Vector vs Raster in PDFs: Why Some Files Stay Sharp Forever

Zoom into one PDF at 800 percent and the letterforms stay razor sharp. Zoom into another and you get a mosaic of soft grey blocks. Both are valid PDFs, and the difference has nothing to do with quality settings. It comes from what the page actually contains. A PDF page is a container that can hold two fundamentally different kinds of graphics, and knowing which one you are dealing with predicts almost everything else about how the file will behave.

Vector content is stored as instructions rather than appearance. The page description says: begin a path at these coordinates, draw a curve with these control points, fill it with this colour, then set the font to Helvetica at 11 points and place the character sequence at this position. Nothing describes a pixel. When you zoom, the rendering engine simply recalculates the same maths at a higher resolution, so curves stay smooth at any magnification. Because instructions are compact, a hundred-page vector report can be under a megabyte, and because characters are stored as text you can select, search, and copy them.

Raster content is the opposite. The page holds a fixed grid of coloured samples, and the number of samples was decided when the image was created. Zooming cannot invent detail that was never captured, so the viewer enlarges each sample and you see blur or blocks. A scanned document is pure raster: the scanner measured light reflectance across a grid and stored the result. To the PDF, what looks like a paragraph is just a region where dark pixels happen to form letter shapes. There is no text there at all, which is why search finds nothing.

This distinction drives real decisions. Compression behaves differently on each type: vector pages shrink mainly through font subsetting and stream compression, while raster pages shrink by downsampling and re-encoding images, which is where visible quality loss comes from. PdfWill compression is tuned to keep text selectable rather than flattening everything to pictures. Converting PDF to Word succeeds on vector text and produces empty boxes on scans. PDF to JPG or PDF to Image rasterises everything, which is a one-way door: once the instructions are gone, they cannot be recovered.

The practical rules follow naturally. Keep documents vector for as long as possible in your workflow, and only rasterise at the final step when you genuinely need a picture. If you receive a scan and need searchable text, run OCR, which analyses the pixel shapes and adds an invisible text layer underneath so the file becomes searchable while still looking identical. If a file is unexpectedly huge, check whether someone printed a vector document to image and rebuilt it as a scan, a surprisingly common way to turn a 400 KB report into 30 MB.