Photography workflows are a chain of format decisions. Shoot RAW or JPEG? Edit in 16-bit or 8-bit? Deliver JPEG or TIFF? Archive as DNG or proprietary RAW? Each choice trades quality against file size, compatibility, or processing speed.

This guide covers the format pipeline from shutter press to final archive, with the actual specs that matter: bit depth, color spaces, metadata handling, and file sizes you can expect at each stage. Written for photographers who want to understand why they're making these choices, not just follow a recipe.

Capture Formats: RAW vs JPEG In-Camera

Your camera records light hitting the sensor. How that data gets stored determines everything downstream.

RAW: The Full Sensor Data

What RAW actually is: Raw sensor data before demosaicing. Each pixel records a single color channel (R, G, or B) through the Bayer filter. The RAW file stores these single-channel readings at 12-bit or 14-bit depth (4,096 or 16,384 brightness levels per pixel), compared to JPEG's 8-bit (256 levels).

Camera-specific formats:

  • Canon CR3: HEIF-based container (replaced CR2 in 2018). Supports C-RAW (lossy compressed, 40% smaller) and standard RAW (lossless). A Canon R5 CR3 file at 45MP is ~25MB (C-RAW) or ~50MB (standard).
  • Nikon NEF: TIFF-based container with Nikon-specific metadata. 12-bit or 14-bit, lossless or lossy compressed. A Z8 NEF at 45.7MP runs 35–60MB depending on compression.
  • Sony ARW: TIFF-based container. Sony's compression is generally lossless. An A7R V ARW at 61MP is ~60–75MB uncompressed.
  • Fujifilm RAF: Unique X-Trans sensor pattern (different from standard Bayer). Some third-party software handles X-Trans demosaicing poorly, producing wormy artifacts. Capture One and Fujifilm's own converter handle it best.

Convert CR2 to JPG | Convert NEF to JPG | Convert ARW to JPG | Convert DNG to JPG

JPEG In-Camera: The Processed Output

When you shoot JPEG, the camera performs demosaicing, white balance, contrast curve application, noise reduction, sharpening, color space conversion, and 8-bit quantization — all in-camera, irreversibly.

What you lose: 6 stops of highlight/shadow recovery (14-bit RAW has 16,384 levels vs JPEG's 256). White balance flexibility (baked in at capture). Color grading headroom (8-bit banding appears with heavy adjustments). Fine detail in shadows (compressed away by the tone curve).

When JPEG makes sense: Sports/press work where volume is extreme (thousands of shots per event) and turnaround is minutes. Casual photography where post-processing isn't planned. Camera buffer speed (JPEG clears the buffer faster for continuous shooting).

RAW+JPEG: Most cameras support dual recording. The JPEG serves as a quick reference; the RAW is there when you need it. The storage cost is ~20–30% more than RAW alone.

DNG: Adobe's Universal RAW

Digital Negative (DNG) is Adobe's open RAW format, designed to solve the proprietary RAW problem. CR3, NEF, and ARW are undocumented formats controlled by camera manufacturers. If Canon stops supporting CR3 in 2040, your files may become unreadable.

DNG conversion: Adobe DNG Converter (free) converts any proprietary RAW to DNG. The original sensor data is preserved (optionally embedded with the original RAW file for safety). Lightroom can convert to DNG on import.

Pros: Open specification (documented, freely implementable). Embeds XMP metadata directly (no sidecar files). Smaller than most proprietary RAWs (15–20% with lossy compression, 0–5% with lossless). Future-proof — Adobe commits to long-term support.

Cons: Conversion adds time to import workflow. Some camera-specific metadata may not transfer. Nikon and Canon have no incentive to support DNG natively. Some purists prefer keeping the original proprietary file untouched.

Convert DNG to TIFF | Convert DNG to PNG

Editing Formats: Working Files

Once RAW data is demosaiced and you're in an editor (Photoshop, Capture One, Affinity Photo), the working format determines how much quality headroom you have for adjustments.

TIFF 16-bit: The Editing Workhorse

Why 16-bit: A 16-bit TIFF stores 65,536 levels per channel versus 8-bit's 256. Heavy curves adjustments, exposure corrections, and color grading on 8-bit files produce visible banding (stair-stepping in gradients). 16-bit gives you the headroom to push and pull tones without degradation.

