The color between the preview and the original image looks so different to me. What cause this to happen?

b923242 said:
The color between the preview and the original image looks so different to me. What cause this to happen?

It's in cmyk. Also it's not the source image, it's (re)exported from pdf source to jpg with maximum quality settings.

Or at least that's what I think what's happened here.

Oh thanks for that, I didn't notice this one is from the VFB pdf, that some poor export execution than.

Hello,

Thank you for your finding. We are still trying to figure out a proper way to export these but I guess I shouldn't have uploaded the pdf >_> and let the amateur have it their way. Ah well, once I have the proper export, I'll clean them all up.

Checkmate said:
Hello,

Thank you for your finding. We are still trying to figure out a proper way to export these but I guess I shouldn't have uploaded the pdf >_> and let the amateur have it their way. Ah well, once I have the proper export, I'll clean them all up.

I do think release the raw file is a good thing. If you say so, the other sonora's arts should had the same problem as well? Tho I only see the issue on this one.

For pool#5993, I use Acrobat, export all images to .tiff files, converted to rgb space by using PS, then rotate if necessary, save to png.

Thank you for sharing that method, I'll explore additional options.

Let's keep calm and don't start uploading those digitals like mad yet.

But that means my rips of the previous Cube game is borked. Pool #5370

Images are embedded in PDF file in two ways: lossy as DCT, basically JPEG, or losslessly as "flate", which is similar to zip.

The majority of the color images will be in DCT format.
You can use `pdfimages -raw` to dump them losslessly (they're in CMYK).

But for these ebooks, the color is somehow inverted if you dump the raw using PDFImages. Not sure if it's a bug of that tool, since PS doesn't have this problem (but PS can't dump them, just open).

Anyway, I have to process them at least once to get the correct color. But since I will convert them to RGB anyway, it doesn't bother me. Just make sure to save as PNG/BMP later to prevent introducing new artifact (but of course, the high compression ratio is lost in process.)

Some other notes:

1) Acrobat will export *pages* instead of images. This may or may not be what you want. You can use `pdfimages` as mentioned above, or Photoshop to extract image resources directly. I prefer the latter since it may have better size/less text etc. See post #521695

2) When converting to RGB, make sure use "Japan Color 2001 Coated" profile (set it first in Acrobat and/or Photoshop.) I've compared with numerous official RGB source to confirm that this is what the industry use.

If you choose to use Acrobat like Twins, you can directly export pages as RGB in process (there is an option). So you don't need to convert it in PS.

Also, Acrobat claims it can extract DCT image without re-compression if and only if the whole page is just one image. Which is NOT the case here (these ebooks are either have multiple images as resource on each, or have *two* images on each page). So this function won't work here, if you export to JPEG, they will be recompressed.

Thank you for having posted your findings fireattack. I did try the -raw one from pdfimages. The problem is that there's vector text on the page which means a reconstruction is required, or the artbook will be an image rip rather than the 1:1 of the said artbook in the format that resembles the original pdf file.

I think we have to compromise here, eventually when digital file is getting popular

Yep, like mentioned depends on what you want.