smpdf

[maadjordan]

First I'd like to introduce SMPDF "Lossless PDF Compressor" (its visually lossless) you can refer to more details at https://github.com/coherentgraphics/...sPDFCompressor and its "community release" at http://community.coherentpdf.com/ there is binaries for different OS.

here is a sample on SWF & FLV pdf files at
http://download.macromedia.com/f4v/v...spec_v10_1.pdf
original = 608,962 bytes
FO(8.40)= 608,962 bytes
SMPDF = 535,223 bytes
SMPDF after FO = 535,223 bytes (because FO could not reduce it).

http://wwwimages.adobe.com/www.adobe...ormat-spec.pdf
Original = 1,724,258 bytes
FO(8.40) = 1,677,408 bytes
SMPDF = 1,488,957 bytes
SMPDF after FO = 1,467,600 bytes

[nikkho]

Sounds interesting. Will take a look.

[SvenBent]

Sounded good until "its visually lossless"

Every PDF optimizer that had said it did optimizations that was indistinguishable from the original was always.. well distinguishable. high res graphics was just not high res afterward when viewing it on a monitor with 100% zoom level.

My question is how visually degrading is it? What tricks does it do is output lossless and what does it do that is visual lossless?

--- edit ---
to asnver some of my own questions

unlike other PDF compressors, Lossless PDF Compressor:
will not resample, reduce, or alter your images in any way
will not prevent your PDF file from being edited afterward
will not remove bookmarks, outlines, or other metadata

Highly enjoy the bolded ( by me) part.

[maadjordan]

I used cpdf to decompress pdf streams and then smpdf to recompress but i got a minor gain till leanify is added to smpdf or cpdf then smpdf is enough https://github.com/johnwhitington/cpdf-source/issues/67

[necros]

hmm, it have an issue - smaller filesize on multiple runs on the same file, bug? last upd. on github was in 2015 8-(

[nikkho]

hmm, it have an issue - smaller filesize on multiple runs on the same file, bug? last upd. on github was in 2015 8-(

Noticed that too. Surprisingly, it does not happen in all PDF I tried.

[maadjordan]

CPDF v2.3 has been released.

https://coherentpdf.com/blog/?p=92

bin for Win,Mac & Linux :

https://github.com/coherentgraphics/cpdf-binaries

[maadjordan]

It seems that unicode file name support was missed during compilation. A temporary windows compile is available through this link

http://www.coherentpdf.com/16thSeptember2019.zip

[necros]

pls support my bug report, author closed and ignores it but it still exist in 2.3 (multiple iterations of optimizing same file give random results)

https://github.com/coherentgraphics/...ries/issues/12

[maadjordan]

I tired before to replicate your case on my files but with no success.

currently the new compile give the following
original 350,140
1st 347,923
2nd 347,921
3rd 347,921
4th 347,918
5th 347,915
6th 347,916
7th 347,918
8th 347,918
9th 347,915
10tg 347,922

but with PDFsizeopt i get 344,569
CPDF after PSO gives 344,674

CPDF can be used for quick optimizations and repair some damaged files but PSO can gain way more reduction on most cases.

the file still has meta data after optimization and no tool can remove them safely.

but still running multiple run be added to File-Optimizer to automate the file size comparing process.

or you can write a batch script to do so. (i think such script has been made in this forum before)

[necros]

this doesn`t answer the question why every iteration gives different size)

[maadjordan]

the conclusion from my post the PSO can get better result than running cpdf several times and the gain much better. The difference in cpdf is mainly the result reordring streams.

[necros]

so why don`t we add internal reordering system to choose less size file as output?