Help repairing photo – scanner fault

wrote in message

Hi
I spent 10 days scanning borrowed family photos at high res. Not until I’d finished did I discover that there is a fault with the scanner, and I can’t easily borrow the photos again.

There are stripes of different ‘greyness’, approx 2 cm wide accross the mostly b&w photos, rather like the different stripes of green in a newly mowed lawn. If you want to see an example, here’s one; I’ve reduced the resolution for the internet.
http://homepage.ntlworld.com/tony.jeffs/junk.jpg

What exactly is the difference between the blue-gray and the black-grey areas?

Any ideas on how about trying to correct it? Is it a realistic possibility?

Thanks tony

Reply

That link really is a bit too small to experiment on properly. But a couple of thoughts – If all the images have the same strips – same width etc, if you make a brightness/contrast adjustment layer, and mask the stripes carefully, you may be able to use the same adjustment layer on all your images, assuming they have the exact same defect. See here: http://www.tv83.net/photorestoration/remove_strips.jpg for a very rough demo.

Hi Caitlin
You’ve got a good point. Once I’ve solved one, I will have solved them all!!!
When I’ve tried, I’ve had a problem getting the edge lines accurate, same as your rough demo, but maybe I could use the healing tool to fix that.

The biggest problem is that I do’t know for sure that it’s a contrast/brightness problem. There’s a purple cast too on the lighter stripes, and I cant quite figure out what the exact difference is between the two areas. It might be something I can fix with curves.

Maybe tomorrow Ill look more carefully at the spectrums for both. It’s late here, so Ill have to stop for today!

Thanks

Tony

wrote in message

Hi Caitlin
You’ve got a good point. Once I’ve solved one, I will have solved them all!!!
When I’ve tried, I’ve had a problem getting the edge lines accurate, same as your rough demo, but maybe I could use the healing tool to fix that.

The biggest problem is that I do’t know for sure that it’s a contrast/brightness problem. There’s a purple cast too on the lighter stripes, and I cant quite figure out what the exact difference is between the two areas. It might be something I can fix with curves.
Maybe tomorrow Ill look more carefully at the spectrums for both. It’s late here, so Ill have to stop for today!

Thanks

Tony

With the stripes for the mask, one thing I also tried after I’d already posted the jpg preview is select each section on the mask and used transform to tweak the edges to align them more accurately. Seemed to work pretty well. I would think this would also be easier on the high res version. The brightness/contrast seemed to fix the colour discrepancy, but again on such a small version it’s hard to tell.

On 8 Sep 2005 15:31:46 -0700, wrote:

Hi Caitlin
You’ve got a good point. Once I’ve solved one, I will have solved them all!!!
When I’ve tried, I’ve had a problem getting the edge lines accurate, same as your rough demo, but maybe I could use the healing tool to fix that.

The biggest problem is that I do’t know for sure that it’s a contrast/brightness problem. There’s a purple cast too on the lighter stripes, and I cant quite figure out what the exact difference is between the two areas. It might be something I can fix with curves.
Maybe tomorrow Ill look more carefully at the spectrums for both. It’s late here, so Ill have to stop for today!

Thanks

Tony

If the lines are coloured you may find the easiest solution is to look in the channels palette. You’ll likely find at least one channel where the lines don’t show.

—

Hecate – The Real One

Fashion: Buying things you don’t need, with money
you don’t have, to impress people you don’t like…

"Hecate" wrote in message

On 8 Sep 2005 15:31:46 -0700, wrote:

Hi Caitlin
You’ve got a good point. Once I’ve solved one, I will have solved them all!!!
When I’ve tried, I’ve had a problem getting the edge lines accurate, same as your rough demo, but maybe I could use the healing tool to fix that.

The biggest problem is that I do’t know for sure that it’s a contrast/brightness problem. There’s a purple cast too on the lighter stripes, and I cant quite figure out what the exact difference is between the two areas. It might be something I can fix with curves.
Maybe tomorrow Ill look more carefully at the spectrums for both. It’s late here, so Ill have to stop for today!

Thanks

Tony

If the lines are coloured you may find the easiest solution is to look in the channels palette. You’ll likely find at least one channel where the lines don’t show.

The lines aren’t coloured (Did you look at the attachment?) They appear in all channels.

Try scanning a neutral gray card, as textureless as you can get, at the same resolution. You now have the stripes without the image. Invert it and put it on a layer above the photo. A blend mode of Soft Light or Hard Light should start to reduce the banding. You’ll probably have to move the stripes layer up or down to match with the stripes on your photo.

Just a suggestion. Let us know how it turns out.
Tom Nelson
Tom Nelson Photography

wrote in message

There are stripes of different ‘greyness’, approx 2 cm wide accross the mostly b&w photos, rather like the different stripes of green in a newly mowed lawn. If you want to see an example, here’s one; I’ve reduced the resolution for the internet.
http://homepage.ntlworld.com/tony.jeffs/junk.jpg

On Fri, 9 Sep 2005 16:52:51 +1000, "Caitlin"
wrote:

The lines aren’t coloured (Did you look at the attachment?) They appear in all channels.

Nope, just read the post. <shrug> I don’t see attachments in my text groups.

—

Hecate – The Real One

Fashion: Buying things you don’t need, with money
you don’t have, to impress people you don’t like…

Thanks for all the suggestions. Tom I couldn’t do exactly as you said since I returned the faulty scanner to the shop, but I could find areas of neutral colours in the photos I’d taken.
The trouble is, the fault isn’t consistant. Although it is to do with the heads, smooth photos are less a(e)ffected than stippled photos, and lighter photos less than high contrast photos. It isn’t even constant in different regions of the same photo.

So – Conclusion, after trying everything you & I could think of I’ve given in and finally borrowed the photos again. It was easier than I thought, although some of the better ones in frames will require further negotiation!

Thanks again.

Tony