SVG vs PDF for Print: Choosing by What Survives

A commercial printer will ask for a PDF. Almost nobody will ask for an SVG, and the ones who do are usually handing it to a laser cutter. The reason is not fashion: PDF carries a page model with real edges, and SVG does not. Everything else about the comparison follows from that.

Both formats are vector-first, both are text-based, both can embed raster images, and both describe an image rather than storing pixels. They were designed twenty years apart for different destinations.

Pages, boxes, and why trim exists

A PDF page has a /MediaBox, which defines the full physical page; a /CropBox, which is what a viewer displays by default; a /BleedBox, which extends past the trim to the area the printer trims from; and a /TrimBox, which is the finished size of the printed piece. A /ArtBox exists too, for the content area within the page.

Those four boxes solve a problem that every print job has. Artwork must extend beyond the final edge so the cutting tolerance does not leave a white sliver. The file has to say, unambiguously, where the final edge is. Trim is usually 3 mm beyond the finished size in Europe and 1/8 inch in the US, and the bleed area must be filled with artwork.

SVG has viewBox, width, height, and nothing else. There is no bleed element, no trim designation, and no standard way to communicate that the outer 3 mm is extra. A viewBox can define a coordinate window, but nothing in the specification says which part of that window is the finished piece. Prepress software has no way to recover the information, so a human has to be told, or the job is set up wrong.

Crop marks and imposition

PDF workflows also standardise how crop marks, registration marks, and colour bars sit on the sheet, and how multiple pages are imposed onto a press sheet. None of that vocabulary exists in SVG. You can draw crop marks as lines, but they are just lines — the RIP treats them as artwork, not as instructions.

Colour: CMYK and spot ink

PDF has supported device CMYK, ICC-based colour, and Separation and DeviceN spot colour spaces from the beginning of its print-oriented life. A PDF/X-1a, PDF/X-3, or PDF/X-4 file is a contract: the colour is defined in a specific way, fonts are embedded, and no external dependency can change what the output looks like.

SVG is sRGB by specification. SVG 2 allows ICC colour profiles through CSS Color, but renderer support is thin and prepress tools generally ignore it. There is no spot colour concept. If a job specifies Pantone 032, an SVG cannot say so in any way a RIP will read.

Practical consequence: a design house that delivers SVG to a printer forces the printer to guess at colour conversion. That guess may be right, but it is a guess, and it will not be consistent across jobs.

Fonts and the embedding contract

PDF's font model is mature. A producer embeds a subset of the font, with the licensing bits that declare whether the file may be further extracted, and the RIP uses the embedded data. Text stays sharp at any resolution and the output is byte-identical to the proof.

SVG can carry text and can reference fonts, but the format does not require embedding and most importers do not honour it. Browser rendering substitutes silently. Illustrator prompts. A cutting machine typically ignores text entirely and requires conversion to paths.

For print specifically, embedding either works (a properly built PDF) or you convert all text to outlines. Outlining in a PDF is acceptable but loses searchability and accessibility. Outlining in an SVG is routine.

Rasters inside vector documents

Both formats can hold bitmaps. A PDF embeds images as XObjects with a defined colour space and resolution; a prepress operator can inspect each image and check its effective DPI against the line screen. An SVG embeds or links images with <image>, and linked images are external dependencies that will silently break if the file moves.

For print, 300 DPI at final size is the usual minimum, and 400–600 DPI for line art with fine detail. There is no automatic check in SVG land, so a linked 72 DPI placeholder sitting inside a print-bound SVG is a common and expensive mistake.

Where SVG is the correct answer

Not every print-adjacent job is a press job. Vinyl cutting, laser cutting, Cricut and Silhouette cutting, and craft plotter work all want SVG or DXF, because those machines need paths and they do not care about page boxes. Screen printing a simple one-colour job often works from a vector file with explicit separation layers, and SVG is fine there as long as the shop can convert it.

Web delivery, application UI, icons, and any context where the image must scale with a viewport or be styled with CSS belongs in SVG. That is not a compromise; it is the format's purpose.

