RAW Format

RAW files contain the complete sensor readout — the exact light values captured by each photosite on the camera sensor. When a camera saves a JPEG, it applies noise reduction, sharpening, color processing, and compression, permanently baking in these decisions. A RAW file preserves the original data, allowing the photographer to make all processing decisions in post-production with full flexibility. Common RAW formats include CR3 (Canon), NEF (Nikon), ARW (Sony), and DNG (Adobe's universal RAW format).\n\nWedding photographers almost universally shoot in RAW because the ceremony cannot be re-photographed. If the photographer misjudges exposure or white balance during a fast-moving ceremony in mixed lighting, RAW files provide the latitude to recover the images in post-processing. A JPEG shot at the same settings might have unrecoverable blown highlights or crushed shadows. The editing flexibility of RAW is an insurance policy against imperfect capture conditions.\n\nThe tradeoff for RAW flexibility is file size and workflow complexity. A single RAW file may be 25-80 megabytes compared to 5-10 megabytes for a high-quality JPEG. A wedding shoot of 2,000 images in RAW requires 50-160 gigabytes of storage, compared to 10-20 gigabytes in JPEG. RAW files also require processing before sharing — they cannot be directly posted to social media or sent to clients without first being converted to a standard display format.\n\nMagic Eraser processes uploaded images regardless of original capture format. While the AI Enhance tool is effective on all image types, photos originally captured in RAW and then processed to high-quality JPEG or PNG preserve more editing latitude than heavily compressed images. The AI can recover more detail from well-captured source material. For best results, users should upload the highest quality version available of any image they want to edit or enhance, as the AI builds upon the information present in the source file — starting with a well-exposed, minimally compressed original gives the enhancement algorithms the richest possible data to work with when applying improvements.