Untitled Document

Hi all,

Does anybody have a good trick method of reversing a color negative into a decent color positive. I know than normally this would be accomplished in the scanner. In this case I have a friend who wants to shoot the negative with a slide duplicator plus digital camera so there is no built-in clut for reversing the color negative.

I tried a couple of LAB curves but the problem is a little more complex – it seems like there should be some kind of channel operation or Overlay blending trick to help remove the severe orange cast, put the contrast back into the image and amp up the "true" colors in the original scene. He's got several thousand images like this so he's looking for some semi-automatic way of dealing with this.

regards,

Lee Varis

http://www.varis.com

888-964-0024

___________________________________________________________________________

Exactly reversing the "a" and "b" curves in Lab mode would reverse the colours. Then to get rid of the cast from the orange tint in the film, revert to RGB, create a Curves adjustment layer and use the white, black or grey eyedropper on a correspondingly appropriate part of the image. There will usually be some fine-tuning of colour balance to do thereafter.

Mark Segal

___________________________________________________________________________

The problem, of course, is that the orange mask complicates the job of reversal. Here's a procedure that will get you to a good starting point.

To copy and invert color negatives using a digicam:

1) Use daylight or a relatively cool light to illuminate the negs, not tungsten, or your blue channel will be dark and noisy. Copy all your negs with the same exposure if possible, using any fixed white balance setting - indoor is good. Don't rely on the camera's autoexposure or auto white balance.

2) Copy one negative with known pure white and clear orange mask - the film leader has this - now you know why the processor has been returning this piece to you all these years :-).

3) Create a curves layer, and click the white eyedropper on an area of pure orange mask, and the black eyedropper on a maximum black section of negative. Invert the ends of the RGB curve.

4) Save the curve and use it to invert your other images to a positive. You will probably want to make slight adjustments to the curves for each image.

Mike Russell

www.curvemeister.com/forum/

___________________________________________________________________________

Lee,

In the book Photoshop CS (there is a newer CS2 edition) by Blatner & Fraser, Chapter 7 page 341 there is a full explanation of how to do this. In your friend's case though, he has to use the Invert comand to get the neg to positive since he is not using a scanner. He can batch his images by droping them into a dropplet created for this purpose.

All the best!

Paco Márquez

661 McKinley St.

MIramar

San Juan, PR 00907

787-721-8554

787-587-7384 Cel.

http://www.pacomarquez.com

___________________________________________________________________________

On Jul 14, 2006, at 2:43 PM, Mike Russell wrote:

Erh... yeah, I tried this and similar things, like Marks suggestion:

Have either of you actually tried to do this with a C41 negative? The problem is non-trivial!

I cannot upload the images because A. it is not my job and B. they involve heavily protected, "brand" iconic imagery from a company that will not hesitate to sue the living daylights out of me.

I've gotten closest by using curves to neutralize the clear orange base to white and the darkest brown part of the negative to black and then inverting. At this point I take it into LAB and amp the saturation with a curves layer, trying to hit something thats neutral then use blending options to ramp off the saturation boost in the shadows. The problem is the the color balance has some strange cross-overs and certain memory colors are almost completely missing – blue skies tend to go gray! It is possible that some color fading has occurred because some of the imagery seems dated. The few images I've tried have radically different curves so it doesn't look like there

will be a generic conversion possible.

Anyone with real experience doing this have any suggestions?

regards,

Lee Varis

http://www.varis.com

888-964-0024

___________________________________________________________________________

I've done it, just not for a while. If I recall correctly, I sampled the rebate and made a new solid color layer over the orig image. I filled that layer with the rebate color and inverted it. Then I set the layer to Color mode at 50%. That cancelled out the rebate color and the rest was normal curves on an inverted base image.

C-41 negs are not exact inverts of positive images. That's one of the reason it's so hard. You have to directly cancel the rebate color before you do anything else.

Did you try John Brownlow's method?

Martin Evening also has a method in one of his Photoshop For Photographers books.

_andrew webb

---------------------------

WebbWorks Words & Pictures

http://www.webbwork.com

___________________________________________________________________________

Hi Lee,

I missed your original post so don't know how quickly you have to do this or how many you have to do.

I did this once many years ago, the results weren't exactly great and it required a couple of steps.

I copied the negs by laying them on a light table so I could include the edges of the neg.

Bringing them up in photoshop was done with an action performing the following steps:

1. Invert (Ctrl I )

2. Levels with clipping set to 0 and with a pause so I could use the eye droppers.

3. Shadow eyedropper set to 0 and click on the edge of the neg with the orange mask.

4. Highlight eyedropper set to 244 and click on the best highlight. Might have to click on a few spots to get it close

5. Midtone eyedropper set to 153 and click on some sort of neutral grey. Need a lot clicks usually to get this right.

6. Crop, save and on to the next one.

It is possible to crop the negs so there is no border and have auto levels set to find light and dark colours with 0 clipping and then do the colour balance with the midtone eyedropper which could save you a step. That wasn't available in photoshop when I had to do this.

