What Is Compression?

Compression tells you how the data inside your file was squeezed down, or whether it was squeezed at all. It is the reason two files holding the same picture, song or video can differ enormously in size.

Lossless and lossy compression

Compression is the reduction of data so that a file needs fewer bits than the original, and it comes in two families.

Lossless compression looks for repetition and statistical redundancy and writes the same information down more efficiently. Nothing is thrown away, so unpacking gives back every original bit. ZIP archives work this way, and so do PNG images, FLAC audio and the LZW and Deflate methods used inside TIFF files.

Lossy compression goes further and discards information judged to be least important: color detail the eye is poor at noticing, sound you were unlikely to hear anyway. The saving is far larger, and it is permanent. The difference is visible in an image or a video and audible in an audio file when you compare it with the original. JPEG, MP3, AAC and H.264 all work this way. Because the loss is permanent, it also stacks: open a lossy file, edit it, save it again, and the damage adds up with every round.

What the value looks like

Either the field states that nothing was compressed, with a value such as Uncompressed or none, or it names the method that was used, spelled the way the file format spells it. Values you are likely to meet include:

• JPEG and JPEG (old-style), the two ways JPEG compression is recorded
• LZW, Deflate and PackBits, the lossless methods found inside TIFF files
• manufacturer-specific entries in raw camera files, such as Nikon NEF Compressed

One file can show more than one Compression row, because a picture file often carries a small preview thumbnail alongside the full-size image, and each is compressed in its own right. In video files the compression is normally named in the codec rows instead of in a field called Compression.

Why it matters

This is the trade between file size and quality, and knowing which side a file sits on tells you what you can still do with it. Lossy is right for sharing, sending and streaming. Lossless is right for editing and for anything you intend to keep as a master, because you can re-save it as often as you like without it degrading. If a file you are archiving reads as lossy, it is not really a master, and no later step can restore what was dropped.

It answers compatibility questions too. Some older programs and some print workflows expect a plain uncompressed TIFF and will refuse or mishandle one compressed with LZW or PackBits, so this row explains a file that opens in one place and not another.

Compression describes how the data is stored. It holds no information about a person, a place or a device.