DSLR Color Negative Scanning- Removing Orange/Magenta Tint

There are many threads that contain discussion(s) regarding the tint and the cause.

Thought I would start one exclusively for solutions to remove the tint.

I have taken a small step in the direction of DSLR scanning: Nikon 7100 / 40mm DX Macro lens / ES-2 Scanning set / Lightroom preset for curve inversion.

Like others, I quickly found out that control sliders work in reverse after negative inversion.....but if you export the inverted image and reimport everything works as you normally expect.

Initial results are quite good, and I now have images that appear better than the original 1hr lab prints.

I will have to be more careful about cleaning and loading negatives, and improve my light sources etc.

But the biggest problem is that damn tint ! I have tried to use eyedropper adjustments- some work perfectly, some just change the tint.

Is there a plug in to address this ? Is there a log of settings for various film types ? What are others doing to remove the tint ?

Hopefully this thread can become an ongoing point of reference for this topic.

There are quite a few previous threads on this subject, and as I've suggested before, using the camera's custom White Balance control can remove most of the orange contrast-mask and make the inversion process easier.

(The orange mask isn't there for fun or annoyance. It serves to correct deficiencies in the dyes used in the film. So cannot be completely removed with a simple filter mask, since it changes density when adjacent to other colours.)

I use flash as the illuminant for the negative, which provides a lot of room for colour-correction. Using an already 'orange' continuous light source to illuminate the negative very much limits your options.

I've found that using flash as the illuminant (with a CT of around 6500K) and setting the camera WB to around 3400K + 3M is in the right ball-park for most Kodak C-41 films. If you have a blank frame or some unexposed and processed film leader, you can use this to take an exact Custom Balance of the film mask, and obviously that's more precise.

With the base black level made neutral after tone inversion, you can then often use auto-colour or auto-levels to get a reasonably close end result.

Alternatively, you can just feed your camera 'scan' into ViewScan and let it work its magic.

Like others, I quickly found out that control sliders work in reverse after negative inversion.....

Not in PhotoShop or Gimp they don't. And sliders are a crude way of doing things. I suggest you get to grips with the Curves tool, which is far more versatile.

Conversion of color negatives has been discussed exhaustively on PNET in recent months. The simplest method is to use Silverfast HDR, which is the core software used to convert scans. Make sure the camera color balance is set according to the copying light source, preferably using the empty slide holder. Secondly, crop away any border area from the image, as it tends to adversely affect the color balance.

There are manual ways t do the conversion, as well as a completely useless option in the Curves menu of Photoshop. There are too many variables to consider. Even small changes in exposure affect the color balance. The same nominal emulsion behaves differently between 35 mm and roll film. Silverfast HDR is equipped to handle these variations, unless you're looking for a career opportunity.

Personally, I wouldn't trade these conversions for the most primitive digital camera, but my archives go back over 50 years, and are largely unscanned, before my kids (now adults) were in diapers and digital imaging was still fantasy.

Leica M3 + Ektacolor 100, Silverfast HDR conversion

As I said, it ain't easy being negative. Magenta is an easy touch-up, but I left the image in the as-is condition. Negative color is like cooking - season to taste.

I've used a Gretag-MacBeth color card in the past, which I found problematic at best, even using a profiling tool.

There are quite a few previous threads on this subject, and as I've suggested before, using the camera's custom White Balance control can remove most of the orange contrast-mask and make the inversion process easier.

 

(The orange mask isn't there for fun or annoyance. It serves to correct deficiencies in the dyes used in the film. So cannot be completely removed with a simple filter mask, since it changes density when adjacent to other colours.)

 

I use flash as the illuminant for the negative, which provides a lot of room for colour-correction. Using an already 'orange' continuous light source to illuminate the negative very much limits your options.

 

I've found that using flash as the illuminant (with a CT of around 6500K) and setting the camera WB to around 3400K + 3M is in the right ball-park for most Kodak C-41 films. If you have a blank frame or some unexposed and processed film leader, you can use this to take an exact Custom Balance of the film mask, and obviously that's more precise.

 

With the base black level made neutral after tone inversion, you can then often use auto-colour or auto-levels to get a reasonably close end result.

 

Alternatively, you can just feed your camera 'scan' into ViewScan and let it work its magic.

 

Not in PhotoShop or Gimp they don't. And sliders are a crude way of doing things. I suggest you get to grips with the Curves tool, which is far more versatile.

