Yet another file size question

[bracewell4213]

I hesitate to post, as there are so many discussions, but I am stumped.

A client file had way too many pasted vector shapes and I thought that's what was bogging down the file and making the exported PDFs large.

I thought I would replace with .ai and also Simplify the path. Nope. The vector art, pasted into illustrator, and Simplified to 2.2 mb and linking in ID didn't help. A single page PDF with minimal text and the graphic is over 9MB.

I tried making a 1MB PSD and that didn't help much. 

I made sure Output included no Profiles. Using default Press Quality as a benchmark.

I eventually deleted all the other pages in the file, nope. I saved as IDML, which really helped reduce the ID file size, but had no impact on the single back blank PDF.

I'm attaching the IDML, the .ai, and a series of PDF tests. 

What is the content on the "blank 5.pdf" that makes the file 9.2 MB?

( tried uploading .ai and it says content type doesn't match extension, which makes no sense. Adobe does not recognize it's own file type? I will upload as PDF, but it was placed as .ai)

[Jeffrey_Smith]

blank 5.pdf is just text, and I cannot find anything that is causing the bloat. Copy and paste text into new document creates a smaller file, but there is still a lot of bloat.

[rob day · Accepted solution]

Hi @bracewell4213 , This looks like it might be the ancestor metadata bug discussed here:

https://community.adobe.com/t5/indesign-discussions/file-size-is-too-big/td-p/9370587#10084949

If I check the ID file’s metadata info I get this for Raw Data:

Can you share the placed YellowRedFooter.psd file?

[bracewell4213]

Hi Rob,

I'm not sure who marked this topic Solved, wouldn't that be up to the posting person? I appreciate the help, but don't consider this solved quite yet...

I read through that thread and it seems all about .PSD metadata and having a document with a huge amount of images in it. I am not seeing exactly how I replace the metadata in the InDesign file. 
   From the File Info > Raw Data screen, my choices are to Import, Export, or Template Folder. There are no XMP templates in my Adobe Application support folders. I checked my older Mac and these same folders are empty there also.
There are mentions of Bridge in a search for "default xmp template for indesign", but I don't use Bridge or have it installed, nor do I want to , as Adobe CC is already taxing my computer's memory and SSD.
   So to solve this, I need to find an alternative XMP template to replace the excessive data put there by Canto, correct? Thanks for your insight, this is far down the rabbit hole. The bigger issue is replacing all the insane vectors with .ai, thereby improving InDesign's sluggish response in this file, but before doing so, the XML default would be the place to start.

[rob day]

A single page PDF with minimal text and the graphic is over 9MB.

Brad’s answer is correct for the metadata part of the file size problem. There is also a very large vector object on a masterpage that is adding to the file size.

The vector object on your A-Master has 24,100 anchor points— saving the vector art as an image and linking to the image will reduce the file size. Here is the result of a script that counts a selection’s anchor points:

On the metadata, If I Export the ID document’s Raw Data—File>File Info>Export—you can see the XMP text alone is 9MB. The bulk of the file size is CantoMetadata. Do you use Canto Digital Asset Management, or did you inherit this file?:

If you replace the vector art with the attached .psd, import the attached default.xmp and do a Save As, the file size will drop to 1MB.

[Brad @ Roaring Mouse · Accepted solution]

Your files all contain a massive Canto Metadata section. That's what's driving your size up, even on a blank page. I don't use Canto so not sure how to advise to avoid that. If you delete the Canto metadata item in your PDF properties, and resave, your files drop down to almost nothing.

[rob day]

Also, I just noticed on your parent page named A-Master, there is still a pasted vector path at the bottom of the page with 1000s of anchor points. You should consider converting the vectors into a linked image file:

If I replace the ID file’s Metadata with the default InDesign metadata template and remove the complex vector object, the file size is 1MB. See attached

[bracewell4213]

Hi Rob,

As I mentioned in my post, that's what started this, I made a .ai to replace the footer image. Before I edited the master for the whole book (which is many pages), I deleted the footer on my test page and placed the .ai so I could run a before/after test. Since my tests PDFs never got smaller, as I describe in my post, I have not yet gotten to the point of placing the .ai on a master page. There are many other masters in the original and all will need new .ai to replace these vectors, but first I need to solve the metadata in InDesign issue.

