Print-ready PDF/X-3 or EPS from a raster logo

The print files described here were produced by the tracer's print pipeline (Ghostscript 9.55 with lcms2 for the PDF/X conversion) on its test package of CC0 and public-domain images, 18 October 2026. Whether a given press accepts a file is still the printer's preflight to decide; this guide tells you what the file contains so that conversation is short.

A print shop wants a vector file with exact geometry, device colour it can trust and no transparency for its RIP to guess at. From a raster logo you get there by tracing it with the Print purpose of the image to SVG vectorizer and downloading PDF: the tracer keeps the geometry as traced, flattens any transparency onto paper white, converts the colours to CMYK with the FOGRA39 characterisation and writes a PDF/X-3 file. EPS is there for sign software and older RIPs, and SVG for your own checks on screen.

What PDF/X-3 is and why it is the one to send

PDF/X is a family of ISO standards (ISO 15930) that restricts PDF to what a press workflow can reproduce without surprises: fonts embedded, no live transparency in the older parts, page boxes defined, and an output intent that says which printing condition the colours were prepared for. PDF/X-1a allows only CMYK and spot colour; PDF/X-3, the 2002 part the tracer writes, additionally allows colour-managed data with ICC profiles, which is the right fit for a logo that started life as an sRGB image.

The output intent in the tracer's files is FOGRA39, the characterisation data behind the ISO Coated v2 (ECI) profile for coated paper on offset presses. It is the most widely assumed condition in European commercial printing, and a printer who works to something else (a newsprint profile, a large-format inkjet) can convert from it. If the shop asks for a specific profile, say which one you have and let them re-separate; do not try to fake it by editing the file.

Transparency flattening and why it matters

A logo with a transparent background looks fine on screen and causes a specific kind of trouble in print: the RIP has to decide what the see-through pixels sit on, and different RIPs decide differently. Some composite onto white, some onto black, some rasterise the whole page at a low resolution to resolve it. The tracer removes the question. In the Print purpose it flattens transparency onto paper white before export, so the file contains opaque, fully specified shapes and nothing else. If you want the logo on a coloured background, add that background in the layout program where the logo is placed.

Flattening is also why the Print purpose does not run the cut clean-up: print keeps every thin line and every small shape the tracer found, because a press resolves far finer detail than a knife. The engine used for print is the stacked vtracer 1.0 engine, chosen on the test set because it reproduces layered artwork, logos with shapes on top of shapes, with the fewest seams.

When to choose EPS, and when SVG

EPS is Encapsulated PostScript, the file format sign-making and embroidery software grew up with. It has no transparency, no colour management of its own and no page boxes to speak of, which is exactly why some older workflows still prefer it: there is nothing in it to misread. Choose EPS when the shop or the software asks for it by name. Choose PDF otherwise; it carries the output intent and the page size, and every modern RIP reads it directly.

SVG from the Print purpose is the same geometry in sRGB, meant for checking on screen and for the web. Do not send it to a press as the print file; browsers and layout programs disagree about SVG units and colour, and it has no output intent. The SVG tools page can convert it to PNG or JPG if you need a proof image to email.

Setting the size

A PDF has a page size, so the tracer needs to know how big the logo is. Enter the design width in millimetres on the tool page and the height follows the image ratio; the page is the design's bounding box. Without a width the tracer assumes 300 dpi, which makes a 1,000 px logo 84.7 mm wide. For a business card that is too big and for a banner far too small, so type the width you need. Scaling a vector in the layout program afterwards is lossless, but a page size that already matches saves a question.

Bleed is not added by the tool. Bleed belongs to the page the logo sits on, not to the logo, so add it in the layout program along with the trim marks the printer asks for.

What a good source looks like

The tracer reproduces what it sees. A flat, sharp logo with clean edges and a few solid colours traces into exact shapes. A JPEG with compression blocks around the edges traces into shapes with blocky edges, because to the tracer those blocks are the edges. A gradient becomes a staircase of flat tones; the tracer keeps gradients it is sure about, and the result panel lists how many it found. Use the largest and cleanest copy you have, crop to the logo with the crop tool, and avoid anything that was screenshotted.

One thing the tracer cannot restore is text. Letters are traced as outlines, which is correct for print (no fonts to embed, nothing to substitute) but means the type is no longer editable. If the shop needs to change a word, they need the original artwork, not the trace.

Before you send it

• Open the PDF and check the preview against the original; the result panel's status should be Done.
• In Acrobat, run Preflight with a PDF/X-3 profile, or ask the printer to; the file should report its output intent as FOGRA39 (ISO Coated v2) and no transparency.
• Confirm the page size is the size you entered, and that any bleed was added in your layout.
• Mention that colours were converted from sRGB, so the shop knows a bright screen orange will print as the nearest coated CMYK orange.

The same job in the Cut purpose produces a very different file, built for a knife rather than a press; the cut guide covers it. For a 3D-printable version of the logo, the traced SVG can go straight into the image to STL tool.