Color mask is present for a reason, but makers could not know how it would impact conversion 30-40-50-60yrs after processing. It serves a purpose, but an unintended consequence is a definite tint after scanning. It is an issue that has to be dealt with, and it is a nuisance

Sliders in LR work in the opposite direction for scanned and inverted slides. Exposure, Black levels, Fill Light, Recovery etc. Curves are more useful and I use curve adjustments to process. But if one does use sliders then one has to remember they work in reverse, or export and re-import the file to have everything working as it should.

I have not tried ViewScan or Silverfast...will investigate. Thanks.

If you use Silverfast HDR, do the conversion first, then the adjustments. Since you're working on a positive copy, the sliders work normally.

A way to do the conversion manually is to...

• Create a RAW image of the negative, using manual white balance adjusted to the light source.
  
• In Photoshop Levels (histograms), adjust the RGB sliders to fit each color respectively (removes orange mask). Auto by channel may work.
  
• Invert (ctl-I) the colors and save as a new file
  
• Make fine adjustments to the converted file in Photoshop or Lightroom
  

Make sure the camera color balance is set according to the copying light source

Create a RAW image of the negative, using manual white balance adjusted to the light source.

By all means shoot RAW, but sorry Ed, I have to strongly disagree that setting the camera WB to the light source helps with the contrast mask in any way.

The orange contrast mask should be neutral black in the final positive, but it's red/yellow shifted - blue deficient - in the negative. This means that the shadows after inversion - highlights in the negative - of each colour channel don't align and need quite drastic correction to bring them into alignment. The overall result is that the blue channel needs a lot of extra gain and contrast, which can reduce its colour refinement and increase noise.

By balancing the camera WB for the contrast mask, part of the load of aligning the three channels is shifted into the camera, and off the purely digital manipulation done by the image editor.

All digital cameras have their sensor and filter array optimised for close to a daylight White Balance. So presenting them with an uncorrected tungsten-balance image, which a colour negative effectively is, puts a strain on their ability to get the best dynamic range from each channel. Whereas correcting the WB to match the mask equalises the RGB channels to a great extent.

However, this requires that the illuminating source has a high enough colour temperature for the lack of blue in the mask to be properly compensated. Otherwise most cameras are still going to struggle with applying a combined CT (illuminant + mask) below 2400 K or thereabouts.

That's why purpose-built film scanners always have a daylight equivalent light-source. Otherwise the blue channel signal is weakened and noisy.

By all means shoot RAW, but sorry Ed, I have to strongly disagree that setting the camera WB to the light source helps with the contrast mask in any way.

Strictly speaking, the WB should be set using the film holder without film present. It may not matter if it is a precise match as long as it doesn't change depending on the film and image content. The orange mask varies, as you noted above, with the density of the image - more image, less mask - which creates a bias which is very difficult to overcome later.

I have settled on using a screw-based LED lamp balanced for daylight. However I have had equal success using incandescent bulbs, halogen lamps and warm LED bulbs. The process should be as consistent as possible up to the point of conversion. That said, I find that using auto exposure helps more than it hurts, since color negatives taken under poor light tend to be rather thin unless exposed for the shadows (letting the highlights care for themselves).

LED replacement bulbs have a relatively smooth spectrum, except for a bump at 480 mmicrons, which is the excitation wavelength used for the phosphor. Perhaps a little more blue helps.

Scanners don't necessariy have a white light source. The composite light of a Nikon scanner has a pink cast, and some flatbed scanners are bright green. Drum scanners use an incandescent light source.

Strictly speaking, the WB should be set using the film holder without film present. It may not matter if it is a precise match as long as it doesn't change depending on the film and image content. The orange mask varies, as you noted above, with the density of the image - more image, less mask - which creates a bias which is very difficult to overcome later

Well, I've tried camera scanning with a 'normal' White Balance based on the illuminant, and with a Custom WB that compensates for the orange mask. The compensated images take far less adjustment after inversion than the 'straight' copies, and consequently appear to have better colour quality.

That the mask varies in density with the density of other dyes is totally irrelevant here. That happens regardless of the illuminant CT, or camera WB.

What mask compensation does is to align one end of each RGB curve, and to provide a common 'pivot' point for the gamma adjustment of each curve. Otherwise the curves are staggered horizontally as well as vertically.

The other advantage is that the necessary alignment is achieved in camera by adjusting the gain of each RGB channel, rather than digitally, which just throws bits away.

