What is DCT Compression?
DCT compression applies the discrete cosine transform to express image or video samples as frequency coefficients. Quantizing and encoding those coefficients reduces data size, usually with some loss of detail.
How DCT Compression works
A DCT maps correlated spatial samples into a set of cosine basis values, concentrating much of a block’s visible information into relatively few low-frequency terms. Encoders reorder and quantize those values so zeros and small magnitudes compress efficiently with entropy coding. The transform is reversible at adequate precision; most quality loss enters during coefficient quantization and later rounding. It appears in still-image and interframe-video pipelines as one stage among prediction, rate control, and coding.
Key facts
- Baseline JPEG operates on small sample blocks and separates a block-average DC term from detail-bearing AC terms; differential and run-length coding exploit their typical distributions.
- The transform itself does not require loss: irreversible quality reduction arises mainly when coefficients are coarsely quantized, while lossless codecs generally use different reversible paths.
- Independent block processing can expose grid boundaries at low bitrates, while aggressive suppression of high frequencies produces ringing near edges and removes fine texture.
When DCT Compression matters
Select quantization settings by balancing file size against blocking, ringing, and lost fine detail. Compatibility matters because JPEG and many block-based video codecs use related but distinct DCT schemes.