Intracoding

Intracoding, often referred to as intra-coding in the context of video and image compression, is a data encoding technique that encodes information within a single data unit without relying on data from other units. In video, intracoding encodes a frame using only information from that frame, exploiting spatial redundancy rather than temporal redundancy with reference to other frames.

In practice, intracoding uses block-based transform coding, intra-prediction, quantization, and entropy coding. The encoder predicts each block from neighboring pixels inside the same frame (for example, from the top or left blocks), then applies a transform such as a discrete cosine transform, followed by quantization and entropy coding. The decoder reconstructs each block using the same predictions and then performs the inverse transform and de-quantization. This approach allows a frame to be decoded independently of others, facilitating random access and error resilience.

Intracoding is contrasted with inter-coding, which uses temporal prediction from reference frames to encode differences between frames. The main advantages of intracoding include simpler decoding, robust error handling, and faster random access, especially important for keyframes or scenes with rapid changes. Its main limitation is lower compression efficiency for sequences that exhibit strong temporal redundancy, where inter-coding can achieve smaller bitrates by exploiting motion and temporal similarity.

Applications of intracoding are widespread. In video coding, intracoding is used for I-frames (keyframes) in standards such as MPEG-2, MPEG-4, H.264/AVC, and HEVC. In still image compression, fully intra-coded formats like JPEG rely on intraframe coding, encoding each image independently without reference to other images. Intracoding remains a fundamental tool for reliable, flexible data compression.