The reason my results were not great is that it was impossible to hold the 35mm negs flat, so sharpness was a problem, and the camera I was using, an nc2000, had no way of setting an accurate white balance for the light source. If you're using a newer Nikon or Canon you should be able to set the colour temperature the same as the light source so this would help a bit. There is no way to have pre sets correcting colour as every roll of film has been processed differently and has a different set of characteristics.

Good luck,

John

John Denniston

www.dennistonphoto.com

www.dirtbikephoto.com

___________________________________________________________________________

Lee, I would have made similar suggestions as have been posted so far. The last time I had to do this was a few years ago with a Crosfield drum scan of a Kodak 400VC neg. I doubt the method I used then would help over what has been said so far, as it was similar (using an inverted sample of the film base and curves etc).

My only other thought would be scanner software, such as Silverfast or ViewScan, which may work as or have a stand alone component that processes images that have been acquired at an earlier point (not sure if a different scanner would be a problem with missing required metadata etc). Perhaps there is stand alone software for working with negs that is not part of a film scanner or other scanner hardware package.

Perhaps some sort of channel blending may also be reqired. As this involves numerous images and unique image and lighting conditions I would be looking for some automated method before writing Photoshop automations, but knowing that whatever method is used there will still be need to work on the images later in Photoshop using non automated methods.

Best,

Stephen Marsh.

___________________________________________________________________________

From: "Lee Varis"

.....

Absolutely. Realizing that my advice to you may well be gratuitous, I have done reversals on a number of negatives, both for my own experimentation and for others.

As is often the case, lack of an actual image will render discussion theoreical, and perhaps futile.

Perhaps the sky was indeed gray. An actual example would lend a needed air of authenticity to any explanation I might offer.

Certainly this is possible, but I think the culprit is likely to be a weak blue channel. The best light source for copying a negative is flash, filtered daylight, or another high temperature source. Tungsten light is guaranteed to result in a weak blue channel in the negative.

You're very welcome, indeed.

Mike Russell

www.curvemeister.com/forum/

___________________________________________________________________________

It's been a while... but I got the most "scanner like" results with ease by simply inverting the image then using Auto Levels as described in this excellent article by Bruce Fraser on creativepro.com. This might be similar to what Paco referenced in the Real World Photoshop CS book.

http://www.creativepro.com/story/feature/17164.html

Bob Smith

Accurate Image • Bob Smith Photographer • Waco Texas USA

http://www.accurateimage.org

___________________________________________________________________________

Lee, just for the record, I don't recommend anything to anyone before I've tested it myself. Furthermore, the technical principles underlying the procedure I suggested to you are unambiguous. Mike Russel's approach is based on similar principles, but do it in a different and perhaps more sure-footed manner. The facts that certain aspects of a procedure are "not your job" or you can be sued for this or thatr are regrettably not technical matters with which members of this forum can assist you despite the best will in the world.

Mark Segal

___________________________________________________________________________

Personally, I've converted 35mm color negs using a scanner (albeit years ago) and was successful in doing so.

Why insist on doing it the hard way? If your friend has as many conversion to do, then buying a scanner is the way to go. Either way it will be a slow and involved process and the cost will definitely reflect on his bottom line not to mention his nerves.

If LEE VARIS! is having a hard time with this, God help your friend!

Another solution would be to get the best quality prints possible from all these negs and then copy those. Of course it will be a trade of in quality vs. time and effort.

This looks like a search for a magically simple solution... which might not exist.

All the best!

Paco Márquez

661 McKinley St.

MIramar

San Juan, PR 00907

787-721-8554

787-587-7384 Cel.

http://www.pacomarquez.com

___________________________________________________________________________

Lee,

Further to my earlier response - which was intended to deal with the question about actually trying what one suggests - of course the issue itself is interesting and much has been written about it. I went through much of this stuff about a year and a half ago when I was trying to see whether I could successfully invert negatives without buying an expensive piece of software to do it for me, because profiling a film scanner for colour negative media is not possible (and I know you are not dealing with a film scanner, otherwise your problem would be solved). After trying everything else I researched, I ended-up buying Silverfast because the volume of work I needed to do simply justified the time-saving of using a high quality film scanner with Silverfast (this now for legacy material, because all my current work is digital, thank goodness).

Anyhow, your question tweaked my curiosity again, so I did a Google search and came up with this recipe below, which actually provides some programming. FRANK ADMISSION: I HAVE NOT TRIED THIS, but the guy who wrote it obviously has, and on the visual evidence - it appears to work.

http: //www.aim-dtp.net/aim/techniques/negative_film/index.htm

I'd be curious to know whether this is helpful to you.

