Learn
Image File Formats Explained
If you work with images in any capacity, you have encountered the moment: you need to export, convert, or deliver a file and the format menu offers a dozen options. Some are familiar, some obscure, and the wrong choice can mean bloated file sizes, lost transparency, degraded quality, or a file your recipient cannot open.
This guide covers the major image formats beyond camera-specific ones (for RAW, JPEG, and HEIF, see RAW vs. JPEG vs. HEIF). The goal is practical: what each format does well, where it falls short, and which one to reach for in common situations.
Lossy vs. lossless compression
Before diving into individual formats, it helps to understand the two broad approaches to making image files smaller.
Lossless compression reduces file size without discarding any image data. When you decompress the file, you get back every pixel exactly as it was. Think of it like a zip archive for image data: the information is reorganised more efficiently, but nothing is removed. PNG and TIFF both use lossless compression.
Lossy compression achieves much smaller files by permanently removing information the algorithm judges to be less perceptible. A lossy encoder analyses the image and discards fine details, subtle colour transitions, or high-frequency texture data that most viewers would not notice at normal viewing distances. The result is a file that looks nearly identical to the original but is a fraction of the size. The tradeoff is irreversible: once the data is discarded, you cannot recover it. JPEG is the classic lossy format; WebP and AVIF offer lossy modes as well.
The distinction matters most when you are working with source files. If you repeatedly open, edit, and re-save a lossy file, each save cycle discards more data, and quality degrades. This is why designers and photographers keep lossless originals and export lossy versions only as a final step.
PNG: the workhorse for graphics
PNG (Portable Network Graphics) uses lossless compression and supports full alpha transparency, meaning individual pixels can be partially or fully transparent. This makes it the default choice for screenshots, UI elements, logos on coloured backgrounds, graphics with text, and any image where crisp edges and exact colour reproduction matter.
For photographs, PNG files tend to be very large, because the complex colour gradients in photos do not compress efficiently with lossless methods. A photograph saved as PNG might be five to ten times the size of a visually equivalent JPEG. For that reason, PNG is rarely the right choice for photo delivery, though it is perfectly appropriate for archiving a photo if you want a lossless copy in a widely supported format.
PNG supports 8-bit and 16-bit colour depth, indexed colour (for smaller palette-based images), and interlacing for progressive loading. It does not support animation in its base specification, though the APNG extension adds animation support and is now widely supported in browsers.
If you are managing design assets, icons, or UI components, PNG is likely the format you encounter most. Keeping those assets organised and searchable saves time when you need to find the right version of a logo or illustration.
TIFF: the archival and print standard
TIFF (Tagged Image File Format) is a flexible, lossless format that supports high bit depths (16-bit and 32-bit per channel), multiple layers, and various colour spaces including CMYK. It is the standard delivery format in print production, medical imaging, scientific imaging, and archival work.
TIFF files can be very large. An uncompressed TIFF of a high-resolution photograph might run to hundreds of megabytes. TIFF does support internal compression (LZW or ZIP, both lossless), which reduces file size somewhat, but the resulting files are still substantially larger than lossy alternatives.
The format's strength is its fidelity and flexibility. A TIFF can store far more colour information than most display technologies can show, which is precisely the point for print workflows where the file needs to survive colour space conversions, editing, and output to high-end printers. If a print shop asks for TIFF delivery, they are asking for maximum quality and editability.
For anyone managing large TIFF archives, cloud storage with enough capacity and proper organisation becomes essential. A single photo shoot delivered as TIFFs can consume gigabytes.
SVG: vectors that scale infinitely
SVG (Scalable Vector Graphics) is fundamentally different from the formats above. Rather than storing a grid of pixels, SVG describes shapes, paths, and text using mathematical instructions in XML. A circle is defined by its centre coordinates and radius, not by thousands of coloured pixels arranged in a circular pattern.
This means SVG images scale to any size without quality loss. A logo rendered as SVG looks equally sharp on a phone screen and a billboard. SVGs also tend to be very small files, since a few lines of code can describe shapes that would require millions of pixels in a raster format.
SVG is ideal for logos, icons, illustrations, diagrams, charts, and any graphic composed of defined shapes and lines. It is not suitable for photographs or images with complex, continuous-tone detail; those require raster formats.
Because SVGs are essentially code, they can be styled with CSS and manipulated with JavaScript, making them particularly useful in web development. They also work well as brand assets because a single SVG file can serve every size requirement without maintaining multiple raster exports.
WebP: Google's modern format
WebP was developed by Google as a more efficient alternative to both JPEG and PNG for web use. It supports lossy compression (achieving files roughly 25 to 35 per cent smaller than equivalent-quality JPEGs), lossless compression (producing files smaller than PNGs), animation, and alpha transparency.
Browser support for WebP is now effectively universal across modern browsers. It has become a common choice for web images, and many content management systems and image CDNs automatically convert uploads to WebP for delivery.
