Compression technique storing only differences between consecutive frames, not full images — drastically reduces file size and bandwidth. H.264, H.265 standard; uncompressed proxies bypass it.
Anyone working with video material will inevitably encounter the concept of interframe coding — and where it hurts: data volume. Instead of storing each frame as a complete image, the codec only remembers the differences between successive frames. This sounds theoretical, but in practice, it's the only thing that makes sense for modern workflows. A 10-second shot in 4K — uncompressed, that's 80 GB per second. With interframe coding? Suddenly, it fits on an SSD.
The mechanics work like this: The codec identifies a keyframe — a complete image — and stores it. The subsequent frames are encoded as differences: only the pixel changes. In a static scene with moving talent, there's no need to re-store all the background information every time. H.264 and H.265 (HEVC) work on this principle — that's why files in ProRes or DNxHD material are so much smaller than uncompressed stuff. The catch: The more volatile the motion, the more cuts or scene changes, the more frequently a new keyframe needs to be set, and the savings shrink.
On set or in post, this makes a difference that cannot be ignored. For offline editing, one consciously works with compressed interframe compression — ProxyGen with H.264 saves storage and makes editing fluid. For color grading or VFX-intensive shots, one then switches to intraframe coding or even uncompressed material — here, latency bandwidth and frame accuracy are more important than storage economy. A motion graphics shot with fast transitions? Interframe codecs can be problematic because keyframe distances create artifacts. Here, it's better to switch to I-frame only in the edit or use ProRes 422 HQ.
The insidious part: When editing later with compressed interframe compression, the difference data breaks at the cut points. That's why one exports to the final cut in higher quality. For live streaming or real-time transmission, interframe coding is indispensable — compression makes it possible. For long-term archiving: rather not, too susceptible to decoding errors over decades.