Initializing... drag & drop files here
Supports: RAF
This page turns Fujifilm raw stills into an AV1-coded video clip: one .raf becomes a short held shot, several become a slideshow. The result is genuinely small — AOMedia describes AV1 as "an open video codec designed to provide high-quality video compression with greater efficiency than previous codecs", and a still image is close to the best case for it. The one thing worth understanding before you start is what the .av1 file really is, because the extension is a codec name, not a container.
| Property | Value |
|---|---|
| Full name | Fujifilm RAW image file |
| Vendor | Fujifilm (X-series and GFX bodies) |
| Sensor readout | 14-bit on current X-series bodies, per Fujifilm's X-T5 specification |
| Variants | Uncompressed, compressed and lossless-compressed raw |
| Colour filter array | X-Trans on most X-series bodies; conventional Bayer on GFX |
| Depth reaching the encoder here | 8 bits per channel — the raw decoder develops at 8-bit by default |
| Audio | None. It is a still image |
| Decoded by | A LibRaw-based reader that lists .raf natively, so no intermediate export is needed |
| Property | Value |
|---|---|
| Container written | Matroska — the same structure as an .mkv, under an .av1 filename |
| Video codec | AV1 |
| Encoder | SVT-AV1 |
| Pixel format | yuv420p (8-bit, 4:2:0) |
| Frame rate | 1 fps — a still has no motion to sample |
| B-frames | Disabled; identical frames gain nothing from them and they add a start-time offset |
| Default clip length | 5 seconds per image |
| Audio track | None. An image source has no audio, and the audio controls are hidden accordingly |
| Maximum frame size | Long edge clamped to 4096 px before encoding |
| Transparent areas | Filled with the Background Color setting, which defaults to white here |
| For playback | Rename to .mkv. Nothing dispatches on a .av1 extension |
Matroska is an open, EBML-based container that can hold many video, audio and subtitle tracks, which is why it is the natural home for an AV1 bitstream. If you would rather have the correct extension from the start, use RAF to MKV and pick AV1 as the codec — same bytes, sane filename.
.raf files onto the page or click "+ Add Files". Only .raf inputs are accepted. Upload order becomes clip order when several files are merged.Because the source is a still image rather than a video, the panel deliberately hides everything that needs a source bitrate to work from. Three modes render:
Every bitrate and target-file-size mode is hidden on this page, so do not go looking for a "target 8 MB" field. Size is controlled by resolution, image duration and the quality mode. A 5-second, 1 fps AV1 clip of a single photograph is five nearly identical frames, so it compresses extremely well regardless — expect a file far smaller than the source raw.
Because .av1 is not a container extension, it is a codec name. The bytes we write are a Matroska stream, which is exactly what an .mkv holds, but players dispatch on the extension and nothing is registered for .av1. Rename the file to .mkv and it opens normally. If you would rather skip that step, convert to RAF to MKV and select AV1 as the video codec instead.
A photograph has no audio track to carry, so there is nothing to encode. The audio controls are hidden on this page for exactly that reason, rather than offering a codec that would do nothing. If you need a soundtrack, add it afterwards in a video editor.
Every image is normalised to a clean PNG before it reaches the video encoder, and that step clamps the long edge to 4096 pixels. Fujifilm lists the X-T5 at 7728×5152, so a full-resolution frame arrives in the video at 4096×2730. Any 3:2 Fujifilm frame wider than 4096 lands on that same size. This is a deliberate guard — very large still frames are unreliable input for the video pipeline — and 4096 is well beyond what any normal playback target needs.
The encode runs at 1 frame per second, and there is no frame-rate control on this page. That is not a limitation in practice: a static image held for 5 seconds is the same picture whether it is stored as 5 frames or 150, and encoding it once per second produces a dramatically smaller file. Use Image Duration to change how long the shot lasts.
The raw decoder develops Fujifilm raw at 8 bits per channel by default, and the encoder then writes yuv420p — 8-bit with the colour channels at half resolution. So neither the extra bit depth nor full-resolution chroma survives into an AV1 clip. That is normal for consumer video, but it means this is a delivery format, not an archival one. Keep the .raf masters.
AV1 generally produces smaller files at the same visual quality and carries no licensing baggage, but it encodes more slowly and older hardware may lack a decoder. H.265 has broader hardware decode support, especially on Apple devices and TVs. For a handful of still frames the size difference is small in absolute terms, so pick based on what has to play the file — RAF to HEVC is the H.265 equivalent of this page and has exactly the same container caveat.
Yes, and it is the default. Merge strategy ships on "Merge images", so every .raf you upload becomes one shot in a single video, each held for the Image Duration you set, in upload order. Choose "Video per image" if you would rather get one file per photograph.
Because the Background Color option defaults to white for image sources. It fills two things: any transparent region in the source, and the letterbox bars if you pick a target frame shape that does not match the photograph's aspect ratio. A raw file has no transparency, so in practice this only affects letterboxing — set the colour to black in Advanced Options if you want cinema-style bars. Files are uploaded over an encrypted connection, processed on our servers, and deleted automatically after a few hours, with no sign-up and no watermark; the RAF converter hub lists every other target this format can reach.