Lossless compression is a class of file encoding that reduces storage size while preserving every bit of the original image data, allowing a mathematically perfect reconstruction of the source when the file is decoded. Unlike lossy compression, which permanently discards information, lossless schemes work by finding statistical redundancy in the pixel values and representing repeating patterns more efficiently. The result is smaller than an uncompressed file but typically larger than a lossy equivalent of the same image.
The format most photographers encounter first is DNG, which offers an optional lossless mode when ingesting or converting raw files. PNG, TIFF (with LZW or ZIP packing), and most native manufacturer raw formats from Canon, Nikon, Sony, and Fujifilm also use lossless schemes by default. A 14-bit lossless compressed raw from a 45-megapixel sensor will typically shrink to roughly 50 to 70 percent of its uncompressed size, depending on scene content. Flat, smooth subjects like skies compress more efficiently than busy ones full of foliage or fine texture, because predictable values are easier to encode in fewer bits.
Common algorithms include Huffman coding, run-length encoding, LZW, ZIP/DEFLATE, and predictive coding tuned for image data. Manufacturer raw codecs often combine a predictor with entropy coding so that only the difference between a predicted value and the actual value is stored. Because no quantization is performed, repeated saves do not compound damage the way they do with JPEG; an image can be opened, edited, and re-saved indefinitely without degradation.
For archival masters, lossless is the responsible default. Raw originals, master TIFFs out of Photoshop, and DNG conversions retain full dynamic range, bit depth, and tonal range, leaving room for future re-edits as software improves. The tradeoff is storage: a single 100-megapixel medium-format file can occupy 120 megabytes or more even after compression, which is why disciplined backup and ingest workflows matter as libraries grow.
Some cameras offer a “lossless compressed” raw alongside “compressed” and “uncompressed” choices. The lossless option is almost always the right pick: identical image quality to uncompressed, with file sizes a third to half smaller and faster buffer flushing during burst mode shooting. The only reason to choose uncompressed is when older third-party software cannot decode the manufacturer’s lossless variant, a problem that has largely faded as Adobe, Capture One, and other tools have caught up.
Confusion sometimes arises around “visually lossless” labels used by formats like JPEG XL, HEIC, or some video codecs. Visually lossless is not the same as mathematically lossless: it means losses fall below the threshold of human perception under typical viewing, but the file no longer decodes to bit-identical pixels. For master files intended for printing, heavy editing, or long-term archive, only true lossless qualifies. For final delivery on the web, visually lossless or carefully tuned lossy output is usually the more practical choice.