Every photographer makes decisions about file formats constantly, yet most of us operate on habit, hearsay, or whatever our camera defaulted to when we first pulled it out of the box. The problem is that choosing the wrong format at the wrong stage of your workflow can quietly destroy your editing flexibility, balloon your storage needs, break compatibility with clients and labs, or degrade your images in ways you won’t notice until it’s too late. This guide breaks down what each major format actually does under the hood, what it preserves, what it throws away, and when you should reach for each one across capture, editing, delivery, and long-term archiving.

Understanding What Formats Actually Control

Before diving into specific formats, it helps to understand the fundamental variables that differentiate them. Compression is the big one, and it comes in three flavors. Uncompressed formats store every pixel at full fidelity, resulting in the largest possible files but zero quality degradation. Lossless compression uses mathematical tricks to reduce file size while preserving the ability to reconstruct the original data perfectly, similar to how a ZIP file works. Lossy compression permanently discards information the algorithm predicts you won’t miss, trading quality for dramatically smaller files.

Bit depth determines how many distinct tonal values each color channel can represent. 8-bit files offer 256 levels per channel, which translates to roughly 16.7 million possible colors and is perfectly adequate for final delivery. 16-bit files provide 65,536 levels per channel, giving you vastly more headroom for aggressive edits before you start seeing banding or posterization. The difference between 8-bit and 16-bit is invisible in a finished image, but it becomes critical the moment you start pushing shadows, shifting white balance, or doing heavy color grading.

Color space embedding determines whether a file carries information about how its colors should be interpreted. Formats that properly embed ICC profiles ensure your colors display consistently across different software and devices. Formats that strip or fail to preserve color profiles can cause unpredictable color shifts that seem to appear out of nowhere. Metadata preservation covers everything from EXIF data recording your camera settings and capture date to IPTC fields for copyright and captions to XMP sidecars tracking your edit history. Some formats preserve all of this faithfully while others strip it partially or completely.

Raw Files: Your Digital Negatives

Raw files aren’t images in any conventional sense. They’re the unprocessed data your sensor captured before the camera applied demosaicing, white balance, tone curves, or color processing. This makes them maximally flexible because no interpretive decisions have been baked in yet. Every manufacturer uses their own proprietary raw format, which is why you see .CR2 and .CR3 from Canon, .NEF from Nikon, .ARW from Sony, .RAF from Fujifilm, and so on.

What raw preserves is substantial: the full bit depth your sensor captured (typically 12 to 14 bits), linear or near-linear tonal data before any curve is applied, white balance stored as adjustable metadata rather than a permanent conversion, and the maximum dynamic range and color information your camera recorded. This is why raw gives you so much latitude to recover highlights, lift shadows, and shift colors without the image falling apart.

The tradeoffs are equally real. Raw files are large, often 25 to 60 megabytes or more depending on your camera’s resolution and the compression setting you choose. They require raw processing software for full-quality rendering and editing. Your operating system may display previews through built-in codecs, but these are limited and inconsistent across camera models, and they don’t give you access to the actual editing latitude raw provides. They’re completely unsuitable as delivery formats because clients, labs, and social platforms can’t use them. And the proprietary nature of each manufacturer’s format raises legitimate archival concerns. Will the software ecosystem still support your files in 30 years? Nobody knows for certain.

Many cameras offer lossless compressed raw, which shrinks files significantly without any quality penalty. Others provide only lossy compressed options, which sacrifice a small amount of data for even smaller files. In practice, the quality difference from lossy raw compression is rarely visible, but if storage isn’t a constraint and your camera offers both, lossless is the safer choice. Uncompressed raw is almost never worth the storage penalty for working photographers. Shoot raw for everything if your workflow supports it, keep your raw files as archival masters, but never attempt to deliver raw to anyone.

DNG: Adobe’s Answer to Format Fragmentation

Adobe created the Digital Negative format to solve the problem of dozens of proprietary raw formats that might become unreadable as software evolves. DNG is an open specification, meaning its structure is publicly documented and theoretically any developer can build tools to read it without licensing fees or reverse engineering.

The critical thing to understand about DNG is that it comes in several flavors with very different characteristics. Camera Raw DNG files are converted directly from proprietary raw formats and maintain essentially the same editing latitude as the original. Linear DNG files contain demosaiced data, making them slightly less flexible but still high quality. Lossy DNG files apply compression that discards some information, resulting in smaller files but reduced editing headroom. Knowing which type you’re creating matters enormously.

