How to impose a saddle-stitch booklet

You fold the printed proof, staple it through the spine, and page 3 is on the back of page 6. Or the booklet reads correctly but the outer pages have crept in so far that the folios on the inner spread are a good three millimetres short of where they should be, and the trim has clipped one of them.

Both are ordinary. The first is a fold-order problem and it is fixable in a minute. The second is creep, it is physics rather than software, and it needs a decision from you before you print. This post covers the method for both.

Why the page count must be a multiple of four

A saddle-stitch booklet is made of sheets folded in half and nested inside one another. One sheet, folded once, gives you four printable page faces: two on the front of the sheet and two on the back. There is no way to get three faces out of a folded sheet, and no way to get five. So the finished page count is always 4, 8, 12, 16, 20 and so on.

If your document is 14 pages, you are printing a 16-page booklet with two blanks, and where those blanks go is your decision, not the software's. Two blanks at the very back is the lazy answer and it usually looks like an error. Better options are a blank inside the back cover and a blank facing the title page, or one deliberate colophon page and one blank. Add the pages to the source PDF before you impose, so that what you hand to the imposition tool is already the real page count.

Choosing a signature size

The signature size is how many finished pages come off one press sheet. FastImpose defaults to 4pp, which is the plain saddle-stitch case: one press sheet, folded once, four pages, and the whole booklet is a nest of those sheets.

Larger signatures fold more times. An 8pp signature is one sheet folded twice, giving eight pages, and 16pp is folded three times. The reason to use them is straightforward: fewer, larger press sheets is cheaper and faster than many small ones, and it is how a booklet gets produced on a large-format press rather than a digital sheet-fed one.

The reason not to use them is equally straightforward. A folded-twice signature has to be trimmed on the head and the fore-edge to open, the folds have to be accurate to a millimetre or so or the pages will not square up, and if your finishing is a bench guillotine and a stapler you do not want it. Choose by your finishing kit, not by the page count.

Fold order, and what ends up next to what

The rule for a nested booklet is that the two page numbers on any one side of a sheet always add up to the total page count plus one. In a 16-page booklet that sum is 17: the outermost sheet carries 16 and 1 on the front, 2 and 15 on the back. The next sheet in carries 14 and 3, then 4 and 13. The innermost sheet is 10 and 7, 8 and 9 — the only true reading spread in the whole book.

For an 8pp form the same rule applies to the outer form: it carries pages 8 and 1 together, because in an eight-page booklet the sum is 9. Page 8 sits to the left of page 1 on the sheet, so that when the sheet is folded, page 1 falls on the front of the booklet and page 8 falls on the back.

1. Outer sheet, front: 8 and 1. Outer sheet, back: 2 and 7.
2. Inner sheet, front: 6 and 3. Inner sheet, back: 4 and 5.
3. Nest the inner sheet inside the outer, fold both together, staple through the fold.

You do not need to compute this — Booklet mode does it — but you do need to be able to check it, because a booklet that has been imposed against the wrong page count reads perfectly for the first two spreads and then goes wrong in the middle. Fold a proof and read it.

Creep: what it is, and what FastImpose does not do about it

When sheets are nested and folded together, the inner sheets have to travel around the outside of the ones inside them. Each nested sheet pushes the pages of the sheets inside it further out towards the fore-edge. The innermost spread sticks out furthest. Trim the finished book square and the innermost pages lose the most from their outer margin — this is creep, also called shingling or push-out.

Two things make it worse, and they multiply:

• Stock thickness. A 300 gsm cover creeps far more per sheet than 80 gsm text stock. Bulky uncoated and any board are the usual culprits.
• Page count. Creep is cumulative across the nest. A 12-page booklet on 100 gsm is often below the threshold anyone would notice. A 64-page booklet on the same stock is not.

A rough working estimate: the total push-out at the innermost spread is about the caliper of one sheet multiplied by the number of nested sheets, and each individual page shifts by roughly half of that. On a 48-page booklet — twelve sheets — on stock around 0.1 mm caliper, you are looking at somewhere near 1.2 mm at the innermost spread. That is enough to notice on a folio sitting 5 mm from the trim, and enough to lose a page number sitting 3 mm from it.

What to do instead

Three workarounds, in order of how often they are the right answer.

