few seconds of actual work.

Here the WebM is decoded in the tab, one frame is written out, and nothing carrying either file moves. For material destined for a records system — compliance recordings, evidence, case files, anything under a retention policy — that is not merely faster. It is usually the only version of the workflow that could be described to whoever set the policy.