preview becomes non-responsive when using quartz filter

dialabrain wrote:

Can you explain how?

I downloaded the yellowed page from this site: https://community.element14.com/members-area/personalblogs/b/blog/posts/cleaning-old-scanned-book-pages-with-linux-bash-commands

Then I saved it as a PDF file. Then I duplicated that PDF file. Then I opened my first PDF copy and added the duplicated page.

This is not a very good test. I would have preferred an original, in-the-wild example, but I didn't want to search too hard. From experience, I know that doing anything like this with multi-page documents is quite difficult, so I didn't bother.

Edit: I tried again and this time had better luck. I found a PDF very close to what the OP has. See https://commons.wikimedia.org/wiki/File:Noli_me_t%C3%A1ngere_%28Jos%C3%A9_Rizal%29.pdf

I had no problem exporting it to a PDF with a black and white quartz filter. And with a real document, my output was much better. It was still poor quality, but the OCR seems to work fine. But then, the OCR also worked fine on the original. I still couldn't do the kind of colour manipulation I would like on a multi-page document. But other quartz filters worked fine. In theory, it's possible to create custom quartz filters, as described here: Add or modify Quartz filters in ColorSync Utility on Mac - Apple Support, but I didn't test that.

But to get back to the original test, I also downloaded the greyscale, heavily skewed image. The built-in OCR was able to read most of that. I'm glad this technology didn't exist a decade ago. I would have spent way too much time tweaking PDFs in grad school.

Can you explain what you mean by a large PDF? Do you mean in size or number of pages? I'm not certain what you were told but, to my knowledge, quartz filters apply to images, not documents.

lhfrank wrote:

It's a 288 page document, 15.7 MB. It's a very yellowed copy of a book of music. I thought the filters would work on documents too!

Quartz filters do work on PDF documents. However, they are designed for images. I don't have any documents with yellowed pages to test. I made one with 2 pages and it worked fine. But it wouldn't surprise me if it failed on a 288 page document. For very technical reasons, it just isn't designed for that.

But it's really a moot point. Even if it had worked, the output would be unusable. Black and white is what it says, black and white. You'll be left with 288 white pages, each filled with a random pattern of connected dots. No text.

If you want to extract the text, I recommend Preview's built-in text extraction tool. It has no problem with yellowed pages. Unfortunately, it does have a problem with documents. It only seems to work on one page at a time. And it doesn't seem to work on PDF documents at all. (Or at least, not on all PDF documents.)

Remember, this is Apple "Preview", not Apple Photoshop. Apple understands that people want more. They just bought Pixelmator. But even Pixelmator is designed for images, not documents. Apple simply doesn't support this kind of data extraction or conversion at scale. There is a reason why people sue Google and the Internet Archive for this, and not Apple.

I'm afraid you're going to have to search the internet "grey market" of command-line PDF tools with OCR capabilities.

etresoft wrote:

Quartz filters do work on PDF documents. However, they are designed for images. I don't have any documents with yellowed pages to test. I made one with 2 pages and it worked fine.

Can you explain how?

Honestly I don't even know how you would apply the filters on an image in Preview. It's easy enough to correct the colors of an image in Preview. Maybe someone else knows.