Video Frame Capture

How the Video Frame Capture works

Pausing a video and pressing Print Screen gives you the player's controls, the window border, a cursor and whatever resolution the window happened to be. Grabbing the frame itself gives you the exact pixels the video contains, at the video's own resolution, with nothing on top. That difference is why editors, support teams and analysts pull frames rather than screenshot them, and it is what this tool does, in your browser, without uploading the file anywhere.

How video frame capture works

Every browser ships a hardware-accelerated video decoder. A <video> element can be seeked to any timestamp and then drawn onto a canvas as a still image, which is then saved as PNG or JPEG. The file never leaves your machine, and because only the part of the file around your timestamp is decoded, grabbing a frame from a four-gigabyte screen recording takes about as long as grabbing one from a short phone clip.

Where frame grabs are used

• Thumbnails. YouTube recommends a 1280 × 720 thumbnail, and a frame from a 1080p video already exceeds that. Grab a strong moment, then add text in the YouTube thumbnail maker.
• Bug reports and documentation. A frame from a screen recording shows the exact state of an interface at the moment an error appeared, at full resolution, without re-running the steps.
• Sport and motion analysis. Coaches step through a clip a fraction of a second at a time to check a golf swing, a running stride or a line call. At 30 frames per second each frame is 0.033 s apart, so a time of 12.40 and 12.43 gives neighbouring frames.
• Evidence and records. Dashcam, doorbell and CCTV footage often needs a clear still of a number plate or a face for an insurance claim. Use PNG so nothing is lost.
• Social posts, slides and articles. A still from your own footage is often better than stock imagery, and it is yours to use.

Why the captured time may not match what you asked for

Video compression stores complete images only occasionally, as keyframes, typically every one to ten seconds, and describes everything in between as changes from them. To show an arbitrary frame, the decoder starts at the previous keyframe and works forward. Most MP4 files carry an index that makes this exact. Files from some screen recorders, files cut off mid-write and some WebM streams have no usable index, so the browser lands on the nearest keyframe instead. When the frame captured is more than a quarter of a second from the one you asked for, the note under the result tells you how far it drifted. That is a property of the file, not something any tool can fix; remuxing the file with a tool such as FFmpeg usually adds the missing index.

How to read the result

• Duration is the length of the video. A time beyond it cannot be captured.
• Asked for is the time you entered, and Captured at is where the decoder actually landed. If the two differ noticeably, see the keyframe note above.
• Video size is the resolution stored in the file, such as 1920 × 1080. This is the real resolution, not the size of the preview on screen, so a 4K source gives a 3840-pixel still.
• Frame size is the resolution of the exported image: the same as the video at Full size, half each dimension at Half size.
• Source is the size of the video file, and Output is the format and size of the still. A 1080p PNG is typically 1–4 MB; the same frame as JPEG is usually a few hundred KB.

The downloaded file is named after the video and the captured time, such as holiday-12-40s.png, so a series of grabs sorts in order.

Choosing PNG or JPEG

PNG is lossless: the file holds exactly the pixels the decoder produced. Use it when the still will be edited, cropped, compared, printed or used as evidence. JPEG is saved at 92 % quality, which looks the same to most eyes at a fraction of the size, but it compresses pixels that have already been through lossy video compression once, so blocky edges and banding in dark areas get slightly worse. Use JPEG for web thumbnails and email, where size matters more. If you need a smaller PNG later, the guide on how to compress an image for the web covers the options.

Tips and pitfalls

• Pick a sharp frame. Moving subjects blur within a single frame, especially in low light where the camera used a slow shutter. Step forward and back a few hundredths of a second and keep the sharpest one.
• Quality starts with the source. A heavily compressed video gives a blocky still no matter what format you save. Frames from a high-bitrate original look far better than frames from a re-uploaded copy; see choosing a video bitrate for why.
• Overlays are not in the video track. Subtitles, player controls and on-screen timecodes added by a player never appear in the still. Subtitles burned into the picture do.
• Codec support is your browser's. MP4 with H.264 and WebM with VP8 or VP9 work almost everywhere. HEVC, used by many iPhones, and AV1 depend on the browser and platform, and a file the browser cannot decode fails with a clear message rather than a blank image. Converting it to H.264 MP4 first is the reliable fix.
• Half size is for thumbnails. It halves both dimensions, so the file is roughly a quarter of the size, which is plenty for previews and chat messages.

Related tools