File sizes: 16-bit doubles the file size versus 8-bit. A 45MP image (8256x5504) at 16-bit RGB is ~260MB uncompressed, ~130MB with LZW. With layers, it grows fast. Use LZW or ZIP compression — both are lossless with zero quality impact.

When to drop to 8-bit: After all edits are finalized and you're exporting for delivery. The final JPEG or WebP output is 8-bit anyway. Converting to 8-bit at the end (after adjustments are applied) preserves all the benefit of 16-bit editing.

Convert TIFF to JPG | Convert TIFF to PNG | Convert TIFF to WebP

PSD: Layered Photo Composites

PSD files preserve the full editing state: layers, masks, adjustment layers, smart objects, and blend modes. Use PSD when your edit involves compositing (combining multiple images), complex masking, or non-destructive adjustment stacks.

Size limit: PSD maxes out at 2GB. For files larger than that (common with 100MP+ cameras at 16-bit with many layers), use PSB (Photoshop Big, identical features, 300,000x300,000px limit).

Compatibility: PSD is Photoshop-specific. Affinity Photo reads most PSD features. GIMP handles basics but drops adjustment layers and smart objects. For cross-editor compatibility, flatten to TIFF.

Convert PSD to JPG | Convert PSD to PNG | Convert PSD to WebP

Delivery Formats: Client and Web Output

Delivery format depends entirely on how the recipient will use the image.

JPEG: The Universal Client Delivery Format

Quality settings: JPEG quality 85–92 is the sweet spot for client delivery. Below 85, compression artifacts become visible on close inspection (especially in skin tones and gradients). Above 92, file size increases sharply with minimal visible improvement. Quality 100 is roughly 3x the file size of quality 90 with no perceptible difference.

Color space: sRGB for all client deliveries unless specifically requested otherwise. Adobe RGB is theoretically better for print, but most consumer software and devices assume sRGB. An Adobe RGB image viewed in an application that doesn't understand color management looks desaturated and dull. Embed the ICC profile always.

Sizing: Full resolution for print clients. 2048px on the long edge for web/social media. 800–1200px for proofing galleries. Strip EXIF GPS data for privacy before delivering to clients (keep camera/exposure data).

WebP: Web Publication

If your photos end up on a website (portfolio, blog, online gallery), WebP at quality 80–85 produces files 25–35% smaller than equivalent JPEG with comparable visual quality. For a typical 2000px wide portfolio image: JPEG quality 85 = ~350KB, WebP quality 82 = ~240KB.

WebP's advantage over JPEG is most pronounced at lower quality settings. At quality 60–70 (bandwidth-constrained scenarios), WebP looks noticeably better than JPEG at the same file size.

Convert JPG to WebP | Convert PNG to WebP | Convert TIFF to WebP

Consumer print labs (Shutterfly, Mpix): JPEG sRGB at 300dpi. Most consumer labs convert everything to sRGB JPEG internally anyway. Sending a CMYK TIFF just means the lab's auto-conversion might produce worse results than your own sRGB JPEG.

Professional print labs: TIFF 8-bit, AdobeRGB or lab-specific ICC profile, 300dpi at output size. Some high-end labs accept 16-bit and do their own profiling. Ask for their ICC profile and soft-proof against it before submitting.

Fine art / giclée: TIFF 16-bit, AdobeRGB or ProPhoto RGB, 360dpi (Epson) or 300dpi (Canon). These printers can reproduce a wider gamut than sRGB, so sending sRGB files means leaving color on the table. The print shop should provide their paper-specific ICC profile.

Color Spaces for Photography

Color spaces define the range of colors available. Choosing the wrong one means either leaving quality on the table or producing broken output.

sRGB: The Safe Default

sRGB covers about 35% of the visible color spectrum. It's the assumed color space of the web, most consumer displays, most software, and most devices. If no ICC profile is embedded, software assumes sRGB.

Use for: All web delivery, client JPEG proofs, social media uploads, email attachments, any scenario where you don't control the viewing environment.

Adobe RGB: Print Editing

Adobe RGB covers about 50% of the visible spectrum — notably wider in cyan-green tones. The difference is most visible in saturated teal/turquoise, forest greens, and certain blues.

