Thanks. I added some metadata and it blew up to 14M, which is unfortunate. Chart I does seem to be missing in both the PDF and the online version; I suspect that it’s missing from the physical copy at UMich (‘pocket’ sounds like something that might go missing).
The object streams for indirect objects have been unpacked and stripped away, leaving their contents uncompressed. Use qpdf to regenerate compressed object streams:
qpdf --object-streams=generate in.pdf out.pdf
(The --stream-data=compress option is already set by default.)
While you are at it, might as well re-linearize the PDF for online readers with low bandwidth:
Got the whole PDF from HathiTrust. I think Chart I is missing from the scan.
Thanks. I added some metadata and it blew up to 14M, which is unfortunate. Chart I does seem to be missing in both the PDF and the online version; I suspect that it’s missing from the physical copy at UMich (‘pocket’ sounds like something that might go missing).
The object streams for indirect objects have been unpacked and stripped away, leaving their contents uncompressed. Use
qpdfto regenerate compressed object streams:(The
--stream-data=compressoption is already set by default.)While you are at it, might as well re-linearize the PDF for online readers with low bandwidth:
That seems to work. I tried
gs, Gscan2pdf, andpdf2djvubut they all either didn’t reduce size or segfaulted.