Document recreation

‹ All services

We rebuild documents that have no source files left: scans, PDFs, material from software nobody opens any more. You get a fully editable file, faithful to the original and ready for translation and further layout work.

When the source file no longer exists

The studio that produced the original has closed down. The files are on a drive nobody can locate. A twenty-year-old document exists only as a printout. The end client sends a PDF and the news that this is all there is.

In every one of those cases the work stops: the content cannot be updated, the file cannot go to translation, the next language version cannot be produced.

Why automated tools are not enough

PDF-to-InDesign converters work on digitally generated files and simple layouts. Given a scan they return an image instead of text, because they do not recognise characters. Right-to-left scripts and CJK are outside their scope. Embedded fonts and linked files are left untouched. In longer documents they break tables, split paragraphs at arbitrary points and lose the style hierarchy.

From a distance the result looks like the original. At the first edit it turns out to be unusable.

How we do it

We start with character recognition — OCR, followed by correcting whatever the recognition got wrong. We check it against the original, because an error introduced at this stage will travel through every language version that follows.

Then we rebuild the document: frames, paragraph and character styles, tables, numbering, cross-references, image positions. Graphics are extracted and replaced with material at the right resolution and, where needed, redrawn as vectors. Fonts are matched to the original or to your specification.

Finally we compare the rebuilt file against the source, page by page.

What you get

A file in the format you need for the next stage: InDesign, FrameMaker, Word. Fully editable, with styles, structure and links intact — not a collection of text frames dropped onto a page.

That file goes straight to translation and comes back to layout without extra work. On a publication localised into a dozen languages, that is the difference between rebuilding once and correcting a dozen times.