Vector vs raster PDF: what’s the difference?
By Owain Wedlake, Founder, Nest Ready · Published
A vector PDF holds a drawing as instructions: lines, arcs and text, each with exact coordinates. A raster PDF holds a picture of the page, made of pixels. At normal zoom the two look the same. The difference shows when you zoom in, try to pick out a line, or need a DXF: geometry can be read straight out of a vector PDF, but in a raster PDF it has to be traced from the pixels and scaled from the dimensions written on the drawing.
What is a vector PDF?
In a vector PDF the page is described as geometry. When a drawing is exported or plotted to PDF from CAD, each line, arc, circle and piece of text goes into the file as a drawing instruction with exact coordinates. Most drawings that come straight from a customer’s CAD or drawing management system are vector.
Three details matter if you want a DXF from one:
- Arcs are stored as curves. PDF has no arc or circle as such, so CAD writes them as smooth curves (Béziers) that software has to recognise as arcs again.
- Text may not be text. Some CAD fonts are exported as short strokes rather than letters, so a vector PDF can show words that can’t be searched or selected.
- The page is drawn at the sheet’s scale. A view drawn at 1:2 on an A3 sheet is half size in the PDF too. The geometry is exact, but it still has to be scaled to true size.
What is a raster PDF?
A raster PDF holds an image of the page: a grid of pixels, each one black, white, grey or a colour, with nothing to say which pixels belong to a line, a letter or a speck of dust. Raster PDFs usually come from:
- a paper drawing put through a scanner, or a copier’s scan-to-email;
- a photo of a drawing saved or exported as a PDF;
- “print to PDF” from a JPG, PNG or TIFF, or a drawing printed as an image;
- a vector drawing printed, marked up by hand and scanned back in.
How much detail survives depends on the resolution, in dots per inch (dpi). At 300 dpi each pixel is about 0.085 mm across on the paper. On a view drawn at 1:5, that one pixel stands for about 0.42 mm of the real part, before any blur or skew.
How to tell whether a PDF is vector or raster
Open the file in any PDF viewer. The first two checks settle most drawings in under a minute.
- Zoom in a long way, to 800% or more, on the edge of a hole or a corner of the title block. In a vector PDF the lines stay crisp however far you go. In a raster PDF the edges turn soft, stepped or fuzzy, and specks or paper texture may show. A good scan can look clean at 200%, so don’t stop early.
- Try to select a single line or a word of text. Drag the text selection tool across a word in the title block. In a vector PDF you can usually highlight it letter by letter, and a PDF editor will often pick out a single line on its own. In a raster PDF the click selects nothing, or the whole page as one image.
- Search for a word you can see, such as the part number, with Ctrl+F (Cmd+F on a Mac). If it’s found, the page holds real text, which points to vector. Treat the result with care: some CAD exports turn text into strokes, so a vector page can fail this test, and a scan that has been through text recognition (OCR) can pass it.
- Look at the file’s properties. In most viewers, File › Properties (or Document Properties) shows the application that created the PDF and the PDF producer. A CAD program points to vector; a scanner, a copier, a scanning app or an image editor points to raster. These fields are only as good as the software that filled them in, so use them to back up the zoom test.
- Check every page, and watch for mixed ones. A drawing set can mix vector pages with scanned ones. A single page can mix too: a vector drawing with a scanned signature, approval stamp or logo pasted in, or a scan with a typed title block or revision note added on top as vector text. For a DXF, it’s the part’s own linework that counts, so zoom in on the profile and holes, not just the title block.
| Vector PDF | Raster PDF | |
|---|---|---|
| Sharpness at zoom | Lines stay crisp at any zoom | Edges go soft, stepped or fuzzy |
| Selecting a line | Words, and in a PDF editor single lines, can be picked out | The page selects as one image, or not at all |
| Searching text | Usually found, unless the CAD export turned the text into strokes | Not found, unless the scan has been through text recognition (OCR) |
| Typical file size | Often small for a part drawing; grows with hatching and detail | Often larger, depending on resolution, colour and compression |
| Where it usually comes from | Exported or plotted from CAD | A scanner, a photo of the drawing, or print to PDF from an image |
| How geometry is recovered for a DXF | Read from the file’s own lines and curves, then scaled to true size | Traced from the pixels, with the scale worked out from the written dimensions |
Why does it matter when you need a DXF?
A DXF for cutting is geometry: lines, arcs and circles at true size. How you get there depends on what the PDF holds.
From a vector PDF, the geometry is already in the file, so a straight edge comes out straight and a hole comes out round. What’s left is sorting and scaling. Dimension lines, centre lines, text, hatching, the border and the title block are geometry too, and only some of the page is the part, which then has to go from the sheet’s scale to true size.
From a raster PDF there is nothing to read, only pixels. The part has to be redrawn or traced, and two questions come with that:
- Where is the edge? A drawn line is several pixels thick, with blurred edges, so a trace has to decide where the true line runs, and a feature only a few pixels across can be lost. Auto-trace tools also trace everything else on the page, often turning one clean edge into many short segments.
- How big is it? A scan’s pixels carry no dependable size: the same sheet scanned at 200 and 400 dpi gives two different pixel counts. The scale note doesn’t settle it either, because the paper may have been printed fit-to-page or copied at a different size before it was scanned. The dependable reference is the dimensions written on the drawing.
So a vector PDF is usually quicker and more certain to turn into a DXF than a scan of the same drawing. When a customer sends a scan, ask whether they have the CAD-exported PDF or the native CAD file. If they don’t, the options are set out in how to convert a scanned PDF to DXF.
How does Nest Ready handle vector and raster PDFs?
Nest Ready classifies each page of a PDF on its own, so a drawing set that mixes CAD-exported pages with scanned ones doesn’t need sorting before you upload it.
- Vector pages are read from the PDF: the geometry comes from the file, and the model chooses which closed outlines are parts. More on this route in PDF to DXF for engineering drawings.
- Scanned pages are traced from the pixels; the model only reads the text. The scale comes from the dimensions written on the drawing and stands only when at least two readings agree. “SCALE 1:N” notes are ignored. Each part shows its stated accuracy, such as ±0.6 mm, and scans worse than 1.0 mm are refused. More in scanned PDF to DXF.
- Pages it can’t read reliably are refused, with the reason and a fix: a blank page, a page too dense to trace, a scan whose writing is too small to read reliably (the fix is a rescan at 200 dpi or more), or a scan too coarse for the drawing’s scale.
Either way the DXF is R12, with LINE, ARC and CIRCLE entities only: the outer profile on layer OUTER, holes on layer INNER, no text or dimensions, at true size in
millimetres. A part from a vector page counts as one part against your allowance; a part from a
scan counts as 5.
The limits: anything uncertain is flagged for review, and parts from scans are flagged by default, but flagged parts you haven’t decided on still download. Check every DXF against the drawing before cutting. Nest Ready produces part geometry only, with no toolpaths, cutting parameters, machine code or nesting.