What is Video Compression?
Video compression reduces the data needed to represent moving images by removing spatial, temporal, or perceptually insignificant information. A codec defines how an encoder represents frames and a decoder reconstructs them.
How Video Compression works
Modern encoders exploit similarity within a frame and across neighboring frames, then quantize transformed prediction errors and entropy-code the result. Intra pictures provide self-contained reference points, while predicted pictures reduce size by depending on other decoded images. Lossless modes preserve sample values; delivery formats generally use controlled loss to reach practical rates. Compression runs between acquisition or editing masters and storage or distribution, with decompression required at playback and many analysis stages.
Key facts
- Long prediction groups can improve compression efficiency, but they increase random-access work and allow corruption from a damaged reference picture to affect later frames.
- Rate control may target a fixed quality, an average rate, or a constrained transmission rate; each mode distributes bits differently across simple and complex scenes.
- Repeated lossy encodes compound quantization damage, especially around text, grain, and sharp edges, so derivatives should be generated from a retained master rather than one another.
When Video Compression matters
Choose codec settings by balancing visual quality, bitrate, encoding cost, and decoder support. Aggressive lossy compression saves bandwidth but can introduce artifacts that later re-encoding amplifies.