**List of steps to reproduce** (step by step, including full links if applicable):
Hello, when I transfer large djvu files (more than 500 pages) it is common that the text layer of the pages does not match the page images.
*I download a pdf file from Gallica
*I extract the pages to process them with Briss and Scantailor,
* I supress the two disclamer pages of Gallica
*I make the Ocr and the djvu with Abbyy finereader 15.
*I check the djvu with DjView.
*I upload the djvu file on Commons
*If there is a mitchmatch I remake the ocr of djvu created with finereader with Tesseract and if fix the problem, but the Ocr is less performant than Finereader one.
exemples : https://commons.wikimedia.org/wiki/File:La_Revue_Ind%C3%A9pendante,_tome_1_-_mai_%C3%A0_octobre_1884.djvu
I precise that i check the correct matching of images page and tex-layer produced by finereader, with DjView 4.11 (DjVulibre) and there is no problem.
Sorry for my poor english.
**Software version (if not a Wikimedia wiki), browser information, screenshots, other information, etc.**:
*Abbyy Finerader 15