Image Compression Glossary

JPEG Joint Photographic Experts Group

A lossy image compression format standardised in 1992. Achieves small file sizes by permanently discarding image data the human eye is least sensitive to, using discrete cosine transform (DCT) encoding in the YCbCr colour space. Does not support transparency. Universally supported across all browsers, operating systems, and upload portals. Best suited for photographs and images with gradual colour transitions.

Typical use: Photographs, web photos, camera images, social media uploads — anywhere small file size matters more than perfect quality.

Compress a JPEG →

PNG Portable Network Graphics

A lossless image format that preserves every pixel exactly. Uses deflate compression on the raw pixel data — no information is discarded. Supports an alpha channel for full or partial transparency. Produces larger files than JPEG for photographs but is the correct choice for graphics, logos, icons, and screenshots where pixel accuracy and transparency matter.

Typical use: Logos, icons, screenshots, graphics with transparency, UI elements — when pixel-perfect quality or transparency is needed.

Compress a PNG →

WebP Web Picture format

A modern image format developed by Google, released in 2010. Supports both lossy and lossless compression modes, and supports transparency in both modes. Lossy WebP typically produces 25–35% smaller files than JPEG at equivalent visual quality. Lossless WebP is typically 26% smaller than PNG. Supported in all major browsers since 2020.

Typical use: Modern websites, web apps, mobile sites — smaller alternative to JPEG and PNG where browser and portal support allows.

Compress a WebP →

AVIF AV1 Image File Format

A modern image format derived from the AV1 video codec, standardised in 2019. Typically achieves 20–50% smaller files than JPEG at equivalent visual quality. Supports transparency, HDR, and wide colour gamut. Encoding is computationally intensive compared to JPEG and WebP. Browser support: Chrome 85+, Firefox 93+, Safari 16.1+.

Typical use: High-efficiency web images, HDR photos — 20–50% smaller than JPEG at similar quality, where encoding speed is not a constraint.

Compress an AVIF →

HEIC High Efficiency Image Container

The container format used by Apple for photos captured on iPhone and iPad, introduced with iOS 11 in 2017. Uses HEVC (H.265) compression and typically produces files half the size of JPEG at equivalent quality. Browser and OS support is limited compared to JPEG — not supported in Firefox.

Typical use: Apple Photos (iPhone/iPad default since iOS 11) and some Android and camera devices — roughly half the file size of JPEG.

Compress a HEIC →

HEIF High Efficiency Image File Format

The file format standard on which HEIC is based, developed by the Moving Picture Experts Group (MPEG). HEIF defines the container structure; HEIC is Apple's specific implementation using HEVC compression. The two terms are often used interchangeably in consumer contexts, though they are technically distinct.

Typical use: The standard container format that HEIC is based on — used by Apple and some other devices.

KB Kilobyte

A unit of digital file size equal to 1,000 bytes under the SI (decimal) definition used by storage manufacturers, operating systems, and most file size specifications. Distinct from kibibyte (KiB), which equals 1,024 bytes under the IEC binary definition. Upload portals and file size requirements universally use the decimal definition.

Typical use: Measuring small file sizes — passport photos, thumbnails, upload limits.

MB Megabyte

A unit of digital file size equal to 1,000,000 bytes (1,000 KB) under the SI decimal definition. Distinct from mebibyte (MiB), which equals 1,048,576 bytes. A typical smartphone photograph saved as JPEG is 2–6 MB. A typical passport photo requirement is 20–200 KB — roughly 10–100 times smaller than a full-resolution phone photo.

Typical use: Measuring larger photo and document file sizes.

Lossy compression

A data compression method that permanently removes information to achieve smaller file sizes. The removed data cannot be recovered — decompressing a lossy file does not restore the original. JPEG, WebP (lossy mode), and AVIF are lossy formats. The degree of quality loss is controlled by a quality parameter: lower quality means smaller file and more visible artefacts. Repeated re-encoding of a lossy file compounds quality loss each time.

Typical use: JPEG, WebP (lossy mode), AVIF — when the smallest possible file size is needed and some quality loss is acceptable.

Lossless compression

A data compression method that reduces file size without discarding any information. The original data is perfectly reconstructed on decompression. PNG and WebP (lossless mode) are lossless formats. Lossless compression works by finding more efficient ways to encode the same data — for example, replacing repeated byte sequences with shorter references. Cannot achieve as small a file size as lossy compression for photographic content.

Typical use: PNG, WebP (lossless mode) — when every pixel must stay exact, such as graphics, medical images, and archives.

EXIF Exchangeable Image File Format

A metadata standard embedded inside image files, defined by the Japan Electronics and Information Technology Industries Association (JEITA). Stores information including camera make and model, lens focal length, aperture, shutter speed, ISO, GPS coordinates, date and time, and image orientation. The orientation tag (values 1–8) records the physical rotation of the camera at capture so software can display the image correctly.

Typical use: Camera metadata stored inside photos — date, GPS coordinates, orientation, and camera settings.

Megapixel (MP)

A unit of image resolution equal to one million pixels. Calculated by multiplying image width by height in pixels: a 4,000 × 3,000 pixel image is 12 megapixels. Megapixels measure spatial resolution — the amount of detail captured — not file size. A 12 MP image can be stored as a 500 KB JPEG or a 12 MB TIFF depending on compression. Typical smartphone cameras produce 12–50 MP images.

Typical use: Describing camera and image resolution — for example, a 12 MP smartphone photo is 4,000 × 3,000 pixels.

MozJPEG Mozilla JPEG Encoder Project

An open source JPEG encoder developed and maintained by Mozilla. Produces smaller files than the standard libjpeg encoder at the same quality level by using progressive encoding, trellis quantization, and optimal Huffman coding. Progressive JPEG encodes the image in multiple passes of increasing detail, allowing browsers to display a low-quality preview before the full image loads. Released under a permissive open source licence (BSD/MIT/zlib).

Typical use: Better JPEG encoding for websites — produces smaller JPEG files than standard encoders at the same visual quality.

Try MozJPEG compression →

Oxipng

An open source lossless PNG optimiser written in Rust by Joshua Holmer. Reduces PNG file size by trying multiple deflate compression strategies and row filter combinations to find the smallest valid PNG encoding of the same pixel data. Supports optimization levels 1–6; higher levels try more strategies and take longer. Produces no change in image quality — the output is pixel-identical to the input. Released under the MIT licence.

Typical use: Optimising PNG files without any quality loss — used in web performance tools and build pipelines.

Try Oxipng compression →