The difference would probably be moot if a true 16 bit colour depth was available from the camera. But it isn't.

Silverfast HDR works best if the color balance of the negative image is consistent. That balance can change if the lens, lamp or diffuser is changed, so setting the white balance according to the actual conditions is important.

RAW image don't have a "white balance" per se. WB is just a number used to render the RAW image for display. In other words, you always "throw bits away."

Any image with a color depth greater than 8 bits is stored as a 16 bit/channel file. This affects how roundoff errors in rendering tend to cause posterization, revealed as gaps in the histogram. It is doubtful whether the difference between 12, 14, or 16 bits/channel means much outside of the vary densest areas of the image. In fact, the DMax of negative film is fairly low (2-3), well within the dynamic range of a modern digital camera.

What mask compensation does is to align one end of each RGB curve, and to provide a common 'pivot' point for the gamma adjustment of each curve. Otherwise the curves are staggered horizontally as well as vertically.

That is essentially the same process as the manual method I described above. I don't see any way that could be accomplished by adjusting the white balance. A related shortcut is setting the orange border around the image as the white point. That doesn't work either, since the density of the mask is variable within the image, and is present even in the subject's highlights. (The border is whiter than white following conversion.)

Through all this discourse we have ignored the fact that the colors and mask density vary from one film manufacturer to the next, as well as between film formats (35 mm, etc) within the same nominal emulsion. Silverfast HDR provides templates for most of the variations, making conversion more convenient and accurate. The PNET culture dwells on the search for free stuff. IMO, $100 is better spent on a good solution than many hours of trial and error, not to mention beating the subject to death in the forum.

Correct, raw's don't have a WB whatsoever (or per se) :). They have a suggested WB in the form of metadata but considering the values define a large range of possible colors, that every camera will report the same WB sightly differently or that cameras don't measure WB as the proper tools to do so would (a Spectrophotometer), it's not anything anyone needs to consider.

And yes, the orange mask vary considerably by manufacturer and through exposure even from the same file stock. The old Imacon Scanner software (FlexColor) did a very respectable job inverting the orange mask, the trick was to make a marque, hit the auto correction, move that around until you produced a desired result, then save that as a preset. It would work with other neg's shot in a similar fashion but not all of course; back to trial and error using the marque and auto. But once you hit a sweet spot, the results were quite good.

That is essentially the same process as the manual method I described above. I don't see any way that could be accomplished by adjusting the white balance.

Off course it adjusts the RGB histograms, you only have to examine them in the curves tool to see that. Setting a WB to that of the light source + mask makes the mask area 'white' and therefore neutral after inversion.

It also stops the 3 separate RGB histograms being quite so spread out horizontally (as represented in curves), and effectively gives a bit more adjustment headroom for difficult negatives that might otherwise 'cramp' the red or blue channels.

Indeed the mask varies between make and type of film. However, some approximate compensation aligns the histograms better than no compensation at all, but what I suggested was to set a custom WB from a blank frame or processed leader. Then there's no need to poke about with the eyedropper tool in an attempt to find absolute black in the image, nor to include a film margin in the scan.

Silverfast HDR provides templates for most of the variations, making conversion more convenient and accurate.

But as pointed out by another poster, your example of Silverfast HDR's output was far from perfect, and if every image needs some hand tweaking, then it really isn't serving its intended purpose.

FWIW, I find that some film types just don't scan well as regards colour accuracy - maybe some incompatibility between dye-transmission and sensor filters, or between illuminant spectrum and dyes.... or maybe just careless processing... whatever.

Such problematic negatives often respond well to a small hue rotation in the red and/or yellow channels using the hue-sat tool. That way grass can be kept grass-green without turning the sky cyan (or tree trunks and rocks becoming magenta).

I don't know the inner workings of Silverfast, but I suspect that a hue rotation targeting specific dye colours is outside of its remit.

Since when did negative scans not need tweaking?

Ed, you just don't understand. It's MAGIC, okay. Once you have it "dialed in" then any negative, regardless of lighting, emulsion, whatever will be perfect "straight out of scanner"! If not, then you are not a photographer, because you will have to CHEAT!

Negative color film is attractive for its tolerance to varying light conditions and exposure. However that tolerance does not extend to the darkroom (or Lightroom). furthermore that exposure tolerance is mainly for overexposure, so thin, underexposed negatives can be very difficult to render well. This extends to B&W negative too. I generally increase exposure for negative film for practically everything, either by "exposing for the shadows" or using a lower ISO setting, as much as one stop.

