Translation and professional typing (OCR-DTP) of scanned German instructions for calibrating a spectrometer into the Russian version while maintaining the layout

Initial situation - the laboratory uses a Bruker Vertex 70 spectrometer, all service documentation for calibrations and diagnostics is available only in German. There are 3 PDF files with scans of printed instructions (about 180 pages in total), the quality is different - distortions, shadows from the binding, small font and tables in places. Documents are needed in the form of Russian-language files for internal training and quality audits, preserving the structure, numbering, cross-references, formulas, notations and tables.

Technical environment and subject area - analytical instrumentation, optics/IR spectroscopy, calibration procedures (Wavenumber calibration, aperture alignment, detector checks), deviation logging, ISO 17025 requirements. Skills in working with OCR and DTP, terminological accuracy and understanding of the meaning of the procedures are required so as not to “break” the instructions with literal translation.

What you should get is a set of Russian-language documents in DOCX format and the final PDF, visually as close as possible to the original. In parallel - a table of DE-RU terms (minimum 80 terms) with uniform translations and notes on controversial cases.

Execution steps
1 - Preprocessing of scans - alignment, background cleaning, splitting/gluing pages if necessary, preparation for OCR.
2 - OCR with manual proofreading - restoration of text, formulas, Greek symbols, indices/degrees, special characters and units of measurement; neat set of tables.
3 - Translation of DE-RU - maintaining the uniformity of terms, preserving warnings/notes, correct transmission of operating modes and parameters, without changing the meaning of numbers, tolerances and units.
4 - Layout and quality control - maintaining the structure, headings, numbering, captions to pictures, links like “see. section 3.2", uniform table style; final check for inconsistencies.

Restrictions
- You cannot change numerical values, tolerances, designations of buttons/menu items of the instrument software and file names - only translate the explanatory text.
- The terms must be consistent throughout the set - one term per translation, unless justified otherwise.
- All controversial points should be recorded in a separate list of questions (up to 20 points) before the final submission.
- Confidentiality - the source and the result should not be published or used in the portfolio.

Measurable acceptance criteria
- 100% of the original PDF pages were transferred to the set while maintaining the logic of the structure - sections/subsections, numbering, signatures, tables.
- OCR errors in critical places (numbers, units, procedural steps, warnings) - no more than 2 for the entire set; all found ones have been corrected.
- Terminological glossary - at least 80 lines DE-RU, applied uniformly.
- Visual match of the final PDF with the original in terms of page structure - acceptable discrepancies in line breaks/fonts, but tables, headings, numbering and signatures must be in the same places and readable.

Entry files - 3 PDF scans; if necessary, I will provide 10-15 pages for a test pass at the start.
Output files - DOCX for each PDF, single PDF, glossary in XLSX/CSV, list of questions/comments in a separate file.