There is also a hybrid role: a PDF is often the best source for making an SVG. Vector PDFs from Illustrator or Inkscape carry clean paths that convert into an SVG with no tracing step. PDF to SVG is a lossless geometry transfer for a genuinely vector PDF, and a tracing operation for a scanned one.

Converting between them

Going SVG to PDF is common when a design needs to become a document: an invoice template, a certificate, a proof for a client. SVG to PDF renders the SVG into PDF drawing operators, and the output is sharp and small. What it cannot do is invent a TrimBox, a CMYK colour space, or an embedded font — it converts what is there.

Going PDF to SVG is common when a vector PDF needs to become editable web artwork. A PDF exported from a vector source converts cleanly. A PDF exported from a raster source, or one where the vector content is obscured by a raster overlay, converts to a traced approximation of that raster.

The often-overlooked point is that PDF's content stream and SVG's element tree have a natural mapping: paths, fills, strokes, and transforms all transfer almost one to one. Groups, clipping, and transparency transfer with some care. Gradients and fonts do not always.

Why print tolerances are measured in micrometres

Offset printing runs at a registration tolerance far tighter than most digital work. A misregistration of 0.05 mm between two inks is visible on a solid edge. That precision only works if the file states the intended page geometry exactly, which is what the PDF page boxes provide. A printer can program their imposition software from those values without asking you.

Bleed is the specific case where this becomes expensive. If a full-bleed image carries no bleed and the sheet drifts by a millimetre during trimming, the finished piece has a white hairline along the edge and is discarded. That is a whole run wasted on a file-level detail: whether the artwork extended past the trim line. SVG has no mechanism to declare where the trim line is, so the printer cannot automate the check.

Ink limits and total area coverage

CMYK builds on a press are not just four independent channels. Total area coverage — the sum of the four ink percentages at a given point — has a practical ceiling, typically 300% on coated stock and 240–280% on uncoated. Beyond that, ink does not dry and the sheet sets off. Rich black, built from 60C 40M 40Y 100K, sits around 240% and is standard; a naive conversion of pure RGB black can produce something near 400% and ruin the job.

PDF pipelines can carry a Total Area Coverage limit and prepress tools validate against it. Whether an SVG can even express four-channels-and-a-limit is moot: it cannot express CMYK at all, so the TAC question never arises until someone converts. And the conversion is done blind, because the source gave no guidance.

Multi-page documents and imposition

A brochure, a booklet, a business card grid — all of these are multi-page jobs. A PDF carries page objects and can describe dozens of pages in one file, each with its own boxes. Imposition software then arranges those pages onto a press sheet with the correct creep and gutter adjustments.

An SVG holds one canvas. Representing multiple pages means either separate files or one enormous canvas with the pages laid out — the latter is what some artboards do when exported, and printers complain about it because they have to guess the page divisions.

Long-term archival

PDF/A exists as a defined archival profile with a self-contained, deterministic subset of the format. Nothing equivalent exists for SVG. If a document needs to look identical in thirty years, PDF/A is the format designed for that job. SVG's permanence comes from its simplicity and open specification, which is a different kind of guarantee — the file will always be parseable, but the rendering may depend on which engine reads it.

A checklist before you send a file to a printer

1. Export as PDF/X-4 if the printer accepts it, otherwise PDF/X-1a.
2. Set the document bleed to the printer's stated value — usually 3 mm or 0.125 in.
3. Confirm the TrimBox matches the finished size and the BleedBox includes the bleed.
4. Embed or outline all fonts.
5. Confirm spot colours are named and that black is not a rich build unless intended.
6. Check effective resolution of every raster element at final size.
7. Flatten or note any transparency that the chosen PDF/X level may not support.

If your file is an SVG and the printer wants a PDF, the conversion is mechanical but the prepress decisions are not. Read our guide to print-ready artwork for the setup side, and the format reference if you are deciding between PDF, EPS, AI and SVG in the first place.

For cutting and crafting destinations rather than presses, SVG versus DXF is the more useful comparison, and the SVG optimisation guide covers trimming a file down before delivery.