Pull a Still Frame Out of a DVR Recording as WebP
A .dvr file is a Microsoft DVR-MS recording — the recorded-TV container Windows XP Media Center Edition, Vista, and Windows 7 wrote, holding MPEG-2 video inside an ASF wrapper. This tool decodes it, takes the frame you point at, and writes that single frame as a WebP still. Pick the timestamp, or ask for a run of screenshots across the whole recording and get one WebP per capture.
How to Convert DVR to WebP
- Upload Your DVR File: Drag the recording onto the page or click "+ Add Files". More than one
.dvrcan be queued, and every file is processed with the same settings. - Choose the Frame: Under "Frame Selection", "Specific Frame" is preselected with a "Time (seconds)" box starting at 0. Type the moment you want —
2.100means 2 seconds and 100 milliseconds in. Switch to "Multiple Screenshots" instead and a "Capture Rate" dropdown appears, preset to one frame per second. - Set Quality and Size: "Quality Preset" opens on "Very High (Recommended)". Switch the mode to "Specific file size" to name a byte budget instead — the box is prefilled at 8 MB. "Lossless?" is preset to "No (Recommended)", and "Image resolution" starts on "Keep original" with presets from 16P up to 4320p available.
- Convert and Download: Click Convert and save the WebP. Files upload over an encrypted connection, are processed on our servers, and are deleted automatically after a few hours. No sign-up, no watermark.
Picking the Right Frame Selection Mode
The two modes solve different problems and the second one can produce a lot of files, so it is worth knowing what each capture rate means before you start a two-hour recording.
| Frame Selection | What you get | Output count from a 60-minute recording | Good for |
|---|---|---|---|
Specific Frame at 0 |
The opening frame | 1 | A quick thumbnail |
Specific Frame at 1830.500 |
The frame 30 min 30.5 s in | 1 | Grabbing one exact moment |
| Multiple Screenshots — 10 seconds per frame | A still every 10 seconds | 360 | Skimming a long show |
| Multiple Screenshots — 1 second per frame (preset) | A still every second | 3,600 | Finding an exact moment to re-grab |
| Multiple Screenshots — 0.1 seconds (single frame at 10fps) | Ten stills per second | 36,000 | Frame-accurate inspection of a short clip |
WebP Against the Other Frame-Grab Targets
| Property | WebP | PNG | JPG |
|---|---|---|---|
| Lossy coding | VP8 intra-frame, 8-bit Y'CbCr 4:2:0 | none | JPEG DCT |
| Lossless mode | Yes, RGBA — the "Lossless?" toggle | Always | No |
| Alpha channel | Yes | Yes | No |
| Typical saving | About 30% smaller than JPEG at matched quality; lossless WebP about 26% smaller than PNG | Baseline for lossless | Baseline for lossy |
| Maximum dimensions | 16383 x 16383 px | Effectively unbounded | 65535 x 65535 px |
| Global browser support | 96.18% — Chrome 32+, Edge 18+, Firefox 65+, Safari 16.0+ (iOS 14+), Opera 19+ | Universal | Universal |
| Best when | Shipping the still to the web | Editing or archiving the still | Pasting into anything, however old |
For the same recording, DVR to PNG keeps the grab lossless and DVR to JPG maximises compatibility.
Frequently Asked Questions
Which frame do I get if I leave the time at 0?
The first decodable frame of the recording. On a Media Center capture that is usually a fraction of a second of channel logo or black before the program starts, so a grab at 0 is often less useful than it sounds. Nudging the time to a few seconds in almost always gives a better thumbnail.
Why does my frame grab have comb lines across moving objects?
Because broadcast television is frequently interlaced. A 1080i or 480i recording stores each frame as two fields captured a fraction of a second apart, and a still pulled straight out shows them woven together as horizontal combing on anything that moved. It is in the source, not the encoder, and nothing here removes it. Grabbing during a static shot avoids it; otherwise deinterlace the recording in a desktop editor first and take the still from that progressive export.
Should I turn "Lossless?" on for a frame grab?
Usually not. Broadcast MPEG-2 has already been through a lossy encoder, so lossless WebP faithfully preserves compression artefacts at several times the file size. Lossless is worth it when the frame is mostly flat graphics — a scoreboard, a title card, a channel bug — where the RGBA lossless coder is genuinely efficient. For normal footage, leave it on "No (Recommended)" and let the Quality Preset do the work.
Can I get a whole set of stills instead of one frame?
Yes. Switch Frame Selection to "Multiple Screenshots" and pick a capture rate; each extracted frame becomes its own WebP file. Check the table above before choosing — at the preset one-frame-per-second rate, an hour of television produces 3,600 images.
Does the quality preset touch the frame before it is encoded?
No. Our pipeline runs in two stages: the frame is first decoded out of the MPEG-2 stream as a lossless PNG, and only then encoded to WebP. That means the Quality Preset governs exactly one lossy step, and the extraction itself never adds artefacts of its own.
How large can the WebP be?
The format uses 14 bits each for width and height, so 16383 x 16383 pixels is the hard ceiling. No broadcast recording comes anywhere near it — even a 4320p grab is 7680 px wide — so in practice the only size lever you need is "Image resolution", which starts on "Keep original" and will hand you the frame at the recording's native dimensions.
Which browsers and apps open WebP?
WebP is supported by 96.18% of global browser traffic: Chrome 32+, Edge 18+, Firefox 65+, Safari 16.0+ on macOS, Safari 14+ on iOS, and Opera 19+. Desktop image viewers have caught up more slowly — Windows Photos and macOS Preview both handle it now, but older editing suites may not, which is the main reason to choose PNG or JPG instead.
Why is there no animated WebP option on this page?
Because a WebP target here is treated as a still capture, which is what the Frame Selection controls exist for. To turn a stretch of the recording into a moving image, use DVR to GIF or DVR to APNG instead — those pages expose framerate and trim controls rather than a frame picker.
My .dvr file will not load — what is wrong?
Two common causes. If Windows Media Center flagged the broadcast as copy-protected, the recording only decodes on the machine that made it and no server-side converter can read it. Alternatively the file may not be DVR-MS at all: security recorders, set-top boxes, and dash cameras also write .dvr files in proprietary formats. If VLC cannot identify a video stream in the file, this tool will not be able to either.