PGX Converter

PGX is a simple single-component raster image format defined as part of the JPEG 2000 standard (ISO/IEC 15444) for use in conformance testing and verification of JPEG 2000 codec implementations. Introduced around 2000 alongside the JPEG 2000 specification itself, PGX files store a single image component (one color channel or grayscale plane) with a text header followed by raw pixel data, providing an unambiguous reference representation against which encoder and decoder outputs can be compared sample by sample. The header is a single ASCII line specifying endianness (ML for big-endian, LM for little-endian), signedness (+ for unsigned, - for signed), bit depth (1 to 32 bits), width, and height. The pixel data follows as raw binary values, each occupying the minimum number of bytes needed for the specified bit depth, with one value per pixel. For multi-component images (like RGB), each component is stored in a separate PGX file. The format's deliberate simplicity — no compression, no metadata, no multi-channel support — ensures there are no ambiguities in interpretation that could mask codec bugs. One advantage is verification precision: PGX's uncompressed, exactly-specified representation allows bit-exact comparison of decoded JPEG 2000 output against reference images, essential for certifying that a codec implementation conforms to the standard. The format's role in the JPEG 2000 conformance testing framework means it is implemented by every serious JPEG 2000 codec (OpenJPEG, Kakadu, etc.) and used in the official ISO conformance test suite. PGX files can also be processed by ImageMagick and various JPEG 2000 development tools.