Close
【HELP】 with kantoku's pic
sample view is here
raw is here (101.2MB)

the pic is from E2 vol 36, heavy paper texure in dark area, I am not going to try that :-) when you finish just upload for me, thanks.
I believe that was why no one scans out of that magazine anymore.
WtfCakes said:
I believe that was why no one scans out of that magazine anymore.
for bright pics, that's ok, but for dark background....hmmm
can i give you a suggestion: please flatten image before saving. It can efficiently reduce 15-20% file size. Almost all of your uploads have this problem.
fireattack said:
can i give you a suggestion: please flatten image before saving. It can efficiently reduce 15-20% file size. Almost of your uploads have this problem.
yes, thank you.
That mag uses the worst paper to scan. And hime is using the same paper now, I just gave up.
If I need to filter those strong textures, do strong filter only on bright area, and weak gauss on dark area
better wait for those compilation books they release from time to time.
Akibarika had scanned E2( post #183523 ) but it isn't so bad like yours.
the raw shows she use 400DPI so if u use too high dpi for this scan(I can't download from mediafire so i just guess it through the file size)
I don't think it has any difference since you can just resize it..
milumon said:
Akibarika had scanned E2( post #183523 ) but it isn't so bad like yours.
I'd say it's just as bad.
You can remove most of the paper texture with a strong filer by using GreyC. It destroy some detail so you can either mask it or use the history brush at maybe mid opacity to restore some details such as the blush lines.

I'm doing a very lazy method on how I usually remove paper texture. Of course you gotta overfilter to remove them. If you use the quick mask or the magic wand and feather it. The spatial & the gauss precision are the important factor with iteration 2nd. Iteration should be repeating the filter, but reapplying the filter 3x at 1 isn't the same as setting the iteration at 3.

#1 is the original or what scan we had.
#2 is slide #4 parameters
#3 is slide #4 parameters, but simply reapplied. So 2x application of the same settings.

But like i said, this technique will destroy almost anything that a scanner will pick up even a good amount those weird crappily compressed jpeg ring (Jpeg Gradient Banding).

Probably can keep the angular step at the default 30 so you don't alter the brightness. I use 15 because if you leave it at anything above that you get areas that remain soild and not more opaque and will result in areas of unfiltered noise. Sadly i have not found a way to retain original brightness through denoising while remove textures.

ps: Don't forget to set contour to 1. Anything under .7 will not preserve detail.

ex.


this is just so bad that i wonder if i need to upload it
That's becoming watercolor now. Not worth it.