How a video codec turns motion into compact data
A codec—the software that compresses and decompresses video—does not have to describe every picture from scratch. It can begin with a key frame, a complete image, then encode later frames as instructions for reconstructing what changed.
For example, if a camera stays still while a person walks across a room, the codec can preserve the room’s details and describe the person’s movement separately. It often divides the image into small blocks, estimates where those blocks moved, and records the differences between the predicted picture and the actual one. The player uses those instructions to build each displayed frame.
Why the savings depend on the scene
A static shot is especially compressible: large areas remain predictable, so few new details need to be recorded. A sudden cut, fast camera pan, or burst of visual complexity is harder to predict and generally requires more data. Compression is therefore not a fixed reduction applied evenly to every frame; the amount of information needed varies with the footage.
Many codecs also use frames that refer to both earlier and later pictures. This can save more space, but it means a frame may not be independently readable. The player must first decode the pictures it depends on.
The tradeoff behind seeking and editing
Key frames provide fresh starting points. When you jump to a timestamp, the player can begin at a nearby key frame and decode forward, rather than reconstructing the entire video from its beginning. More frequent key frames can make seeking and editing more responsive, but they also take up storage because each one contains a full image.
That is why a compressed video is better understood as a sequence of reconstruction clues than as a stack of complete photographs. The codec saves space by exploiting what the eye sees as continuity—while keeping enough complete frames and change information for the sequence to play.