Cheers,

Mark Segal

___________________________________________________________________________

My real concern about scanning negatives were never the color, because eventually we will achieve a good result, in long or short term. I see a problem, thou, in noise. On the example you show below, the color is ok, really good, actually, but the noise is very noticeable, specially on the little lake and the walls of the castle. Sometimes I could get rid of it by "scanning"(I digitize with a DSLR) bigger than needed and downsampling, but it's not a real solution. I'd like to know why this does not happen when they are scanned with a cylinder.

Regards,

Rodolpho Pajuaba

___________________________________________________________________________

Lee,

Please take a look at the link:

http://www.fujifilm.com/image_intelligence/about.html.

It's vague but they protect their recipe the same as Coke Cola does.

Other options may be Kodak HR500, Durst Sigma, & Noritsu.

These systems specialize in scanning negatives using software AND hardware to optimize the scan. I think it would be a shame to tackle this project manually and then learn there is already a faster, cheaper, better solution.

I hope this helps

-Stephen Ray

___________________________________________________________________________

On 7/14/06 11:45 PM, "Stephen Marsh" wrote:

If you have to do a anymore than a few, SilverFast is the way to go. They have a NegaFix mode made exactly for processing color negs scanned as color negs (so you don1t have to use it as the driver but rather feed the rendered scan to this module). Works pretty darn well without all the screwing around with curves and such. Can be used as a plug-in or standalone.

Andrew Rodney

http://www.digitaldog.net/

___________________________________________________________________________

It sounds like it may not be amenable to an automated solution because of faded emulsion, which would make every neg a unique case.

The general solution to negatives is two sets of corrections, one for the dark half, one for the light half, merged with a blurred luminosity mask.

Yes, it's not trivial, but look at it this way. The scanner manufacturers had this problem licked 20 years ago, and they had to do their calculations on the fly with 128k of RAM. So it's not like it's the world's most sophisticated algorithm.

Dan Margulis

___________________________________________________________________________

Rodolpho,

The hardware (light sensors) used to capture the light are better and in some cases different; PMT tubes vs CCD devices. This is not to say you won't get noise at all just that the threshold before you see it is much higher.

Lee Clawson

2/\V/\7 Studio

___________________________________________________________________________

I'm curious. Have you been able to identify the film type?

We scan negatives all the time. At the low end a digital minilab does a remarkably good job considering the ease and speed, but at the high end it isn't easy and never routine. We drum scan using fluid mount and we scan in raw RGB as if it were a chrome. In Photoshop we have an action which, among other things, reverses the image and creates a layer set containing several levels and curve layers to 1) neutralize the orange mask to black, 2) set the image exposure to near black for the shadows, 3) set the highlights, 4) change the curve to an 'S' shape and 5) adjust any individual hue errors. Sharpening and noise are handled with the use of Kodak's digital GEM plugin.

Noise from a scan is always more evident than if the negative were printed optically which is something I suggest you have done from this one. Every neg is different and if you're dealing with an internegative, anything is possible. It sounds as if no one is really sure what is possible from the neg in the first place. Get thee to a photo lab (if you can find one!) and have an enlarger print made as a guide.

john castronovo

tech photo & imaging

___________________________________________________________________________

Hi all,

Thank you, everyone, for your suggestions. I guess I was hoping for some magic bullet technique but, based on everyone's response and the various links that were provided it really just boils down to some basic moves that were my first thoughts anyway.

MARK SEGAL wrote:

Timo cracks me up... ever the champion of linear gamma his technique really turns out to be the same basic white point/black point inversion technique that everybody uses only he doesn't know how to use curves so he writes curves functions using Excel to generate different gamma functions (applied as "curves" in the Curves dialog). Very funny...

The other suggestions basically involve auto levels which is fine as far as it goes:

This article by Bruce Fraser is really about using the auto correction options to set the parameters for auto color correction - decent info that dates back to Photoshop 7 but there is nothing here that specifically addresses reversing color negatives.

Mike Russell writes:

Actually, in the two sample files that I have seen, this may be a big factor. The normal inversion techniques would possibly work better if there was more info in the blue channel to begin with. This is a job that my friend is doing at his photo lab. For some reason it seems that they have decided that it would be faster to get the files done by shooting with a Canon 1DS mark II and a slide copier rather than scanning with a drum scanner. I'm not sure they are using flash with the slide copier... maybe they are using tungsten and this alone could account for the difficulty.

I have come up with a recipe that seems to work pretty good and is reasonably easy to automate with an action. The results vary a bit from image to image but simple auto corrections at the end stage solve most of the remaining color cast issues. I'll break down my steps in the next email.

regards,

Lee Varis

http://www.varis.com

888-964-0024

___________________________________________________________________________

Hi all,

OK... here is my recipe -- which seems to be working with these older and possibly somewhat color faded negatives.