The archival case for DNG is straightforward: an open specification is less likely to become unreadable than a proprietary format tied to a single manufacturer’s continued support. DNG also embeds XMP metadata directly inside the file, eliminating the risk of losing sidecar files that travel separately from your images. You can even embed the original proprietary raw inside the DNG as a backup, though this obviously increases file size.

The arguments against DNG conversion are worth considering too. Converting adds time to your ingest workflow. Some photographers are uncomfortable with Adobe being the steward of the “open” specification. There’s a small but nonzero risk of conversion errors. And certain editing software handles native raw files better than converted DNG. If you trust your proprietary format’s longevity and your software works well with native raw, skipping DNG conversion is a reasonable choice. If you’re archiving for decades and want maximum insurance against format obsolescence, DNG conversion makes sense. The middle ground is making an archive of JPEGs. 

JPEG: Universal but Unforgiving

JPEG remains the universal language of digital images. Every device, browser, application, and operating system can open a JPEG without any special software or extensions. This ubiquity comes from a compression scheme that uses discrete cosine transform to divide your image into 8×8 pixel blocks and then discards high-frequency detail the algorithm predicts you won’t perceive. In mainstream photography workflows, JPEG is effectively 8-bit only, though the underlying standard does include higher bit depth modes used in some industrial and medical imaging contexts.

Many export tools use a quality scale from 0 to 100, though the exact mapping varies by encoder. As a rule of thumb, settings between 80 and 90 are usually visually transparent for web delivery. Below 70, compression artifacts become increasingly obvious depending on the content. Even maximum quality settings apply some compression. Lossless JPEG variants exist in the specification, but they’re not what cameras produce or what most software expects, so for practical purposes, standard JPEG is always lossy.

A less visible setting called chroma subsampling can significantly impact quality in ways most photographers never notice. At 4:4:4 subsampling, full color resolution is preserved. At 4:2:0, which many cameras and applications use by default, chroma is stored at quarter resolution relative to luminance. This is usually fine for continuous-tone photographs but can show up on sharp color edges, graphics, and text.

The generation loss problem is perhaps JPEG’s most important limitation. Every time you open a JPEG, make changes, and save it again, you’re recompressing already-compressed data and losing additional quality. The solution is simple: always edit from raw or lossless sources and export fresh JPEGs as needed. Never treat JPEG as a working format where you edit, save, edit again, and save again.

JPEG supports embedded color profiles, and sRGB remains the safest default for web delivery. It preserves EXIF, IPTC, and XMP metadata, though some social platforms and web services strip metadata on upload. Use JPEG for web delivery, client galleries, email attachments, social media, and any situation demanding universal compatibility. Avoid JPEG for working files during editing, for archiving masters you might want to re-edit later, and for anything requiring transparency.

TIFF: The Professional Interchange Standard

TIFF is a flexible container format that has been the standard for high-quality image interchange for decades. Many print labs and publications accept it, some prefer it for fine-art and prepress work, and it moves reliably between most major applications without quality loss.

The format supports multiple compression options. Uncompressed TIFF offers maximum compatibility but the largest files. LZW compression is lossless, widely supported, and provides moderate size reduction. ZIP compression is also lossless and slightly more efficient than LZW but occasionally causes compatibility problems with older software. JPEG compression inside a TIFF container exists but completely defeats the purpose of using TIFF in the first place.

TIFF supports both 8-bit and 16-bit depth. For working files you’ll continue editing, always use 16-bit to preserve maximum headroom. Drop to 8-bit only when creating final deliverables if file size is a concern. TIFF can also preserve Photoshop layers if you enable that option when saving, though this creates very large files. For complex layered work, PSD is usually more practical.

Color space embedding in TIFF is reliable, making it appropriate for color-managed print workflows. Labs typically expect TIFF files in sRGB or Adobe RGB depending on their specifications. Metadata support is comprehensive. Use TIFF for delivering to print labs and publications, for archiving processed masters alongside your raw originals, for moving files between applications when quality must be preserved, and for any workflow requiring 16-bit depth or guaranteed lossless quality. Avoid TIFF for web delivery since browsers won’t display it and files are unnecessarily large.

PNG: Built for Graphics, Not Photographs

PNG was designed for web graphics, not photographs. It uses lossless compression and supports transparency natively through an alpha channel, making it essential for logos, interface elements, and composites where you need to preserve edges against variable backgrounds.

