DCT Coefficient Viewer

JPEG compression divides an image into 8×8 pixel blocks and applies the Discrete Cosine Transform to each one. This converts spatial pixel data into frequency coefficients — the DC component captures average brightness, while AC coefficients capture edges and texture. The DCT Viewer recomputes this transform and visualizes the coefficient distribution. When a JPEG is edited and re-saved, the second compression pass creates periodic artifacts in the coefficient histogram — a telltale sign of manipulation. For compression-level error analysis, combine with the ELA Scanner.

An unedited photo from a camera is compressed exactly once. When someone opens a JPEG, modifies it, and saves it again, the second compression quantizes already-quantized coefficients. This creates a characteristic comb-like pattern in the histogram — peaks at multiples of the new quantization step. The periodicity score quantifies this pattern. A score above 2.5 strongly indicates re-compression. For JPEG quality-level splice detection across image regions, use the JPEG Ghost Scanner.

AC Energy shows the sum of absolute AC coefficients per block as a heatmap. Blocks with different compression history stand out in energy. DC Component visualizes the average brightness of each 8×8 block — useful for seeing the overall luminance structure. Blockiness highlights discontinuities at 8×8 block boundaries — stronger blockiness in some regions may indicate a pasted element compressed at a different quality. For lighting direction analysis, try the Luminance Gradient Scanner.

Forensic analysts use DCT analysis to prove an image was re-saved after editing — a key piece of evidence in manipulation cases. Journalists verify that press photos haven't been re-compressed, which would indicate post-capture editing. Insurance investigators detect doctored claim photos. Researchers studying JPEG compression artifacts use it to understand quantization behavior. For metadata-level verification, combine with the Authenticity Checker. For a complete forensic workflow, visit the photo forensics hub.