Video to GIF

You are trading a good codec for a bad one

Video formats describe motion: a frame is stored as 'this block moved, here is a small correction'. GIF has no such concept and stores frames close to whole, so a clip that was a few hundred kilobytes as video routinely becomes several megabytes as a GIF.

That is not a flaw in the conversion, it is the format. The reason to accept it is placement — GIF plays as an image, so it works in email, chat boxes and forums that will not take a video element at all.

Trim before anything else

Duration costs linearly, and most clips contain two or three seconds that carry the point plus several that do not. Cutting to the useful moment is usually a larger saving than every quality setting combined.

A loop also reads better when it is short. A two-second loop is understood as a loop; an eight-second one is watched once and then becomes a distraction on the page.

Camera movement is the expensive part

Because GIF cannot describe motion, a panning or handheld shot changes every pixel in every frame and nothing can be skipped. A locked-off shot where only one element moves compresses dramatically better.

If the source moves, cropping to the region that actually matters recovers much of the difference — you remove pixels that were being re-stored on every single frame.

Settings, in the order worth adjusting them

Each of these is more effective than the one after it.

• Dimensions: 320-480 px wide is plenty for most uses and cost scales with area.
• Duration: cut to the seconds that carry the point.
• Frame rate: 10-15 fps reads as smooth for screen content.
• Palette: fewer colours, and turn dithering off on flat-coloured content.