Use for: Print editing workflows where the final output is a professional print lab or press with a wide-gamut workflow. Edit in Adobe RGB, deliver print files in Adobe RGB, deliver web files in sRGB (converted at export).

ProPhoto RGB: Maximum Editing Headroom

ProPhoto RGB covers ~90% of the visible spectrum. It includes colors no display can currently show and no printer can currently reproduce. It exists for editing headroom — extreme color grading and wide-gamut archival.

Critical rule: ProPhoto RGB MUST be used at 16-bit. At 8-bit, the huge gamut is spread across only 256 levels per channel, creating visible banding in normal-saturation areas. A ProPhoto 8-bit file looks worse than an sRGB 8-bit file in common color ranges.

Use for: Lightroom editing (it uses ProPhoto internally by default), master archival copies of heavily edited images, handoff to fine art printers who accept ProPhoto.

Display P3: The New Screen Standard

P3 is 25% wider than sRGB, covering more reds, greens, and oranges. Every Apple device since 2016 (iPhone 7, MacBook Pro 2016+) uses P3 displays. Most modern Android flagships and monitors support it too.

For photographers: If your audience views on modern devices and you want the most vibrant colors, consider P3 delivery for web (supported in CSS and Safari/Chrome). But always provide sRGB fallback. A P3 image on an sRGB display looks identical to sRGB — the extra colors simply get clipped.

Metadata: EXIF, IPTC, and XMP

Photo metadata is invisible but critical for organization, rights management, and workflow.

EXIF: Camera and Exposure Data

Contains: Camera make/model, lens, focal length, aperture, shutter speed, ISO, date/time, GPS coordinates (if enabled), white balance, metering mode, flash status, orientation.

File size: Typically 5–50KB. Embedded thumbnails (for quick preview) add 20–100KB. Some cameras embed a full-size JPEG preview in the RAW file (adding 2–10MB).

Privacy: GPS coordinates are the biggest concern. Strip them before sharing publicly. Social media platforms usually strip all EXIF on upload, but direct web publication doesn't.

IPTC and XMP: Rights and Organization

IPTC: Standardized fields for photographer name, copyright, caption, keywords, location, contact info. Embedded in JPEG and TIFF. Survives most file conversions. Essential for stock photography (agencies require IPTC metadata).

XMP: Adobe's extensible metadata format. Stores Lightroom/Camera Raw edit settings as XML. Can be embedded in the file or stored as a sidecar (.xmp file alongside the RAW). DNG files embed XMP directly; proprietary RAWs use sidecar files.

Preservation: Format conversions can strip metadata. JPG-to-PNG preserves EXIF in most tools. JPG-to-WebP preservation depends on the encoder. Always check that copyright and contact IPTC data survives the conversion chain.

Archival Strategy

Long-term photo storage requires formats that will be readable in 20+ years. Proprietary RAW formats are the biggest risk — if the camera manufacturer abandons a format, your files may require specialized software to open.

Recommended Archival Formats

RAW archival: Keep the original proprietary RAW file (CR3, NEF, ARW) AND convert to DNG as a backup. The proprietary RAW preserves manufacturer-specific features; the DNG ensures readability if the proprietary format becomes unsupported.

Processed archival: TIFF 16-bit with LZW compression. Open standard, universally supported, lossless. Include an embedded AdobeRGB or ProPhoto ICC profile. A 45MP archival TIFF with layers is 200–500MB — plan your storage accordingly.

Storage math: A 1,000-photo shoot at 45MP: ~50GB RAW, ~25GB DNG (converted, lossy compressed), ~250GB TIFF masters (16-bit, layered). Total: ~325GB. At $0.004/GB/month on cold storage (Glacier/Backblaze), that's $1.30/month to archive forever.

The photography format pipeline is straightforward once you map each stage to its requirements: maximum data at capture (RAW), maximum headroom during editing (16-bit TIFF/PSD), minimum file size at delivery (JPEG/WebP), and maximum longevity at archival (DNG + TIFF).

The most expensive mistake is format conversion at the wrong stage. Converting RAW to JPEG before editing destroys headroom you can't recover. Archiving only JPEGs means future edits start from a degraded source. Get the pipeline right, and format decisions become automatic.