View Full Version : Purple Sky and Green Ground ~ Portra 400 and Epson 700 workflow

Frank Petronio
14-May-2012, 08:28
I've been using Portra 400 since, umm, Portra 400NC, and scanning mostly on an Epson 700. While I do not do a lot of landscape, I've been finding that a lot of mine seem to end up with green(er) land while the skies have too much magenta. Or I'll correct the sky towards green to get a cleaner blue and the land will go magenta. The yellow-blue balance seems OK but we know how yellow affects the green....

Is it me or is this film prone to cross curves? I can fix it by isolating one or the other but it would be nice to know my film and workflow is getting the color relationship right without making heroic measures. Thanks

John Rodriguez
14-May-2012, 11:11
I run into that as well. It's not just Portra 400, I've seen it with Ektar and Portra 160 too, although the effect seems to be strongest with 400. I've only noticed it in sunrise/sunset lighting, and have assumed that the result was somehow due to the red layer getting overexposed. I haven't tried using a cooling filter yet to confirm, I just wind up doing local corrections.

Brian C. Miller
14-May-2012, 11:18
Frank, what color calibration have you done with your film? Have you photographed a calibration target and run it through your software to produce an ICC file for that film type?

14-May-2012, 12:28
I would consider buying an iT8 target (transparency not printed) and profile the scanner. For me that made the biggest difference on my HP G4050 (I've only scanned 120 film so far).

Short of that you might look at ColorPerfect http://www.c-f-systems.com/Plug-ins.html a plug-in for PhotoShop. For me the iT8 target makes the biggest difference.

Frank Petronio
14-May-2012, 12:40
No, in fact I don't know how to profile my film. Can you point me in the right direction? I have one of those little XRite Color Checker Passport things I've used for digital and a Spyder 3 Calibrator.

Brian C. Miller
14-May-2012, 13:24
I use Profile Prism (http://www.ddisoftware.com/prism/). It is used to create a profile for your printer, scanner, and "input device," aka film (or digital camera).

Photograph the target using whatever you consider to be "normal" lighting. I use outdoors noon day sun.
Have the film developed at your normal lab, and then scan the target.
Run the scanned image through Profile Prism to produce the profile
When you scan that film in the future, select that profile for your scanner.

That should reduce your need for drastic color correction.

14-May-2012, 13:48

Selective Color adjustment layer.

Peter York
14-May-2012, 13:49
Is C-41 amenable to profiling with an iT8 target? I thought this process only worked with E-6 films.

J. Fada
14-May-2012, 14:05
Profiling the film would be your best bet unless your lab is causing the crossovers. I don't know if Epson software has the facility but Vuescan does. I have never profiled a film through Vuescan though so someone else would have to answer how it goes. Personally I think you are better off with a profile built by yourself using your film and processor but you probably don't have the equipment you need for that. Maybe someone will offer to do it for you that is near you.

I don't use ColorPerfect but I have played with it and it works well. You may want to give that a go. It simplifies scanning too since you don't have to jerk around with adjustments in the scanning software. This may be another reason to use Vuescan if you don't already.

For crossover problems I have found that splitting the file into highlights and shadows makes it easy to fix. I do this in LAB mode but it should work fine in any color space. I know you are experienced in Photoshop but if for some reason you don't know how to make this split- in the channels palette on the bottom there is a little dotted circle. Click on that for the highlight 50% and use that as your mask. Invert it for the shadow mask. You can figure out the rest. Maybe that will help someone else too.

14-May-2012, 21:34
Most likely you have the film base colour set wrong in your scanning/inverting software; that will always cause crossovers. What scanning software are you using and/or what software are you performing the inversion in?

Try with a demo copy of VueScan at least and first tweak the "Film base colour" sliders so that your shadows are neutral and only after that point, adjust highlight colour. The "Lock Colour" checkbox is your friend.

15-May-2012, 11:14
Is C-41 amenable to profiling with an iT8 target? I thought this process only worked with E-6 films.

C-41 (or any color negative film) and E-6 (or any slide film) use the same lens and same data to produce the image. The software will invert that image from a negative, but the premise is the same. Get your spectrum of colors to match the correct spectrum of colors and you'll have the correct spectrum of colors. If your color spectrum is correct, then you invert, you'll have correct spectrum for the inverted colors.

Tobias Key
15-May-2012, 11:30
I've had a lot of sucess by scanning the neg as a raw DNG without any adjustments from the scanner at all. Then I open the scan in CS4 invert, select curves, select white black and grey points and normally I'm pretty much there. There's lot's about this method online.