• Design the margin to absorb it. This is the honest answer for most short booklets. Keep live matter — folios, rules, anything with an edge the eye can measure — at least 8 mm from the trim rather than 4 mm. Creep of a millimetre is invisible on a page whose margins are generous and asymmetric anyway. It only shows when something straight sits close to the cut.
• Apply the shift upstream. If you own the source document, apply the compensation in the layout application before you export the PDF — InDesign's Print Booklet has a creep field, and exporting an already-shifted PDF means the imposition step just places pages that are already correct. You are using two tools for the two jobs each does.
• Trim to the fold, not to the nominal size. For a job where the content is forgiving, fold and stitch first, then trim the fore-edge on the guillotine against the folded book rather than to a measured dimension. You lose a little more from the inner pages, which is exactly what creep did anyway, but the finished edge is square and no page is short.

And the pragmatic threshold: if your booklet is 32pp or under on text-weight stock, creep is very unlikely to be the thing that ruins the job. Do the fold test, look at the innermost spread, and only reach for compensation if you can see it.

The cover spine option and the stock weight field

A self-cover booklet — cover printed on the same stock as the text — needs nothing special. Turn the cover spine option off and the imposition is uniform through the whole book.

A separate, heavier cover is different. A 300 gsm cover wrapped around a text block has to bend around the stitched edge, and the fold does not happen at a line but across a small width of material. The cover spine option adds an allowance at the spine for that. If you have used a heavy cover without one, you will have seen the result: the front cover artwork sits a fraction to the left of where it should, and the back cover a fraction to the right.

The stock weight field — 1 to 1000 gsm, defaulting to 120 — only becomes editable once cover spine is switched on. Be clear about what it does: it records the stock you specified. It is not fed into the sheet geometry. It is there so the setting travels with the job and the value is not lost between the estimate and the press. Do not enter a weight expecting the layout to change in response.

Why sheet numbering is unavailable in Booklet mode

If you try to switch sheet numbering on with Booklet mode selected, FastImpose refuses the job rather than printing it. That is deliberate, and the reason becomes obvious the moment you picture the finished piece.

Sheet numbering exists so that a stack of press sheets can be kept in order on the bench — a number printed in the waste area outside the trim, one per sheet, so the operator can see at a glance that nothing has been shuffled. It relies on the number being cut away.

A booklet signature is not a sheet that gets cut apart. It is folded and trimmed as one piece and it becomes several pages of a finished book. A per-sheet number has nowhere to live: put it in the margin and it is inside the bound book on someone's page 11. There is no waste area to hide it in, because the whole sheet becomes product. Hence the validation. If you need to keep track of a booklet run, use the slug line instead — it can carry the job name and go in the area you trim off.

The walkthrough

1. 1

   Get the source PDF to a multiple of four

   Add the blank pages yourself, in the position you want them, before importing. Do not let a filler page land somewhere you did not choose.

2. 2

   Pick the sheet size and check it against the finished page

   Two A5 pages side by side need A4. Two A4 pages need A3. Then add the bleed and the mark distance on top and confirm it still fits — an A4 sheet is exactly two A5 pages with nothing left over, so if you want bleed and crop marks you need a sheet larger than A4. That is usually the point at which a Custom sheet size such as SRA3 at 320×450 mm is the answer.

3. 3

   Choose the signature size

   4pp unless your finishing can genuinely fold and trim larger. This is the field people change without needing to.

4. 4

   Set the cover spine if the cover is a different stock

   On for a separate heavier cover, off for a self-cover. If you switch it on, the stock weight field unlocks — record the actual gsm there so the value stays attached to the job.

5. 5

   Set bleed and marks

   The bleed allowance is a single value applied all round; the default is 0 mm. Crop marks default to 5 mm long, 3 mm clear of the trim, 0.25 pt thick. If your bleed is 3 mm the marks will start exactly where the bleed ends — see the crop marks post for why that particular pairing causes trouble.

6. 6

   Turn duplex on

   A booklet is printed both sides by definition. If your press or driver flips on the long edge you will also want the 180° back-flip; the duplex post has a one-sheet test that settles which.

7. 7

   Export, print one copy, fold it and read it

   Every single time, before the run. Check the page order end to end, then open the innermost spread and look at the fore-edge margin against the outermost. The difference you see there is your creep, measured on the actual stock rather than estimated.