I did some searching on the Wayback Machine and was able to recover the dead link: Auto rotate with a vector. After downloading and trying Version 3 of the script, I see what you mean about this being a very handy utility. It sure beats rotating the image manually to get rid of the skew.
Thanks for tracking that down, Iâm using it already!
Or you could simply use the measure and rotate tools.
Rotate works pretty well with the âcorrectiveâ version, where you can align the grid to the image. But I like being able to pick two specific points that I want to align. Sometimes the grid doesnât match up with what Iâm trying to guide by.
I just bought a HP DeskJet 2630 all-in-one but honestly itâs scanner is a disaster. Even at 1200 dpi, the image is made of big dots.
This is the image that I scanned using it
The image looks like I scanned it, then I printed it, and then I scanned it again. Horrible.
And this is the image that I got from scanning at the internet cafe at the corner of the street - they use a Konica Minolta* C220 / Develop ineo+ 220
I remember some 15 years ago I had a regular scanner (probably BENQ) that I paid some 50 or 60 Euros for it and it was quite ok. But this HP scanner is horrible.
Can anyone recommend me a decent scanner?
Thanks
Strange as it may seem, the top image from your scanner is probably the more accurate. The dots youâre seeing are called a moire pattern (https://en.wikipedia.org/wiki/MoirĂ©_pattern#Printing_full-color_images) and are pretty normal when scanning an album cover. What I do is scan at 1200 dpi, then use a filter called descreening in GIMP to smooth out the color before scaling it down.
The scanner at the cafe may have some software doing a similar process automatically, but youâre going to have a more accurate end result if you do it manually (you can see a color difference between the two images, too, which means that scanner is probably doing more than just descreening).
What we are seeing here is not a 1200 dpi scan. A CD cover at 1200 dpi would be much larger than 1000x1000px.
I wouldnât call the top image more accurate at all. This moirĂ© is not a pattern that is on the actual cover. Itâs an additional artefact.
My guess would be that the posted image is a screenshot from an image viewer that introduced the moire at a viewing level.
Ah, no. I downloaded the image, and itâs actually looking much worse than that.
There is something else very wrong here.
Thanks for all the replies. The first image is scanned using my HP DJ 2630, at 1200 dpi, and resized to 1000 px height.
At 300 dpi, the image looks much worse, of course.
I am trying to upload the 300 dpi version now, it has 1.1 MB and 1,400 x 1,400 px , but I get the following error:
Sorry, that file is too big (maximum size is 3072kb). Why not upload your large file to a cloud sharing service, then share the link?
Everything goes wrong with me at moment, not my lucky days it seems. Iâll take a break and come back later. Iâll play some StarCraft (Remastered!), maybe later Iâll face less disturbances in the field
The moire you are getting here is really intense.The type of scanner you are using is often pre-set to produce an overly sharpened image. In addititon to this you may have used the wrong algorithm for scaling it down.
When I scale down my own scans using photoshopâs âbicubic (smooth gradients)â algorithm, the moire usually is on the subtle side.
Descreening before scaling down will certainly help, but make sure to pick a suitable algorithm for downscaling as well. It will keep things a lot smoother.