The Print-Ready Manuscript: A Technical Guide to PDF/X, Bleed, and Colour Compliance

PDF/X-1a: The Manufacturing Contract

PDF/X is not a separate file format. It is a constrained subset of PDF defined by ISO 15930, first published in 2001 and revised several times since. The "X" stands for exchange — the standard exists to guarantee that a PDF can move from originator to printer without ambiguity. PDF/X-1a (ISO 15930-1:2001, updated as ISO 15930-4:2003) is the most restrictive and most widely accepted variant. It mandates that all fonts be embedded, all colours be specified in CMYK or spot-colour space, no transparency be present, and no external content references exist. The file must be entirely self-contained.

The rationale is industrial, not aesthetic. An offset lithographic press exposes four plates — cyan, magenta, yellow, and key (black) — each corresponding to one ink channel. If a PDF contains RGB data, the raster image processor (RIP) at the print facility must convert it, and different RIPs produce different CMYK approximations of the same RGB value. PDF/X-1a eliminates this variable by requiring the conversion to happen before the file leaves the originator's control. The Ghent Workgroup, an international consortium of prepress professionals, recommends PDF/X-1a as the minimum exchange standard for commercial print (Ghent Workgroup, "PDF/X-1a Guidelines," 2019).

Why RGB-to-CMYK Conversion Is Non-Trivial

RGB and CMYK are fundamentally different colour models. RGB is additive: red, green, and blue light combine to produce white. CMYK is subtractive: cyan, magenta, yellow, and black inks absorb wavelengths from white paper to produce colour. The gamuts do not overlap perfectly. Saturated blues and greens that display vividly on an RGB monitor have no exact CMYK equivalent — they must be approximated, and the approximation is governed by an ICC (International Color Consortium) profile that maps one colour space to the other.

The standard ICC profile for North American commercial print is GRACoL 2006 (General Requirements for Applications in Commercial Offset Lithography), maintained by Idealliance. For European work, the Fogra39 profile (ISO 12647-2:2004) is the equivalent standard. IngramSpark's published specifications require PDF/X-1a with a CMYK colour space and recommend the GRACoL profile. Amazon KDP is more permissive — it accepts RGB PDFs and performs server-side conversion — but this permissiveness introduces unpredictability. A cover designed with a specific navy blue in RGB may arrive at the reader as a dull, greenish approximation after KDP's automated conversion. Professional practice dictates controlling the conversion oneself.

The conversion engine most widely available outside commercial prepress suites is Ghostscript, an open-source PostScript and PDF interpreter maintained by Artifex Software. Ghostscript's pdfwrite device, combined with a CMYK ICC profile and a PostScript preamble defining output intent, can convert a Typst-generated PDF to a compliant PDF/X-1a file. The command is not intuitive — it requires specifying the output intent, the ICC profile path, colour conversion strategy, and several compatibility flags — but it is deterministic and automatable. PagePerfect's pipeline uses precisely this approach, executing the conversion as a post-compilation step before the file reaches the user.

Bleed: The Geometry of the Guillotine

Bleed is the extension of printed content beyond the intended trim edge of the page. It exists because guillotine cutters — the industrial paper trimmers that cut printed sheets to final size — operate with a mechanical tolerance of approximately 1 to 2 millimetres. If the printed image ends exactly at the intended trim line and the cut lands 1.5mm inside that line, a white strip appears at the edge of the finished page. Bleed prevents this by extending the image 3mm (or 0.125 inches, the American convention) beyond the trim on all four sides.

The PDF specification accommodates bleed through a hierarchy of page boxes. The MediaBox defines the total extent of the page including bleed. The TrimBox defines the intended final dimensions after cutting. The BleedBox sits between them, marking the region where bleed content exists. A properly constructed PDF/X-1a file contains all three boxes, and the difference between TrimBox and BleedBox must be at least 3mm on each side. IngramSpark's submission requirements are explicit on this point: files without the correct TrimBox and BleedBox will be rejected at upload.

For book interiors, bleed is typically relevant only when content extends to the page edge — full-bleed images, coloured backgrounds, or decorative rules that touch the margin. Standard text-only interiors do not require bleed because the text block sits well within the trim area. However, the cover always requires bleed: the cover image must extend 3mm beyond the trim edge on all four sides, plus account for the spine width (which varies with page count and paper stock) and any wrap-around artwork. As discussed in "The Geometry of Authority," the margins of a printed page are not arbitrary white space — they are engineering tolerances made visible.

Font Embedding and the Self-Contained File

PDF/X-1a requires all fonts to be embedded in the file. This means every glyph used in the document — including ligatures, small caps, and mathematical symbols — must be physically present in the PDF as a font subset or complete font programme. If a font is referenced but not embedded, the RIP will substitute a default (typically Courier), destroying the typographic design.

Typst embeds fonts by default when generating PDF output, a key reason it is preferred for professional typesetting. However, embedding is not the same as licensing. Some commercial fonts prohibit PDF embedding in their licence terms (the fsType flag in the OS/2 table of the OpenType specification). A preflight check must verify not only that fonts are embedded but that their embedding permissions allow it. The preflight system discussed in "The False Economy of the Software Default" addresses this as part of a broader validation pipeline.

The Preflight Pipeline

Preflight is the systematic verification of a PDF against a set of compliance criteria before it is submitted for printing. The term originates from aviation — the checklist a pilot completes before take-off — and its adoption by the prepress industry reflects the same philosophy: catch errors while correction is still cheap.

A thorough preflight for book manufacturing checks at minimum: PDF/X-1a conformance (output intent, colour space, font embedding), trim and bleed box dimensions, page count divisibility (signatures for offset; even page counts for POD), image resolution (minimum 300 DPI for halftones, 1200 DPI for line art), text-to-trim distance (no text within 3mm of the trim edge), and spine width calculation based on page count and paper stock. Each of these checks can be automated, and each catches errors that would otherwise result in rejection, reprinting, or — worst case — a print run of books with misaligned covers, cropped text, or colour shifts.

The alternative to automated preflight is manual inspection by the printer's prepress department, which introduces delays of one to five business days and costs that are ultimately borne by the author. An automated pipeline that validates at the point of export — before the file leaves the author's control — eliminates this round-trip entirely.