In the mid-60's there was a race to increase the sensitivity of film. Tri-X was 160, then 225, and ultimately 400, with little evidence that the chemistry had actually changed. I wasn't into color until much later, but I suspect the same transformation occurred, because when exposed as advertised, color negatives tend to be too thin. Don't lend to much credence to the sacrosanct status of ISO standards. ISO standards are simply consensus standards - mutually agreed between stakeholders, not an engraved platinum bar in Paris. (The much-touted ISO 9000 standards are based on record-keeping, not production quality or effectiveness.)

But I digress. Exposure has a profound effect on color balance. Prevailing white balance has a profound effect on both color and exposure (incandescent lighting produces thin negatives, possibly due to blue deficiency). When it occurs in reversal film, we toss those slides because there is no way to fix them. In negative film we attempt to recover if possible. In a studio, whether for portraits or product photography, we have complete control over lighting and exposure. We may not have to tweak every image, because most of the tweaking is done up front when setting up the process. The rest of us have to work with the hand we're dealt.

Cheating? Ansel Adams did not build his reputation with contact prints.

That's why I like to shoot chromes like Velvia 50. You see what you get immediately, know whether it's exposed properly, and then can scan and adjust colors more quickly.

That's why I prefer digital. The colors are accurate and you can evaluate the quality of images, making any necessary adjustments before leaving the site.

Velvia 50 (or 100) is fantastic for certain scenes. It is not, however, forgiving of either exposure or white balance. While it is great for scenery and wildlife, skin tones are generally unpleasant, and shooting under overcast skies results in colors beyond correction.

Since when did negative scans not need tweaking?

And someone's forgot the major differences between global and selective corrections (tweaking). Scanning software, unlike Photoshop and the like are global.

The low gamma (contrast index) of color negative films gives them the exposure tolerance, and maybe some color tolerance.

In the early days of color negative films, there was only one type (that is, optimal color balance).

After not so long, it was found that not all color balance errors can be corrected in printing,

and so tungsten balanced color negative films were created. (Type L is the one I know.)

(Though maybe not all the types, A, B, F of reversal films.)

At about the same time, black and white films started recommending blue bulbs for flash.

In, I believe, the late 1950's, ASA changed the scale, reducing the safety margin for

underexposure, increasing ASA values. But yes, there has been a slow increase in available

ASA values over the years.

The low gamma which gives the increased tolerance for film, is then corrected with a higher

gamma in printing. (Or scanning.) It decreases the exposure and color tolerance when printing.

The rule for color printing is to find the filter that, when you look through, corrects the color,

then add one half the density value of the complementary color. Exposure and color

errors are amplified.

I suspect that it is both the mask and the low gamma that complicate using a DSLR

for color negatives. Note that one thing you don't need is high dynamic range.

You do need to accurately record the low dynamic range in the negative.

One can reduce the colour cast problem significantly when 'scanning' colour negatives with a 35mm or digital camera

using a macro 1:1 copy set up and Wratten Colour Compensating (CC) filters.

In the past I have used a combination of a Wratten CC50 Blue and a CC50 Cyan placed over on a small daylight balanced light box.

The rig used the two CC filters overlaid on the light box under a 35mm negative carrier or mask to hold the negative strip.

Simply the combination of the filters plus the clear sections of the orange base/mask produce close to a neutral grey result.

If interested for more details have look at an instructional video I made a while back :

It is not suitable for bulk batch processing but for those interested for occasional copy work and with access to a 1:1 macro set up.

That's why I prefer digital. The colors are accurate and you can evaluate the quality of images, making any necessary adjustments before leaving the site.

 

Velvia 50 (or 100) is fantastic for certain scenes. It is not, however, forgiving of either exposure or white balance. While it is great for scenery and wildlife, skin tones are generally unpleasant, and shooting under overcast skies results in colors beyond correction.

I shoot digital, Ed, as well. But this thread is about film. I also use Velvia 50 with my landscape shooting, not for people, normally. I agree it can be difficult to correct but not impossible. Here's both Velvia landscape and people shots. I think Velvia 50 is exciting. Portra is great for people, but kind of boring for nature and landscapes. Velvia's limited number of stops makes contrast more daring eye candy. With scanning, you even get to play with it's natural benefits.

https://www.flickr.com/search/?sort=date-taken-desc&safe_search=1&tags=velvia&user_id=55760757@N05&view_all=1