Email or username:

Password:

Forgot your password?
Alex's posts Post Back to profile
Top-level
Alex Schroeder

OK, I think I found a website with the answer I was looking for. The key phrase is in the title of this post: "Serving pre-compressed files using Apache". Sometimes searching for stuff is hard just because you don't know what it's called. ๐Ÿ˜…

https://feeding.cloud.geek.nz/posts/serving-pre-compressed-files-using/ 
AddEncoding gzip gz
Options +Multiviews
SetEnv force-no-vary
Header set Cache-Control "private"
<FilesMatch "\.pdf\.gz$">
ForceType application/pdf
</FilesMatch>
1 November at 9:46 | Open on social.alexschroeder.ch
4 comments
Alex Schroeder

And now that I've gone through the directories, it turns out that those PDFs are already pretty compressed! Most of their content is probably compressed image files. I guess I should have known.๐Ÿ˜…
https://alexschroeder.ch/view/2024-11-01-gzip

1 November at 10:08 | Open on social.alexschroeder.ch
Alex Schroeder

I guess I could use Ghostscript to reduce PDF file size! Sure, it wouldnโ€™t be a bit-exact copy but perhaps archival copies donโ€™t need to be.

https://alexschroeder.ch/view/2017-09-06_PDF_File_Size

1 November at 20:34 | Open on social.alexschroeder.ch
Alex Schroeder

First test with Ghostscript to reduce filesize is disappointing. It also turned text in the original to a badly pixelated image. I wonder why. I had hoped it would only affect images. Now I'm reading https://ghostscript.com/blog/optimizing-pdfs.html and my head is smoking.

3 November at 13:05 | Open on social.alexschroeder.ch
Alex Schroeder

OK, based on that blog post, I have the following invocation. This doesn't turn the fonts into images but it does downsample all images to 150dpi if they are at 165dpi or larger, and it forces JPEG if there images are uncompressed or PNG (I think).

Let me know if there's anything else you can think of.

function pdf-shrink --description 'Turn images in PDF files to 150dpi JPG'
# https://ghostscript.com/blog/optimizing-pdfs.html
gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dNOPAUSE -dQUIET -dBATCH \
   -dConvertCMYKImagesToRGB \
   -dGrayImageFilter=/DCTEncode -dColorImageFilter=/DCTEncode \
   -dAutoFilterGrayImages=false -dAutoFilterColorImages=false \
   -dDownsampleColorImages=true -dDownsampleGrayImages=true -dDownsampleMonoImages=true \
   -dColorImageResolution=150 -dGrayImageResolution=150 -dMonoImageResolution=150 \
   -dColorImageDownsampleThreshold=1.1 -dGrayImageDownsampleThreshold=1.1 -dMonoImageDownsampleThreshold=1.1 \
   -dPreserveHalftoneInfo=false -dPreserveOverprintSettings=false \
   -dTransferFunctionInfo=/Apply -dUCRandBGInfo=/Remove \
   -sOutputFile=$argv[2] $argv[1]
end

OK, based on that blog post, I have the following invocation. This doesn't turn the fonts into images but it does downsample all images to 150dpi if they are at 165dpi or larger, and it forces JPEG if there images are uncompressed or PNG (I think).

Let me know if there's anything else you can think of.

3 November at 14:33 | Open on social.alexschroeder.ch
Go Up