downsampling pdf in Indesign & Illustrator does nothing?

[wentworth]

Is it true that downsampling when exporting/saving a pdf in either Illustrator or Indesign does nothing if the design is vector?
It does make sense but for years ive been assuming it did something. Especially in Illustrator where i would assume a high percentage of the design would be vector it seems silly to have it as an option.

[CHM]

I presume you talk about image down sampling. This only changes your images in your pdf.
Vectors can not be changed in pdf, these stay as they where in the original file. This also means if you have a file with a lot of paths (like vectorised text) you file wil remain big.

[wentworth]

yeh i mean under the compression part of the export window. I see now that illustrator mentions bitmap, so obviously not vectors. But Im more used to creating pdf's in indesign and thats not so specific about what the downsampling refers to. So basically If its all vector you may aswell have all set to none right?
Just seems weird that in a program like Illustrator the export to pdf compression refers only to bitmap when its used largely to create vectors.

[Stephen Marsh]

Adobe offer "unified" PDF export options, so that all programs can use the same settings for consistency.

One can save using the "Illustrator Default" setting, which is probably best for PDF files that will be opened in Illustrator and edited (as opposed to PDF files destined for direct output such as PDF/X etc).

I also make Photoshop PDF presets that do not resample and that do not use JPEG, so that image only PDF files can be opened/resaved multiple times without damage.


Best,

Stephen Marsh

[dabob]

yeh i mean under the compression part of the export window. I see now that illustrator mentions bitmap, so obviously not vectors. But Im more used to creating pdf's in indesign and thats not so specific about what the downsampling refers to. So basically If its all vector you may aswell have all set to none right?
Just seems weird that in a program like Illustrator the export to pdf compression refers only to bitmap when its used largely to create vectors.

But some people use illustrator as a total design tool . . . a couple of years ago I got a 16 page book in quark only to find out it was one illustrator file with 16 pages laid out on one really big sheet and then placed and cropped in quark for the 16 page version . . . some people work way to hard!!!!!

[michaelejahn]

Just seems weird that in a program like Illustrator the export to pdf compression refers only to bitmap when its used largely to create vectors.

Perhaps you have not ever created artwork in Adobe Illustrator (example : a package design) where you might have images ( example : go to the store and notice that there a lot of packages that contain images)

In these cases, we are all very happy that Adobe thought to include the ability to not downsample the images that are placed into an Illustrator document.

You might not know this, but many many developers build all sorts of advanced plug-ins for Adobe Illustrator that are very specifically designed to assist in the package design prepress vertical.

There are many things that one can do in that application to images and objects that are not that simple in other applications (like InDesign, which has more tools for multi-page documents)

In Flexographic printing in particular, you often need to apply amorphic distortions.

you might find this interesting,

Flexographic Pre-Press: Calculating Distortion Factors Without a Chart!

hope this helped !

[David Kunkel]

There's an obscure setting in Illustrator to increase the 'flatness' of curved lines. It allows the imagesetter to use a coarser setting when deciding which imagesetter pixels are inside a boundary vs. outside. It can make the difference between imaging smoothly or choking the processor (of course, this was much more likely a decade ago). I wonder if this would affect the size of a PDF printed (not saved) from the Illustrator file.

[Stephen Marsh]

AFAIK this is just a metadata entry, however it used to be applied on an object level in older versions - so I guess that all the metadata entries could add up in older versions (however they are still just text entries). I doubt that this greatly adds to a vector documents file size. I think that the current defaults are application wide and not at the unique object level. Most leave this blank to use the RIP defaults.

I remember a nasty bug back in the day (was it AI9), where curved vectors like circles were coming out like octagons (well not that bad, but it was visible to the naked eye that the curves were not smooth).

Stephen Marsh

[Lukas Engqvist]

What you mention Stephen as the AI9 is the flatness that Michael was talking about. It was not limited to any one software but a general problem in RIPs as one tried to balance computer power with the complexity of designs.

There are many effects that are rasterised from illustrator, besides placed art…*all of wich can be downsapled, some times rendering ugly results. If an object has a feather or a dropshadow it will be rasterised so one should be very careful to interpret illustrator and PDF as size or resolution independant.

There is no all vector or all bitmap world any more

[David Kunkel]

I tend to agree with Lukas that a large file size suggests that the original poster's files aren't all vector but have Illustrator-created raster effects pushing up the file size. Compression won't affect vector art, but since it would take thousands of vector-defined objects to equal the file-bloating capability of one placed (or generated) bitmap, downsampling to the appropriate resolution for the final usage could generate significant size reductions. I'd keep the appropriate compression settings on, just in case.

Back to the digression: I remember reading in the early 90's that one PostScript pattern-filled object laying over another could stop an imagesetter in its tracks, while a big bitmap would simply get slowly but relentlessly digested and imaged. In the former case, the processor would have to do all the math for the lower object, then all the math for the upper object, and then subtract the latter from the former for every point along the edge (where the flatness setting would become a critical success factor).