Why your crop marks land in the wrong place

The proof comes off and the corner marks are sitting in the middle of the bleed, printing over the artwork. Or they are out where they should be but nobody at the guillotine can see them on the uncoated stock. Or they moved when you scaled the job to 96% and now they no longer describe the trim at all.

Crop marks are the simplest thing on an imposed sheet — two short lines at each corner — and they go wrong more often than almost anything else, because a mark is only meaningful relative to something and it is rarely obvious what that something is.

What the three numbers mean

FastImpose draws crop marks and nothing else. Three values control them.

So with the defaults, a mark occupies the band from 3 mm to 8 mm outside the trim line. Everything inside 3 mm is left clear. That clear band exists for a reason: a mark that ran right up to the trim would be cut through, leaving a stub of ink on the finished piece.

Marks that sit inside the bleed

This is the most common complaint and it has an exact arithmetic cause.

Bleed extends the artwork outward from the trim. A 3 mm bleed means the artwork continues for 3 mm past the cut line all the way round. The mark distance also measures outward from the trim. So a 3 mm bleed and a 3 mm mark distance means the artwork stops at exactly the point where the mark starts. They abut. There is not a millimetre between them.

The fix is to set the distance greater than the bleed. In practice:

• 0 mm bleed: 3 mm distance is fine, and marks sit in clean white.
• 3 mm bleed: set the distance to 4 or 5 mm. One or two millimetres of clear paper between the end of the bleed and the start of the mark is all it takes.
• 5 mm bleed: set the distance to 6 or 7 mm — and check the sheet still has room, because you have now pushed the outermost mark 12 mm or more past the trim.

The general rule: distance > bleed, always, with at least 1 mm of daylight. If you only remember one thing from this post, that is it.

The sheet room check

Then verify the sheet is big enough. The space you need outside the trim on each edge is bleed, plus the gap between bleed and mark, plus the mark length. With a 3 mm bleed, a 5 mm distance and a 5 mm mark, that is 10 mm on every edge — 20 mm added to both the width and the height of the imposed block.

Two A5 pages side by side are exactly A4, with nothing left over. Add 20 mm each way and A4 will not hold it. That is when you go to a Custom sheet size — SRA3 at 320×450 mm is the usual answer for A3-sized work with marks, and it is not a preset, so you enter it as a custom size. There is more on the presets in the paper sizes reference.

Marks that appear to move when you scale

You set the scale to 95% to make a job fit, and the marks are no longer where you expected. They are not broken — they are doing exactly what they should, and the confusion comes from what you thought they were attached to.

Scale shrinks the placed page. A page scaled to 95% has a smaller trim box on the sheet, so the trim line itself has moved inwards. The marks follow the new trim, because the trim is what they describe. What they do not do is stay pinned to the artwork's original position, and they do not scale their own 5 mm length and 3 mm distance — those are sheet measurements, so 3 mm stays 3 mm.

The practical consequence is that at 95% the marks are relatively further from the artwork than they were at 100%, and on a multi-up sheet the whole grid contracts while the mark geometry stays constant. It looks like the marks drifted. They did not; the trim did.

Marks the operator cannot cut to

0.25 pt is a defensible default. It is about 0.088 mm, it renders cleanly on a coated stock at 300 dpi and up, and it is fine enough not to leave a visible stub if the cut is a fraction off.

It is also, in three situations, too fine to work with.

• Textured and uncoated stock. A laid or linen finish has surface variation larger than the mark. The line breaks up into dots and someone at the guillotine is squinting at it under a fluorescent tube.
• Toner presses at lower resolutions. A 0.25 pt line is under a pixel at 300 dpi and gets rendered as whatever the RIP decides — which may be a lighter grey line, or an inconsistent one.
• Light or tinted mark colour on tinted stock. A fine black line on cream or grey stock has much less contrast than the same line on white.

The fix is to raise the thickness. 0.5 pt is visible on almost anything and still cuts away cleanly. Going above about 1 pt is counterproductive: a thick mark has a wide centre and the operator has to decide which part of it to cut to, which reintroduces the ambiguity you were trying to remove. And a thick mark that is not cut entirely away leaves a visible black stub on the finished edge.

Ask the guillotine operator. This is a two-second conversation that saves a reprint, and the answer is a property of the stock and the machine, not of the file.

Marks lost because the sheet was trimmed to the wrong box

The last failure is the one where the marks were correct and then disappeared.

A PDF page carries several boxes — MediaBox, CropBox, BleedBox, TrimBox, ArtBox — and different software honours different ones. If the marks are drawn 3 mm outside the trim and something downstream crops the sheet to its TrimBox, the marks were outside that box and they are gone. Same story if a print driver is set to fit-to-page and shrinks the sheet to the printable area: the outermost content, which is the marks, is what gets clipped.

1. Open the exported sheet and check the marks are actually in the file before blaming the imposition. If they are there, the loss happened downstream.
2. In the print dialogue, turn off fit-to-page, scale-to-fit and auto-rotate. Print at actual size.
3. If the sheet genuinely does not fit the printer's imageable area, the sheet is too small for the job — go up a sheet size rather than letting the driver shrink it, because shrinking moves every trim line and the marks stop describing the cut.
4. If a downstream tool is cropping to a box, check the format notes for which boxes the imposed sheet carries.

How FastImpose draws marks, and what that buys

Worth knowing because it explains several of the behaviours above. FastImpose does not stamp marks onto the source pages. Each source page is placed on the sheet as a Form XObject with a transform — a position, a rotation, a scale — and the marks are drawn on the sheet from that same transform.

Three consequences, all of them useful:

• The source PDF is never modified. Nothing is drawn into the client's artwork. The page content is copied through as vector, and the marks are separate sheet-level geometry sitting outside it.
• Marks always agree with the actual placement. Rotate a page, scale it, move it with centre-in-sheet, flip the back of a duplex sheet — the marks are computed from the same transform, so they cannot get out of step with where the page really is. Marks stamped onto a page before placement can, and that is a classic source of a sheet where the marks are 2 mm off in one direction across the whole grid.
• Nothing is rasterised. A 0.25 pt mark is a 0.25 pt vector stroke in the output, so it stays crisp at whatever resolution the RIP works at rather than being baked to pixels at export.

The limit of this is worth being equally clear about: FastImpose draws crop marks only. There are no registration marks, no fold marks, no collation marks and no colour bars. If your press operator needs a colour bar on the sheet, it has to come from the RIP or be placed in the artwork before imposition. Full detail is in the crop marks reference.

A checklist before you commit the run

1. Mark distance is greater than the bleed by at least 1 mm.
2. The sheet has bleed + distance + length of clear space on every edge — 10 mm or so with typical values.
3. Scale is 100% unless you have a specific reason it is not.
4. Thickness suits the stock — 0.25 pt on coated, 0.5 pt on anything textured or uncoated.
5. The driver is set to actual size, no fit-to-page, no auto-rotate.
6. You have printed one sheet and physically looked at a corner.