correcting for uneven background of scanned 35mm negatives

"Alex" wrote in message

Hello all,

I have a few 35 mm negative films that all show the same uneven background –
clearly visible on the unexposed parts of the films. Probably the result of a
not well closed film container before or after use I guess.
My question is: how can I correct this in PS, using a scan of a blank frame and
a scan of an exposed frame? I have experimented a little trying to merge the
two layers containing the ‘blank’ and the image, but not with any success. I
have the feeling this should be relatively easy, if only I knew how 🙂

Hi Alex,

You’ll have to fiddle a bit. Here are two suggestions.

First method: Scan the blank frame, and use it as a mask for a curves adjustment layer on one of the images you want to fix. Click on the mask layer and adjust it with curves, then click on the adjustment layer and curve that layer. Each time you can improve the image just a bit.

Second method: Scan the blank frame, and add it to your target image as a new layer, with the mode set to difference.

Both of these sound good, but the devil is in the details. If you post a link to a blank frame image and an image you want to fix, perhaps some of the people here will take a crack at it.
—

Mike Russell
www.curvemeister.com/forum/

"Mike Russell" wrote:

Hi Alex,

You’ll have to fiddle a bit. Here are two suggestions.

First method: Scan the blank frame, and use it as a mask for a curves adjustment layer on one of the images you want to fix. Click on the mask layer and adjust it with curves, then click on the adjustment layer and curve that layer. Each time you can improve the image just a bit.
Second method: Scan the blank frame, and add it to your target image as a new layer, with the mode set to difference.

Both of these sound good, but the devil is in the details. If you post a link to a blank frame image and an image you want to fix, perhaps some of the people here will take a crack at it.
—

Mike Russell
www.curvemeister.com/forum/

Hello Mike:

I have two images at :
http://s56.photobucket.com/albums/g190/alexalex_02/

These are downsampled scans of 1) a blank portion of the film and 2) a (very underexposed) image. The underexposure clearly show the problem. I have added to Photobuckett a better ‘image02’ to have a real case to work on. I also added a ‘background02’ to go with it.

I already tried the second method you suggest, not really with any success. The result looks pale and has to be changed a lot to look better. I will try your first idea.

Thanks so far!

Greetings, Alex

Hi Alex,

I’m on the road at the moment, and will take a crack at fixing the light leak problem using the examples you provided.

After a quick glance, it seems to me that the light leak problem is relatively subtle. The larger problem is your procedure for inverting the negatives – you are not compensatig for the orange mask of the negative, and the result is an image that is too blue.

I’m about to change cities and may or may not have an internet connection. Perhaps someone else will chime in in the meantime.

BTW – are the tulip images from Kukenhoff Isse?

Mike

"Mike Russell" wrote:

Hi Alex,

I’m on the road at the moment, and will take a crack at fixing the light leak problem using the examples you provided.

After a quick glance, it seems to me that the light leak problem is relatively subtle. The larger problem is your procedure for inverting the negatives – you are not compensatig for the orange mask of the negative, and the result is an image that is too blue.

I’m about to change cities and may or may not have an internet connection. Perhaps someone else will chime in in the meantime.

The tulips are indeed from the Keukenhof park at Lisse! Clever to see that from such a coarse image (the originals are about 2700 x 4000).

I did compensate for the orange mask while scanning: the example is perhaps the worst because of the severe underexposure, and any mismatch in the mask compensation shows more. The tulip image is really not so bad after inverting and level/gamma correction?
The images have been scanned at 16 bits/channel with a linear gamma, so there is still some work to do after inverting. But all that is of later concern. I first want to get rid of the effect of the light leakage. The rest is fairly routine here!

In fact I started to subtract the RGB values of the ‘blank’ pixel by pixel from the RGB values of the image with a simple program (that is not yet finished however —).

Alex
Greetings, Alex

Alex wrote:

<snip>

In fact I started to subtract the RGB values of the ‘blank’ pixel by pixel from
the RGB values of the image with a simple program (that is not yet finished however —).

The program works in principle and it does just about what I expected. The amount of ‘blank’ to subtract has to be twiddled a little I guess, but so far so good.

Greetings, Alex

"Alex" wrote in message

Alex wrote:

<snip>

In fact I started to subtract the RGB values of the ‘blank’ pixel by pixel from
the RGB values of the image with a simple program (that is not yet finished
however —).

The program works in principle and it does just about what I expected. The amount of ‘blank’ to subtract has to be twiddled a little I guess, but so far
so good.

Cool – keep us posted on the results.

—

Mike Russell
www.curvemeister.com/forum/

"Mike Russell" wrote:

<snip>

The program works in principle and it does just about what I expected. The amount of ‘blank’ to subtract has to be twiddled a little I guess, but so far
so good.

Cool – keep us posted on the results.

—

Mike Russell
www.curvemeister.com/forum/

I have a working version that more or less gives the results I expected. The 84 negatives I wanted to process are nearly done.

There is one hitch: the effect seems not to be strictly additive: the amounts of ‘blank’ present in underexposed or well exposed parts are different. I am afraid that the solution is not so simple as I presumed.

For the time being I am satisfied. The negatives are not that terribly important or even of great quality.

Thanks for the support!
Greetings, Alex

great quality.

Thanks for the support!

Greetings, Alex

"Alex" wrote in message
[re subtracting fogged area from negs]

I have a working version that more or less gives the results I expected. The 84
negatives I wanted to process are nearly done.

There is one hitch: the effect seems not to be strictly additive: the amounts
of ‘blank’ present in underexposed or well exposed parts are different. I am
afraid that the solution is not so simple as I presumed.
For the time being I am satisfied. The negatives are not that terribly important or even of great quality.

Hi Alex,

As you say, the effect is not additive for a number of reasons. The toe of the film’s response curve is almost level at the dark end, and there is a technique called "flashing" that takes advantage of this: film is exposed to a small amount of light to "preload" it, and make it more sensitive to subsequent exposure. In color, the issues become more complex because you have a minimum of three layers of emulsion, each with a slightly different characteristic curve.

That aside, I’m glad to hear you had success working outside the box like this. The end result is what matters, and an empirical solution, based on what works, seems to be just what the doctor ordered.
—

Mike Russell
www.curvemeister.com/forum/