File & format glossary

Clear, plain-English definitions of the document, image and encoding terms that come up when working with files — the words behind “lossless,” “DPI,” “alpha channel,” “base64” and the rest.

Lossless compression

A way of shrinking a file that preserves every bit of the original data, so it can be perfectly reconstructed. PNG, ZIP and FLAC are lossless; nothing is discarded, but files stay larger than lossy equivalents.

Lossy compression

A way of shrinking a file by permanently discarding detail the human eye or ear is least likely to notice. JPG and MP3 are lossy — much smaller files, at the cost of some irreversible quality loss.

Raster image

An image stored as a fixed grid of colored pixels, such as JPG, PNG or WebP. Enlarging a raster image beyond its native resolution makes it look blocky or blurry because there are not enough pixels to fill the space.

Vector image

An image stored as mathematical shapes and paths rather than pixels, such as SVG. Because it is described by geometry, it can be scaled to any size while staying perfectly sharp.

Alpha channel

An extra layer of information in an image that records transparency for each pixel — how see-through it is. Formats like PNG and WebP have an alpha channel; JPG does not, which is why it cannot store a transparent background.

Transparency

The ability of parts of an image to be fully or partially see-through, so whatever is behind them shows. Essential for logos and cut-out subjects, transparency is stored in an alpha channel and supported by PNG and WebP but not JPG.

DPI (dots per inch)

A measure of print resolution — how many dots of ink are packed into each inch of paper. Higher DPI means finer detail in print; 300 DPI is a common standard for sharp printed images.

PPI (pixels per inch)

How many pixels are displayed per inch on a screen, the on-screen counterpart to DPI. It describes pixel density, which affects how sharp an image looks on a given display.

Resolution

The pixel dimensions of an image — its width and height in pixels, such as 1920×1080. More pixels means more detail and larger files, but also more room to display or print an image at a larger size without it looking soft.

Aspect ratio

The proportional relationship between an image’s width and height, such as 16:9 or 4:3. Keeping the aspect ratio fixed while resizing prevents the image from looking stretched or squashed.

Compression ratio

How much a file has been shrunk, expressed as the original size relative to the compressed size (for example 10:1). A higher ratio means a smaller file, which for lossy formats usually means more quality traded away.

Metadata

Data about a file that is not its main content — author, creation date, software used, camera settings, location. It is stored alongside the visible content and often survives edits unless deliberately stripped.

EXIF

The metadata standard used by digital cameras and phones to embed details inside a photo — camera model, exposure, date, and often GPS location. Re-encoding or compressing an image usually removes EXIF data, which can be good for privacy.

Color space

The defined range of colors a file can represent, such as sRGB or CMYK. sRGB is the standard for screens and the web; CMYK is used for print. Mismatched color spaces are a common cause of colors looking different in print versus on screen.

Bit depth

How many bits are used to describe the color of each pixel. Higher bit depth allows more distinct colors and smoother gradients — 8-bit gives 256 levels per channel, while 16-bit gives far more, reducing visible banding.

Base64

An encoding that represents binary data using only plain text characters, so it can travel safely through text-only systems like email and JSON. It is not encryption; it makes data larger (about 33%) and is trivially reversible.

Character encoding

The system that maps characters to the bytes stored in a file, such as UTF-8. Using the wrong encoding is why text sometimes appears as garbled symbols; UTF-8 is the modern default that covers virtually all languages.

UTF-8

The dominant character encoding on the web and in modern files, capable of representing every character in the Unicode standard. Saving text files as UTF-8 avoids most “garbled character” problems across languages and programs.

Hash

A fixed-length fingerprint calculated from data by a hash function. The same input always produces the same hash, and changing even one character changes it completely — useful for verifying that a file has not been altered.

Checksum

A small value derived from a file (often a hash) used to detect whether the file was corrupted or changed. Comparing a downloaded file’s checksum with the published one confirms it arrived intact.

OCR (optical character recognition)

Technology that reads the text inside an image or scanned page and turns it into real, selectable, searchable text. Without OCR, a scanned PDF is just a picture of words that cannot be selected or edited.

PDF/A

A stricter version of PDF designed for long-term archiving. It requires everything the document needs — fonts, color profiles, images — to be embedded, so the file will render identically decades from now, with no external dependencies.

RGB

A color model that builds colors by mixing red, green and blue light, used by screens and digital images. sRGB is the most common RGB color space for the web.

CMYK

A color model used in printing that mixes cyan, magenta, yellow and black ink. Because it produces color differently from screen-based RGB, some bright on-screen colors cannot be reproduced exactly in print.

Compression artifact

A visible flaw introduced by lossy compression — blocky patches, smudges or halos around sharp edges and text. Artifacts become more pronounced the harder an image is compressed, and are why JPG suits photos but not crisp graphics.

Bitrate

The amount of data used per second of audio or video, usually measured in kilobits per second (kbps). Higher bitrate generally means better quality and larger files; it is the audio/video equivalent of an image’s quality setting.

Container format

A file format that wraps together several kinds of data — for example video, audio and subtitles — into one file. MP4 and HEIC are containers; the container is separate from the codecs that compress the content inside it.

Codec

The method used to compress and decompress audio, video or image data (the name is short for coder-decoder). A container like MP4 can hold content compressed by different codecs, such as H.264 or H.265.

Rendering

The process of turning a file’s underlying instructions into the visible result you see — for example drawing an SVG’s shapes or laying out a PDF’s pages on screen. The same file can render slightly differently across programs.

Flatten

To merge separate layers or elements of a document or image into a single fixed layer. Flattening a PDF or image bakes annotations and overlays into the page so they can no longer be edited or removed individually.