Ajouter un Filigrane

Watermarking is the practice of adding a visible mark -- typically a logo, brand name, or copyright notice -- to an image to establish ownership and deter unauthorized use. In the age of social media, images can be copied and redistributed in seconds, often without attribution. A watermark makes it clear who owns the image and makes unauthorized use less appealing, as the watermark must be removed (which degrades image quality) before the image can be passed off as original.

Watermarks are used by photographers (to protect portfolio images), e-commerce businesses (to prevent competitors from copying product photos), content creators (to brand social media graphics), and stock photo agencies (to preview images before purchase). The key is finding the right balance: the watermark should be visible enough to deter theft but not so obtrusive that it ruins the viewing experience. A common technique is placing a semi-transparent watermark in the center or bottom-right corner with 30-50% opacity.

Beyond visible watermarks, invisible (digital) watermarks embed information in the image's pixel data that can be detected by specialized software but not seen by humans. These are used by services like Digimarc to track image usage across the web. However, visible watermarks remain the most practical solution for most creators -- they're simple, effective, and don't require any special software to verify.

Image watermarking is performed using the Canvas API's drawImage() and fillText() methods. The image is first drawn to a canvas, then the watermark text or logo is composited on top using globalAlpha for transparency. The canvas is then exported as a new image using toBlob() or toDataURL(). The output format can be JPEG (smaller, lossy) or PNG (larger, lossless). For photographs, JPEG at 90% quality provides a good balance of file size and visual quality.

When watermarking multiple images (batch processing), the browser processes them sequentially to avoid memory issues. Each image is loaded, watermarked, exported, and then the canvas memory is released before processing the next one. For large batches (100+ images), this can take several minutes but won't crash the browser. The processing happens in the browser's main thread, so the UI may be slightly unresponsive during batch operations -- web workers could be used for background processing, but Canvas API support in workers is limited.