What is PGM? Portable Graymap Format Explained

How PGM Works

PGM, the Portable Graymap format, stores a single channel grayscale image. Each pixel is one integer that represents intensity, running from zero, meaning black, up to a declared maximum that means white. There is no color and no transparency, only shades of gray, which makes it the natural middle sibling between the black and white PBM and the full color PPM.

The maxval, or maximum value, is part of the header and defines the range of every pixel. A maxval of 255 gives standard 8-bit grayscale with 256 possible shades, while a larger maxval, up to 65535, allows 16-bit precision for applications that need finer tonal detail than a byte can hold. A reader uses maxval to interpret the raw numbers correctly.

A PGM header consists of a two character magic number, then the width, the height, and the maxval, all written as ASCII text separated by whitespace, followed by the pixel data laid out row by row. Because these fields are explicit, a program always knows the image size, the bit depth, and where the data begins.

Like the rest of the family, PGM comes in two variants. The plain form uses magic number P2 and writes each pixel value as readable ASCII text, while the raw form uses magic number P5 and stores the values as compact binary bytes. Both describe the same image; the raw form is simply smaller and faster to read.

History and Standardization

PGM comes from the same late 1980s toolkit created by Jef Poskanzer, originally called pbmplus and later carried forward as the Netpbm project. It was designed as a grayscale companion to the black and white PBM, using the same header conventions so that the family would feel consistent.

Together with PBM for one bit images and PPM for full color, PGM forms the trio unified under the name PNM, the Portable Anymap family. The three differ only in how much information each pixel carries, which lets tools handle them with shared logic.

Rather than being governed by a formal standards organization, PGM is defined by a short and clear specification that has remained stable for decades. That stability, plus how easy the format is to implement, is why it is trusted as a plain interchange target in image processing work.

PGM Compared to Its Siblings and Real Formats

Compared to its siblings, PGM sits between PBM and PPM. PBM can only store pure black or white, so it cannot express shading, while PPM stores three color channels per pixel and therefore uses roughly three times the data of the equivalent grayscale image. If your image is genuinely grayscale, PGM is the right sized choice among the three.

Compared to practical formats such as PNG and TIFF, PGM applies no compression, so its files are large relative to their content. PNG offers lossless compression and broad support, and TIFF offers flexible compression along with rich metadata, both of which are better for storing or distributing real grayscale images.

The honest tradeoff, as with the whole family, is simplicity against size. PGM is extremely easy to read and write, which is exactly why it is so common in image processing research and command line pipelines, where a program may generate intermediate images that are never meant to be stored long term. For final delivery of grayscale imagery, a compressed format is usually the better choice.