Page MenuHomePhabricator

IA uploader downgrades the quality of uploaded PDF file
Open, Needs TriagePublicBUG REPORT

Description

Compare e.g. the tile page of https://archive.org/details/bohemiaunderhaps00capeiala/mode/2up with the title page of the uploaded file https://upload.wikimedia.org/wikipedia/commons/3/3e/Bohemia_under_Hapsburg_misrule_%281915%29.pdf (or any other pages of the same files)

Event Timeline

Jan.Kamenicek renamed this task from IA uploader downgrades the quality of uploaded file to IA uploader downgrades the quality of uploaded PDF file.Apr 11 2021, 10:52 PM
Jan.Kamenicek removed a project: PDF-Rendering.

The degradation is not due to IA-Upload, it's the Internet Archive's own heavy compression (~90MB of original images becomes 6,5MB of PDF) during PDF generation that causes this. The PDF from the IA is the same: https://archive.org/download/bohemiaunderhaps00capeiala/bohemiaunderhaps00capeiala.pdf

The scan pages you see at the IA are JPEGs generated directly from the JP2 scan photos, so they look a lot better in the IA "book reader" mode.

I guess we could look at generating a new PDF; is that worth it? It seems a similar better-quality file could be achieved by using DjVu.