The format's main limitation is that it is less universally supported in desktop software, print workflows, and older applications. If you need to deliver files to a print shop or embed images in a document that will be opened in various applications, JPEG or PNG remains safer. For web publishing, WebP is an excellent default.
AVIF: the newest contender
AVIF (AV1 Image File Format) is based on the AV1 video codec and offers even better compression than WebP, particularly at low to medium quality settings where it preserves detail and avoids the blockiness that JPEG shows at high compression. AVIF also supports HDR (high dynamic range), wide colour gamuts, and both lossy and lossless compression.
The tradeoff is speed. AVIF encoding is significantly slower than JPEG or WebP encoding, which matters for workflows that process large numbers of images. Decoding is faster but still somewhat slower than JPEG. Browser support has grown steadily and covers Chrome, Firefox, Safari, and Edge in their current versions, but older browser versions and many native applications do not yet handle AVIF.
For content creators publishing to the web, AVIF offers a meaningful file size reduction. The practical approach is to encode AVIF with a JPEG or WebP fallback for older clients, which is what most image CDNs do automatically.
GIF: animation's legacy format
GIF (Graphics Interchange Format) dates to 1987 and remains in use primarily because of animation. A GIF can store multiple frames in a single file, playing them in sequence to create simple animations, which made it the format of choice for animated images on the web for decades.
GIF's limitations are severe by modern standards. It supports only 256 colours per frame, which produces visible banding and dithering in anything beyond simple graphics. Its compression is lossless but inefficient compared to modern alternatives. Animated GIFs tend to produce large files with poor visual quality.
For static images, GIF has been obsolete for years; PNG does everything GIF does, better. For animation, newer options are steadily replacing it: APNG offers full-colour animated images, WebP supports animation with better compression, and short video formats (MP4 with H.264 or H.265) are far more efficient for anything beyond a few frames. Social platforms and messaging apps increasingly use video-based formats internally even when users think they are sharing "GIFs."
BMP: the uncompressed relic
BMP (Bitmap) stores pixel data with no compression at all (or minimal RLE compression). The result is very large files with no quality advantage over PNG. BMP exists primarily as a legacy format from early Windows computing. You may encounter it in very old archives or in specialised hardware and software that predates widespread PNG support. In nearly every modern context, there is no reason to use or keep BMP files; converting them to PNG gives you the same quality in a smaller file.
EPS and PDF as image containers
EPS (Encapsulated PostScript) and PDF (Portable Document Format) are not purely image formats, but they frequently serve as containers for images in print and design workflows.
EPS was long the standard for delivering vector artwork to print. A logo or illustration saved as EPS could be placed into a page layout at any size. EPS has been largely superseded by PDF and native vector formats (AI, SVG), but you may still encounter it in older asset libraries and some print workflows.
PDF can contain both vector and raster content, supports colour management, and preserves layout across platforms. It is widely used for delivering print-ready files, multi-page documents with embedded images, and proofing. A PDF exported from a design application at press quality can serve as a reliable delivery format for design projects.
For a broader look at formats specific to design applications, see Design File Formats Explained.
Choosing the right format by use case
The right format depends on what you are doing with the image. Here are the most common scenarios.
For web publishing, WebP or AVIF with JPEG fallback gives the best balance of quality and file size. PNG remains appropriate for graphics with transparency or crisp edges. SVG is ideal for logos, icons, and illustrations.
For print production, TIFF is the standard for photographic images. PDF is preferred for layouts and composed pages. EPS or PDF works for vector artwork, though native design files are increasingly accepted. Print workflows rely on colour space accuracy, so format choice intersects with colour management.
For archival, TIFF and PNG are the most common lossless choices. TIFF offers greater flexibility (higher bit depths, more colour space options), while PNG is more compact and widely supported. The key is to keep a lossless original from which you can generate any derivative format later.
For UI and app design, PNG and SVG dominate. PNG for raster elements (icons at specific sizes, screenshots, complex graphics) and SVG for scalable elements (logos, simple icons, illustrations). Keeping a well-organised library of design files in these formats means assets are ready for any screen density or platform.
For social media, each platform has its own preferred formats and will typically re-encode whatever you upload. JPEG and PNG are universally accepted. The platform's own re-encoding means there is limited benefit to uploading in cutting-edge formats; what matters more is uploading at sufficient resolution with appropriate PPI settings.
For email, JPEG and PNG are the safest choices for inline or attached images, since email clients vary widely in format support. Keep file sizes reasonable; large attachments cause delivery issues. For sharing larger files, a secure sharing link is more reliable.
Format conversion: what you lose and what you keep
Converting between formats is routine, but understanding what happens during conversion prevents unpleasant surprises.
Converting from lossless to lossy (for example, PNG to JPEG) discards data permanently. The resulting JPEG may look fine, but you cannot recover the discarded detail. Always keep the lossless original.
