Vector vs raster

A vector file stores shapes as math: points, curves, fills, strokes. Blow it up to billboard size and the renderer just recomputes the curves, so edges stay razor sharp. A raster (bitmap) file stores a fixed grid of pixels, so enlarging it invents pixels that were never captured, which is where the mush comes from. File extensions are a rough hint, not a promise: AI, EPS, SVG, and PDF are containers that happily hold placed photos, and a PSD can hold live vector shapes and type. JPG, PNG, and TIFF are pixels all the way down. Open the file and look before you trust the suffix.

The rule is boring and reliable: anything drawn (logos, icons, type, diagrams, dielines) should be vector; anything captured or painted (photos, textures, brush art) is raster and needs enough pixels for the final printed size, usually 300 PPI. This is why print shops keep demanding 'the vector version' of a logo and refuse the PNG you pulled off the website.

Gotcha: putting a raster image inside an SVG or AI file does not make it vector, it just puts a pixel photo in a vector wrapper. Same for auto-tracing a low-res JPG logo: you get wobbly, node-heavy curves that look wrong at large sizes. Transparency is a separate trap: PDF/X-4 and SVG can carry it live all the way to the RIP, while pixel-based effects like blurs, glows, and drop shadows usually rasterize somewhere in the pipeline, so set the raster effects resolution to 300 PPI before you hand off and ask which PDF standard the shop wants.