interframe

Interframe refers to a family of video compression techniques that encode frames by exploiting temporal redundancy between successive images. In interframe encoding, frames are not stored independently; instead, frames are predicted from one or more reference frames, and only the differences (residuals) are encoded. This approach contrasts with intraframe encoding, where each frame is compressed without reference to others.

The core components of interframe coding are motion estimation and motion compensation. The encoder searches for blocks in reference frames that best match blocks in the current frame and records motion vectors that describe how those blocks moved. The encoder then stores the predicted block data plus the residual error. By using reference frames, interframe coding can achieve much higher compression ratios, especially for video with motion and smooth changes over time.

Within interframe schemes, three principal frame types are commonly used: I-frames (intra-coded frames) serve as independent reference points and do not rely on other frames. P-frames (predicted frames) depend on previous frames, typically preceding I- or P-frames. B-frames (bi-directionally predicted frames) reference both past and future frames to achieve additional compression. The sequence of frames together forms a group of pictures (GOP), which determines the ordering and dependencies of frames and affects latency and random access.

Interframe methods are widely used across modern video standards, including MPEG-2, MPEG-4 Part 10 (H.264/AVC), HEVC (H.265), and AV1. They enable efficient streaming, broadcasting, and storage but introduce latency, error propagation risks, and increased decoder complexity. Proper management of I-, P-, and B-frames is essential for balancing compression efficiency, quality, and latency.