SVG to JPG

SVGs are usually logos, icons, diagrams and charts: flat colour, hard edges and text. JPEG is built for photographs, and it handles exactly those things worst — hard edges get a halo of compression artefacts, and small text turns mushy.

Convert to JPEG when something downstream insists on it, which does happen: some upload forms, some print workflows, some older systems. Otherwise use PNG, which is lossless and usually smaller for this kind of artwork.

JPEG has no alpha channel, so every transparent pixel needs a colour. It is filled with white by default; set it to whatever the image will sit on.

Without that step transparent areas come out black, because an empty canvas is transparent black — which is the single most-reported bug in every image converter that exists.

An SVG has no pixels. It is a description of shapes, so "how big" is a question you answer, not one the file does.

Most SVGs declare a nominal size, or a viewBox that gives their proportions — both are read here and shown in the result. Rendering at a scale of 2 gives a result that is crisp on a retina screen; a scale of 1 matches the size the file declares.

Some SVGs declare neither, which is legal and common for icons meant to be sized by CSS. There is genuinely nothing in such a file that says how large it should be, so it falls back to the browser default of 300 × 150 and says so — set a width instead.

An SVG can reference things by URL: a web font, a linked image, a stylesheet. It can also contain scripts, which is why an SVG uploaded to a website is a well-known attack route.

Drawing one as an image puts the browser into a locked-down mode: no scripts run, and nothing external is fetched. That is the browser's own protection and it is the behaviour you want for a file you did not write.

The consequence is visible: text set in a web font falls back to a system font, and any picture the SVG links to by URL comes out blank. Fonts embedded in the file, and images stored in it as data URIs, are unaffected — those are the ones a self-contained SVG uses.

Rasterising is lossy in the way that matters most: the shapes, paths and text stop being objects and become a grid of pixels. Scale the result up and it softens; there is no route back.

Keep the SVG. Convert a copy for whatever needs pixels — an upload form, a document, a system that will not take vectors — and edit the original when the artwork changes.

There is no honest PNG-to-SVG on this site for the same reason. Tracing a photograph into vectors produces a large file full of guessed paths that looks worse than what it started with, and any tool offering it is either doing that or simply wrapping the bitmap in an <svg> tag, which is not a conversion at all.