1. Sample the rebate - the orange base color in the edge of the negative.

2. Create a Solid Color adjustment layer using this color and change the layer mode to Difference.

This inverts the image and removes the orange mask color everywhere, however it forces the black point to "0" and makes it impossible to retrieve any detail from the shadows so...

3. Reduce the opacity of the Solid Color Difference layer to 95%

4. Create a Curves adjustment layer to address the basic color cast and re-set the white and black points to neutralize them and get them to "15" for black and "240" for white.

I also engineer a little mid point correction to adjust the bias back to the warm side as typically there seems to be a consistent cool bias in the overall color (I'm thinking automation here and there won't be an opportunity to find a real neutral in every image).

At this point the color is halfway decent but the saturation is very muted and there is not a lot of color variation anywhere so...

5. Flatten and convert to LAB.

6. Duplicate the Background into a new layer and change the layer mode to Overlay.

This punches up the saturation and color variation but it also crunches the contrast way up so...

7. Double-click to bring up the Blending Options/Layer Styles dialog -- in the Advanced Blending area, uncheck the "L" channel so that the Overlay is only applied to the "A" and "B" channels.

8. Flatten and convert back to the RGB of choice.

At this point some kind of auto color correction may be applied on a case by case basis to fix any remaining color bias issues with respect to neutrals or black and white points. The LAB move is magical and really amps up the color in a natural Kodacolor way. I see the Overlay layer working on these older negs at 100% but newer more colorful negs may need less so you can always reduce the opacity to get the desired intensity.

regards,

Lee Varis

http://www.varis.com

888-964-0024

___________________________________________________________________________

Hi Lee,

I tried your recipe on some negs I have and my results weren't satisfactory. If you have time in the future can you post a couple of your results to your web page so I or anyone else can take a look. I'm kind of interested in this idea as I have a few thousand colour negs I would like to post to one of my web sites but scanning would take too long and all I want really is a low rez image. Copying with a digital camera would make the project practical if there was a way to automate most of the colour correction.

Regards,

John Denniston

www.dennistonphoto.com

www.dirtbikephoto.com

___________________________________________________________________________

on 7/16/06 3:02 PM, Lee Varis wrote:

Lee,

Geez, this is a surprise. Reading the original post I guessed the idea was to save money by not using any other resources. Then I find they have all the equipment and experience a photo lab would have at their disposal.

Seems a bit short-sighted to get the capture done quick and leave your friend in the lurch with no tools/techniques needed for the conversion.

I'm admittedly partial to the conversion by a drum scan, that said, I'd estimate that a batch scan would get to 70% acceptable as finals, 20% needing minor corrections and the other 10% needed attention.

Lee Clawson

2/\V/\7 Studio

___________________________________________________________________________

On Jul 16, 2006, at 10:34 PM, John Denniston wrote:

I'm afraid this "recipe" might only work with these particular vintage color negatives. However, there is nothing particularly radical about the inversion method. As I mentioned before, I'm not at liberty to upload any of these images.

If you upload one of your images I can see if my approach could work -- I'll happily work it up and re-post.

regards,

Lee Varis

http://www.varis.com

888-964-0024

___________________________________________________________________________

On Jul 17, 2006, at 6:55 AM, Lee Clawson wrote:

Hey, I'm with you on this... its an interesting challenge to do it all in Photoshop but its kind of academic as the best method is to use time-tested cluts in the scanner - or something like Silverfast, etc...

All in all, its been interesting.

regards,

Lee Varis

http://www.varis.com

888-964-0024

___________________________________________________________________________

Hi Lee,

This problem of yours has been on my mind the last few days so this evening I did a test on one neg, shooting it with the camera preset to 2500K. In photoshop I used auto levels with "enhance per channel contrast" and no clipping. Then a simple invert.

The results were very good.

Regards,

John Denniston

www.dennistonphoto.com

www.dirtbikephoto.com

___________________________________________________________________________

There is a Photoshop plug-in NegPos, that does this inversion properly. In my experience it is about as good as SilverFast does for precalibrated films. NegPos could be adjusted to practically any film. As author of NegPos shows in his articles, inversion in Photoshop is designed incorrectly, no matter of image mode, and I found no way to do in PS what this plug-in does.

Best results could be achieved if you get RAW scan (using VueScan or SilverFast, or any other tool that could save tiff with raw scanner data), then you could open this file in PS and apply NegPos. You could try a bit muted version of NegPos for 30 days. Download it here http://www.c-f-systems.com/Plug-ins.html <http: //www.c-f-systems.com/Plug-ins.html> .

Sergei Antonov (http://www.flickr.com/photos/santonov

<http://www.flickr.com/photos/santonov> )