PNG comes in several bit depth configurations. PNG-8 supports only 256 colors like GIF, producing tiny files but with severely limited usefulness for photographic content. In common usage, people often refer to “PNG-24” for standard RGB images and “PNG-32” for RGBA images with an alpha channel for transparency, though these aren’t official format designations. PNG also natively supports 16 bits per channel, which is why it sees use in VFX and 3D workflows where lossless high-bit-depth output matters.

The fundamental problem with using PNG for photographs is that lossless compression is inefficient for complex photographic detail. A photograph saved as PNG will be dramatically larger than the same image saved as a high-quality JPEG with no perceptible quality benefit. While PNG does support 16-bit depth and major applications like Photoshop handle it correctly, browser support for 16-bit PNG remains inconsistent, and many web-focused pipelines will downconvert to 8-bit silently. PNG does support embedded color profiles and metadata, but handling varies significantly across browsers and applications. Many assume sRGB regardless of what’s embedded, and many web services strip metadata on upload regardless of format.

Use PNG for logos, graphics, screenshots, and web images requiring transparency. Use it for interface elements and illustrations with sharp edges and flat colors where JPEG artifacts would be visible. Do not use PNG for delivering photographs, archiving photos, or print delivery.

HEIF: The Modern Compression Champion

High Efficiency Image Format uses HEVC video compression technology to achieve roughly 50 percent smaller files than JPEG at equivalent quality. Apple adopted it as the default capture format on iPhones, and it’s gradually appearing in other devices and platforms.

Beyond size efficiency, HEIF supports 10-bit color depth compared to JPEG’s 8-bit ceiling, can store image sequences and depth maps in a single container, and technically supports transparency in some implementations. On Apple devices, HEIF typically uses the Display P3 color space, which is wider than sRGB and can cause oversaturation when images are viewed in applications that don’t handle color management properly.

The compatibility situation is HEIF’s Achilles heel. Windows requires installing extensions to open HEIF files. Older software may not support the format at all. While Safari and some browsers can display HEIF natively, cross-platform support and downstream pipeline acceptance remain the real pain points. Many print labs, stock agencies, and clients simply cannot accept HEIF deliverables. If your iPhone captures in HEIF, that’s fine for personal use and situations where you control the entire pipeline. For professional delivery, convert to JPEG or TIFF before sending files to anyone else until wider adoption is achieved.

Emerging Formats: WebP and JPEG XL

Google’s WebP format offers both lossy and lossless modes with roughly 25 to 35 percent smaller files than equivalent JPEGs. It supports transparency and has achieved good browser support. Traditional photo-editing software support remains weaker. WebP makes sense for web delivery when you control the platform and file size is a priority.

JPEG XL is perhaps the most promising next-generation format, supporting lossy, lossless, and progressive decoding along with HDR, wide gamut, and high bit depth. It can even losslessly recompress existing JPEG files to smaller sizes without additional quality loss. The browser situation is mixed: Chrome added and then removed JPEG XL support, but Safari 17 includes native support across iOS 17, macOS Sonoma, and even older macOS versions like Ventura and Monterey that can run Safari 17. Given Apple’s significant presence in photography workflows, this isn’t a dead format, but adoption remains fragmented.

Practical Recommendations by Workflow Stage

For capture, shoot raw whenever possible. Use JPEG only when you need maximum burst speed, your camera lacks raw capability, or you’re absolutely certain minimal editing will be needed. For editing and working files, always start from raw. If you need to move files between applications, use 16-bit TIFF with LZW compression or PSD if layers matter. Never use JPEG as a working format.

For web and client delivery, JPEG remains the correct choice: universal compatibility, reasonable file sizes, and sufficient quality. Export in sRGB at quality 80 to 90, sized appropriately for intended use. For print delivery, use TIFF unless your lab specifically requests JPEG. Follow their color space specifications and provide full resolution files. For archiving, keep your original raw files as master negatives, optionally convert to DNG for format longevity insurance, and consider archiving finished TIFF masters alongside raw. Never archive only JPEGs because you’ll have eliminated all editing headroom forever.

Understanding what each format actually does removes the mystery from these decisions. Match the format to the task, preserve maximum flexibility for as long as possible in your workflow, and convert to delivery formats only at the final step. If you want to dive deeper into raw processing workflows, Fstoppers offers comprehensive tutorials on both Mastering Adobe Lightroom and The Complete Capture One Editing Guide that cover everything from import to export.