Video to Image Frames

Video to Image Frames samples a video at an interval you choose and returns the frames as a ZIP of numbered JPG or PNG images.

The interval is in seconds, from a tenth of a second up to a minute, and it is applied as an exact frame rate rather than as a count of frames to skip — which is what makes it behave identically on a variable-frame-rate phone recording and on a studio export. A tenth of a second is the floor, so ten images per second is the ceiling here; this is a sampler, not a way to recover all 60 frames of every second. A cap on the number of frames, up to 600 and 60 by default, keeps a long clip from exhausting memory, and when the cap cuts a clip short the result says how much of it was covered. Frames are never scaled up past the source width, and the ZIP is stored without compression because JPG and PNG are already compressed.

Decoding and zipping both happen in the page, so footage you are turning into a dataset or a reference sequence is never uploaded.

Building an image set from video for annotation, pulling reference frames for rotoscoping or tracing, sampling a long recording to scan it by eye, and extracting stills from a timelapse at a fixed interval.

Also known as: split video into images · extract all frames · video to png sequence · frame by frame extractor · convert video to frames · change video to frames