Image Format Converter

Image formats differ along three axes that actually matter: whether they discard information to save space, whether they support transparency, and how widely they are supported. Almost every practical decision follows from those three.

Lossy and lossless are different jobs

JPEG is lossy. It discards detail the eye is least likely to notice, which works extremely well on photographs where color varies smoothly, and badly on sharp edges. This is why screenshots, diagrams, and text saved as JPEG develop a halo of noise around every line. PNG is lossless and handles exactly the opposite case: flat areas of solid color compress superbly. A photograph saved as PNG can easily be several times larger than the JPEG equivalent with no visible benefit.

WebP covers both modes in one format and typically lands meaningfully smaller than JPEG at comparable quality. AVIF compresses harder still, at the cost of slower encoding. Both support transparency, which older formats forced you to choose between alongside efficient photographic compression.

Generation loss is the mistake to avoid

Every save of a lossy format re-runs the compression, and the damage accumulates. A JPEG opened, edited slightly, and saved a dozen times looks visibly degraded even though each individual save seemed fine. Crucially, converting a JPEG to PNG doesn't recover anything, because the detail was discarded at the first save and PNG simply preserves the damaged version losslessly. Always convert from the highest quality original you have rather than from an already-compressed copy.

Transparency and the halo problem

JPEG has no alpha channel at all. Converting a transparent PNG to JPEG doesn't preserve or ignore transparency, it fills it, usually with white or black, which is why logos converted this way arrive with a rectangle around them. If an image needs to sit on varying backgrounds, it needs PNG, WebP, or AVIF.

One last distinction that catches people out: SVG is a vector format describing shapes rather than pixels, so it scales to any size perfectly. Converting a photograph to SVG gains nothing, because the photograph has no shapes to describe and the result is simply the same pixels wrapped in markup.