[Brad @ Roaring Mouse]

As complicated as the vector object seems to be, if you copy and paste it into a new document, and save a PDF, it's less than 1M, so it's not as hoggish as one would think.

The Canto metadata is travelling with the OP's ID file. It's in the .idml file that we got, and even if I save as a full ID document on my system, a resulting PDF still carries it. ONLY if I copy and paste the objects into a brand new document does it vanish.

So, this might be a question for the Canto people. Why is it so large and is it supposed to populate into PDFs, etc, etc.

[rob day]

Hi Brad, you can import and export xmp data. I attached a default new doc xmp above, which you can import into the problem doc via File Info to remove the Canto metadata.

I was looking at the InDesign file size with the embedded vector art and it grows by around 5mb for every copy and every copy degrades performance  considerably. It doesn't seem like vectors are the best choice for this kind of art.

[bracewell4213]

Yes, this is actually the initial problem I was trying to solve. I know all these vectors are bogging down the 48 page file, and this is but one of many (the sister file is 148 pages, some with 6 or 8 "picture frames" done in that same vector style).

   I converted one footer to simplified .ai and did a test to be sure it looked ok and would result in a smaller file, but it did not. That's what led me to step-by-step remove things until I had a one page doc and continued to test each step of removal.
   I had never heard of Canto, this is an inherited file. I'd never really explored XMP at all.

  Thank you for the default.xmp!!!!! I hunted online and couldn't find one anywhere, despite many mentions of it. I'm a bit puzzled why I don't have one. Isn't that supposed to be part of the default INDD install? I put it in the Metadata Templates folder so I can apply it to the other books in this project.

>>> In my 48 page original file, swapping the default.xmp in the File Info>Raw Data reduced my .indd from 105mb to 56mb right away.
Exporting page 4 with the dumb vectors intact, the PDF is now 465kb.

I think the ID file is still sluggish, assuming it's drawing on a lot of RAM to paint all those vectors, so I will probably still pursue replacing with .ai, so at least I'm only operating with a preview. 

   Files from this client routinely come with highly questionable production choices. I suspect some young whippersnapper designer hasn't learned about efficient production work. I will be fixing lots of dumb styles and other things as well, but it will pay off since there are 12 books to the series.

   I agree that Canto is poorly executed if it's bloating an exported PDF to the point where one page is 30mb (from another of my test files). I'll have to see if that xmp data is also in the photos placed in the other books. They are insanely slow to work in as well.
   Thanks for your patience and sticking with me to a working solution!

[rob day]

You can make your own default .xmp by creating a new blank document and Exporting its .xmp

If you feel compelled to use complex vectors rather than a high res image file, move it into Illustrator, save an .ai file, and link rather than embed/paste. If you don't link, InDesign has to process the 24,000 path points, and you'll always have performance issues.

With this kind of textured art, a linked image is going to perform better, and there shouldn't be a quality issue as long as the Effective Resolution is adequate for the output.

[bracewell4213]

Yes, I agree Rob, that's what I wrote in my post that started all this.

"I thought I would replace with .ai and also Simplify the path." 

You've pointed that out a few times now and I've replied a few times that I did put a new .ai in and the resulting PDF was no smaller. I'm very aware of the vectors, believe me, it was the first thing I saw when I inherited this file. That is my plan, and there are additional things at play, like on some pages these are used as photo frames, so I have some choices to make as to achieving the window effect. This is the start of my experimentation, but it was moot when we discovered the problems was the XMP. I appreciate your help with the default.xmp. 

[rob day]

that I did put a new .ai in and the resulting PDF was no smaller

I meant the file size and performance of the InDesign file would improve because linked files preview with a proxy. I wouldn't expect a change in the PDF because you are still exporting the very complex paths via the link.

I would expect an improvement in the PDF file size if you replaced the .ai file with a .psd version of the art. The attached PDF contains the vector art converted to a PSD and is only 660KB. The arts resolution is 400ppi, more than enough to fully resolve the texture.

[Brad @ Roaring Mouse]

Apart from the metadata bloat, my suggestion is to render your footers into a CMYK TIF file at 600ppi, no antialiasing, (optionally with LZW compression) and place that instead of the vector. I did a PDF test with a 24-page document full of text and the PDF file size was a total of 730KB.

Even if you stay with vector, you can simplify it a lot by combining commnon shapes. e.g. select all the white-filled objects and use Pathfinder to Unite them into one shape... that will reduce the number of points by 2/3 alone.