Peter York
15-May-2012, 13:31
I have not been able to find a defiitive answer on this, but I have read that color negative is, at best, a PIA to profile. All explanations seem to point to the characteristics of the orange mask - i.e., that it is not constant due to the scene shot, variances in development, or other factors.

If anyone has a deeper understanding of this, please chime in.

At the beginning of each photo session, if you photograph an iT8 target with color negative, then you can profile the film for each session, but not across all sessions.

It may be the case that a general profile will get Frank in the ballpark.

Henry Ambrose
16-May-2012, 17:18
I've found one of the main issues in scanning color negatives is getting enough exposure on the film.
Color neg goes funky when it does not get enough light.

There is no profiling of a scanner for color negative film.
The standard Epson software should work just fine for color negatives.
Fiddle around with setting good grey/neutral points using the eye dropper.

I used to think of color negative as an "interpretive" material - you make it look like what you want it to look like.

David Higgs
17-May-2012, 08:55
I'm with Henry - you have to make a choice with C41 on colour. This can either be viewed as a PITA or as a great creative tool. Quick and dirty - use PS to choose the white and black points - that get's me 90% of the way. The other 10 % is down to selective colour balance changes - as usual there are a number of ways in PS to do the same thing, and I'm sure Frank knows his way around this.
I've had zero success profiling film as changing light seems to change colour balance so much that unless you shoot a test target for each scene it's pretty much pointless. I also find you have to get creative with contrast curves as C41 tends to come out flat on scans for obvious reasons.
It's a bit fiddly, but the latitude and ease of processing and ease of scanning make it worth it.

Henry Ambrose
17-May-2012, 11:05
To clarify, I refer above to using the eyedropper tools in the Epson software for setting grey or neutral at the scanning stage.
You'll have to futz with it a bit, but before too long you'll get a feel for it.

As David mentions above, changing light conditions make it impossible to come up with a fixed setting. You are dealing with both the various color layers of the emulsion and how they react to varying intensities and color of light - AND- the relationship between the orange mask and those color layers. Different processing regimes change this relationship as well. Mix all that up and you need a human in charge to make it work. Once you "get it" you'll never want to go back to transparencies.

Frank Petronio
17-May-2012, 11:28
Well that is the way I've been scanning color neg since the 90s with a LeafScan 45. I am wondering if this cross-curve is built into the film itself?

17-May-2012, 16:00
1) try color perfect. Here's a tutorial that shows Vuescan at the beginning but the rest is ColorPerfect. http://benneh.net/blog/2010/09/25/vuescan-colorperfect-a-guide/?goal=yes This is by an 8x10 and 4x5 shooter with an Epson V700.
Color Perfect's plugin page http://www.c-f-systems.com/Plug-ins.html (click the Download for PC or Mac about a third of the way down the page)
Links to tutorials on settings for other software http://www.c-f-systems.com/Scanners.html

2) when I say that I profile my scanner I'm talking about the scanner itself, not the film. I never bother profiling the film, there's no real point as has been pointed out by a few here. But if your scanner's colors are off won't that throw the rest of the process off? If it's spot on won't that help (within the confines of the d-max of the scanner)? I scan with a Nikon LS-2000, Canon FS4000US, and an HP Scanjet G4050. While you can tell a slight difference between the three it's not like one jumps out as the worst of them all colorwise (sharpness is a different story). You couldn't say that before I profiled the scanners. With the HP it was heroics indeed for every single shot.

2-Jun-2012, 13:33
I use Colorperfect, seems to help a lot for accuracy and especially repeatability.
I have found the phenomenon you describe often occurs for example in bright sunlight, closeish to midday, and when the negative is a bit overexposed...
Just my findings.

Frank Petronio
2-Jun-2012, 13:45
Yeah exactly, I do shoot in "bad" and extreme light sometimes and this was mid-afternoon in the Utah desert.

Jim Michael
2-Jun-2012, 14:07
Crossovers are a pain. I was just playing around with someone's image the other day that had that very issue. I think the thing that worked best was going into the hue & saturation adjustment layer and selectively adjusting hue on one color. I was probably in LAB mode.

21-Oct-2015, 13:55
This post is a bit old but I was printing Kodak Ektar and Portra 160 (4x5) last year using the RA4 processor and scanning with an Epson V700 and noticed a red or magenta color cast sometimes . Once it was there I would not totally go away in the Ra4 process and difficult to deal with when scanning. In my opinion it happened during film processing as the lab I was using only runs film 2/month. I started using a lab with better quality control and those issues went away. It is also possible that there is something wrong with the film.

Drew Wiley
21-Oct-2015, 15:12
Don't blame the film. The quality control on these products is superb. Look elsewhere.