Thread: libpoppler/pdftops spool size issues when printing large PDF files.

this has been annoying me years, , i'm wondering if knows of fix insane spool sizes generated pdftops (i assume) when trying print pdf files containing bitmap graphics.

see time when printing lecture notes uni classes, few print fine, if there 1 graphic, takes ages (easily 1/2 hour). not issue printers, it's been happening years, , i've experienced on 4 different printers @ work , 2 @ home (bw lasers, colour lasers , colour inkjets toshiba, sharp, brother, hp , fuji-xerox).

examining print queue in windows vm whilst printing ubuntu yielded interesting results. print settings in cases were: a4, colour, duplex, long edge binding. source 23/196 pages of 5.41 mib pdf containing selectable text (i.e. not scan) , pictures.

spool sizes:
windows vm, acrobat reader x (10.1.1) -> 2.24 mib, printing starts immediately.
linux, acrobat reader 9 (9.4.2) -> 28.9 mib, printing slow, pages print received.
linux, evince/document viewer 2.32.0 (poppler/cairo 0.16.4) -> 325.6 mib, seems wait entire file before printing start. holds printer 1/2 hour or more.

text-only documents print fine (i.e. no discernible difference windows) documents raster graphics (e.g. thesis has .eps graphics , prints fast). .tex->pdf , .tex->dvi->ps->pdf both print fine.

checked around , found old bug on launchpad (from karmic days) looks nothing being done about.
https://bugs.launchpad.net/ubuntu/+s...ce/+bug/516280

there many factors involved know start description - not criticism aimed @ you, happens bit of minefield area.

example, pdf may contain pdf transparency constructs (alpha blending, , like), , such transparency not part of postscript imaging model. "print" such pdf using postscript (or other page description languages) transparency needs "flattened" 1 or more opaque imaging operations. in cases, may involve rendering pdf content high-res raster image, , wrapping in enough postscript render image. obviously, full page, high resolution image going larger original vector operations. if take "smarter" approach, , try degenerate transparency effects series of opaque vector operations, file size (in trivial cases) considerably larger.

above made worse fact pdf may not contain actual transparency operations, might still contain pdf transparency constructs (i won't go details unless wants me to!) - , enough trigger "transparency flattening" mentioned above.

pdf generated cairo/poppler particularly prone this: wrap (almost?) marking operations in pdf transparency constructs, when there no transparency on page description. worse, pdf transparency includes bounding box pdf interpreter knows required apply blending operations, cairo pdf sets such bounding boxes full page size, regardless of size of actual marks on page (this not transparency bounding boxes, others well). mind, bad form - bounding box should be: "the smallest rectangle (oriented axes of glyph coordinate system) enclose entire shape". not contrary spec. mean such case "render image" flattening approach taken, end 1 or more high-res images full size of page.

ghostscript, have discussed several times pre-scanning pdfs find, not pdf transparency constructs (which do) whether actual blending required constructs, , if not, disable transparency handling. problem is, there considerable overhead these checks, , general feeling should not penalise "well formed" pdfs other creators handle poorly formed pdfs 1 library.

there other possibilities: "to postscript" converters not preserve fonts, characters end either small bitmaps or being degenerated vector operations. less efficient repeatedly draw outline of character draw once in font, , reuse within interpreter.

also, "to postscript" converters emit level 2 postscript, emit level 1, means things shaded fill patterns have (usually) rendered , emitted image (shaded fills introduced in postscript level 3).

believe oneiric, print spooler uses ghostscript (with ps2write device) instead of poppler based pdftops. *may* worth installing in virtual machine see whether situation improves.

if want make example of problem pdf's available me, can take , see if there obvious reason large spooler size. promise not share it, , delete system when done.

chris

p.s. sorry long post.......

Forum The Ubuntu Forum Community Ubuntu Official Flavours Support General Help [all variants] libpoppler/pdftops spool size issues when printing large PDF files.

Ubuntu