Converting from lossy to lossless (JPEG to PNG) does not recover lost data. The PNG will be a lossless copy of the already-degraded JPEG, and it will be a larger file. This conversion preserves what remains but does not improve quality.
Converting between colour modes (RGB to CMYK, or between colour spaces) can shift colours. This is expected and necessary for print workflows, but it means the converted file may not look identical to the original on screen.
Stripping metadata during conversion is common and sometimes intentional (for privacy when publishing online) but can also remove valuable information like colour profiles, copyright data, and camera settings. Understanding what metadata your images carry helps you make informed decisions about what to preserve.
The safest practice is to keep your original files in whatever format they were created in and generate converted copies as needed. A workspace that stores originals alongside derivatives, with search that can find either, makes this manageable even across large libraries.
Future formats: JPEG XL and the format wars
JPEG XL deserves mention as the most technically ambitious new format. It offers excellent compression in both lossy and lossless modes, supports progressive decoding, wide colour gamuts, HDR, animation, and can losslessly recompress existing JPEG files to roughly 20 per cent smaller. Its feature set makes it a strong candidate to replace JPEG, PNG, and GIF simultaneously.
However, JPEG XL's adoption has been complicated. Google removed JPEG XL support from Chrome in 2023, favouring its own WebP and AVIF formats, while Apple added support in Safari. The format remains in a state of uncertain adoption, technically superior but lacking universal browser support.
The broader pattern is one of fragmentation. Where JPEG once served as a near-universal standard, we now have multiple competing formats with different strengths. The practical response is to store originals in high-quality formats and let your publishing tools handle format conversion for delivery. Most modern web platforms and CDNs can serve the optimal format automatically based on the viewer's browser.
Frequently asked questions
Which image format has the best quality?
Quality depends on the use case, not the format alone. For preserving every detail of the original, TIFF at high bit depth is the most capable format. For web delivery, AVIF and WebP produce the best visual quality at small file sizes. SVG preserves vector artwork at infinite resolution. The key question is whether you need lossless preservation or visually acceptable compression.
When should I use PNG instead of JPEG?
Use PNG when you need transparency, when the image contains text or sharp edges (like a screenshot, diagram, or logo), or when you need a lossless copy. Use JPEG for photographs and images with smooth gradients where some compression is acceptable and file size matters.
Is WebP better than JPEG?
WebP produces smaller files at equivalent visual quality, supports transparency and animation, and offers a lossless mode. For web use, it is a better choice in most situations. For print delivery, professional workflows, and maximum compatibility with older software, JPEG remains more widely supported.
Can I convert a JPEG to PNG to improve quality?
No. Converting a JPEG to PNG wraps the already-compressed data in a lossless container, which preserves what is there but does not recover detail that JPEG compression removed. The PNG will also be a larger file. To get a high-quality PNG, you need to export from the original source file, not from a JPEG derivative.
What format should I use for logos?
SVG is the best format for logos in most contexts. It scales to any size without quality loss and produces very small files. For situations where a raster format is required, export PNG at the specific dimensions needed. Keep the vector original (SVG, AI, or EPS) as your source file so you can generate raster exports at any size in future.
Why are my TIFF files so large?
TIFF stores image data at full quality, often at high bit depths (16-bit per channel rather than 8-bit), and may include multiple layers or colour space information. An uncompressed 16-bit TIFF of a 50-megapixel photograph can easily exceed 300 MB. Using LZW or ZIP compression within TIFF reduces file size somewhat, but TIFFs will always be larger than lossy alternatives. This is a feature, not a flaw: the size reflects the data being preserved.
Is GIF still relevant?
For static images, no. PNG replaced GIF for static graphics years ago. For animation, GIF persists through cultural inertia and universal support, but it is technically inferior to APNG, animated WebP, and short video formats. If you are creating new animated content, consider these alternatives.
What is the difference between raster and vector formats?
Raster formats (PNG, JPEG, TIFF, WebP, AVIF, GIF, BMP) store images as grids of coloured pixels. They have fixed dimensions and lose quality when scaled up. Vector formats (SVG, EPS, PDF with vector content) store images as mathematical descriptions of shapes. They scale to any size without quality loss but cannot represent photographs or complex continuous-tone imagery.
Should I keep original files or just the exported versions?
Keep originals. Original files, whether RAW photographs, layered PSD or AI files, or high-resolution TIFFs, contain the most data and give you the most flexibility for future use. Exported versions (JPEGs for web, PNGs for specific sizes) can always be regenerated from the original, but the reverse is not possible.
What image format should I use for email attachments?
JPEG for photographs and PNG for graphics with transparency or text. Keep file sizes under a few megabytes per image for reliable delivery. If you need to send higher-resolution files or many images at once, a sharing link is more practical than email attachments.