Initializing... drag & drop files here
Supports: DIVX
This converter pulls a still frame out of a DivX video and writes it as a Windows Bitmap — the format Microsoft has shipped since the first days of Windows, which stores pixels plainly and predictably instead of compressing them. That is exactly why people still ask for it: legacy Windows imaging software, embedded displays, machine-vision pipelines and a surprising number of industrial tools want a bitmap and nothing else. By default the converter takes the very first frame of the clip, but you can name any timestamp, or ask for a whole series of stills at a fixed interval.
.divx file or click to browse for it. Several clips can be queued at once, and each is uploaded over an encrypted connection.The pipeline runs in two stages, which explains why the result looks the way it does. First the DivX video is decoded and the requested frame is written out as a lossless PNG — nothing is thrown away at this point, so what lands in stage two is exactly what a player would show you at that timestamp. Then that PNG is converted to BMP using the quality preset you chose.
Two things follow from this. The BMP is a faithful copy of the decoded frame, not of whatever camera or master the video came from: DivX is lossy MPEG-4 Part 2 with chroma subsampling, so any blocking, banding or colour smearing in the source is part of the frame you are saving. And because the intermediate is lossless, the only quality decision that matters is the preset — there is no hidden JPEG-style compression step in between.
If you asked for Multiple Screenshots, the same two stages run once per captured frame, so a long clip at a fast capture rate produces a lot of large files. A ten-minute video sampled once a second is 600 bitmaps.
BMP has no quality dial in the sense JPEG does. Instead the three presets select different bitmap encodings, and the difference in file size between them is dramatic:
| Preset | What is written | Colour fidelity | Relative size |
|---|---|---|---|
| High (default) | Uncompressed 16-bit RGB555 pixels | 32 levels per channel — smooth, but gradients can band slightly | Largest by far; fixed, predictable size |
| Medium | 256-colour palette with Floyd-Steinberg dithering, RLE-compressed, written as the older BMP3 variant | Dithered, which hides most banding at normal viewing distance | Roughly a tenth of High on typical footage |
| Low | 86-colour palette with Floyd-Steinberg dithering, RLE-compressed | Visibly reduced; posterised in smooth areas | Smallest |
Because the High preset is uncompressed, its size is pure arithmetic: 2 bytes per pixel plus a small header, regardless of what the picture contains. We measured it against the encoder settings this page uses — a 640x480 frame comes out at 614,454 bytes, a 1920x1080 frame at 4,147,254 bytes, and a 3840x2160 frame at 16,588,854 bytes. Medium and Low are RLE-compressed, so their sizes depend on how flat the image is.
| Property | BMP (this page) | PNG | JPG |
|---|---|---|---|
| Compression | None at the High preset; RLE on the palette presets | Lossless DEFLATE | Lossy DCT |
| Typical 1080p frame | ~4.1 MB at High | Usually 1-2 MB | Usually a few hundred KB |
| Pixel-exact | Yes at High, within the 16-bit colour grid | Yes, full 24-bit | No |
| Transparency | Alpha only in later header versions, with patchy reader support | Yes, well supported | No |
| Metadata | Essentially none | Yes | EXIF |
| Created by | Microsoft | PNG Development Group / W3C | Joint Photographic Experts Group |
| Pick it when | Legacy Windows software, embedded systems or machine vision demands a bitmap | You want a lossless still at a sane size | You are sharing or embedding the image |
At the default High preset, exactly 2 bytes per pixel plus a 54-byte header, because those pixels are stored uncompressed as 16-bit RGB555 — 614,454 bytes for a 640x480 frame, 4,147,254 bytes for 1080p and 16,588,854 bytes for 4K. Content makes no difference; a black frame and a detailed one are byte-for-byte the same size. The Medium and Low presets switch to a palette with RLE compression and come out roughly an order of magnitude smaller, at the cost of colour fidelity. If you want a lossless still without the bulk, DivX to PNG is the better target.
No — the High preset writes 16-bit RGB555, which gives 32 levels per colour channel rather than 256. In practice that is invisible on most video frames, since the source is already chroma-subsampled lossy video, but it can produce faint banding across a smooth sky or a slow gradient. If you need genuine 24-bit-per-pixel colour from your frame, use the PNG target instead: PNG stores full truecolour losslessly and still produces a smaller file than a 16-bit BMP.
Yes. Leave Frame Selection on Specific Frame and type the time into the Time (seconds) field — decimals are accepted, so 5.25 seeks to five and a quarter seconds. The frame returned is the one the decoder produces at that position, which for an inter-frame codec like DivX means the nearest decodable frame at or around your timestamp rather than a mathematically exact instant. Leaving the field at its default of 0 gives you the first frame of the clip.
Switch Frame Selection to Multiple Screenshots and pick a Capture Rate. The dropdown covers one frame every 0.1, 0.2, 0.3 and 0.5 seconds, then one per second (the default) through to one every ten seconds. The converter walks the clip from start to finish at that cadence and returns a numbered sequence. Choose the interval deliberately: at BMP sizes, a fast capture rate on a long clip produces a very large download.
Because you are looking at decoded DivX. MPEG-4 Part 2 is a lossy codec that discards high-frequency detail and stores colour at reduced resolution, so blockiness around motion, ringing at hard edges and smeared colour are all baked into the source frames. Saving to an uncompressed format preserves those artefacts perfectly rather than removing them. Nothing in the conversion adds new loss at the High preset — but nothing can subtract the loss that is already there either.
Only awkwardly. Alpha arrived with the BITMAPV4HEADER in Windows 95 and the undocumented V3 header before it, and reader support has been inconsistent ever since — plenty of software silently ignores the alpha channel or renders it as black. It is a moot point here in any case, because DivX video has no alpha channel to preserve. If you need a still with real transparency, use PNG.
Yes, under Image resolution. Keep original is the default; Resolution Percentage scales by a ratio, Preset Resolutions snaps to a standard height, and Width, Height or Width x Height take explicit pixel values. Scaling down is a sensible way to control BMP's size, since size is a direct function of pixel count. Scaling up is rarely worth it — a DivX rip is typically 480p or 576p, and enlarging it just interpolates pixels that were never recorded.
Because BMP has essentially no metadata container. The format stores a header describing dimensions, bit depth and compression, and then the pixels — there is nowhere to record a capture time, camera settings or a colour profile. If you need the still to carry metadata, DivX to PNG or a JPG target is the right choice, since both support metadata blocks.
It is uploaded over an encrypted connection, processed on our servers, and deleted automatically after a few hours. Files are never shared or made public, no account is needed, and the output carries no watermark. Only the frames you asked for are produced — the rest of the video is decoded in passing and discarded along with the upload.