Shortest route to an accurate scan

[Daniel Moore]

After untold hours of reading, learning new software, trial and error, I'm still doubltful as to knowing the shortest route to an accurate scan. Profiling with an IT8 recorded on the same type of film had climbed to the top of my list but I've yet to produce a scan I'm satisfied with. I'm limited with a V700, for now, for 6x7 and 4x5, but some folks are content with that and I again feel I'm missing something. Including a Colorchecker SG was another route I thought to take, and do own one, but that would serve to neutralize color casts, not exactly producing accurate sunsets as one example of a desirable color cast being neutralized, and no longer being accurate. So, realizing what an expansive topic this could grow to in order to meet it's needs, I'm hoping the opposite can be achieved, a bit of wisdom or boiled down order of operations with an eye to rapid and accurate color and tone. Or should I resign myself to spending the second half of my life in photoshop tweaking curves and levels until my mental scan and the real one begin to line up? I've always loved choosing film for it's look and rendering of a scene, but I'm so tired of adjusting by eye to try and get it back.

[Peter De Smidt]

Have you tried the IT8 profile route? It is the fastest way to the most accurate color scans.

[Daniel Moore]

I have, I bought a Wolf Faust recently, IT8 of provia, my most used film, and my results are ok, but that's the problem, there's no life in them. The scans appear 2 dimensional and flat, especially, and this is a big, big, big issue, when compared to the same scene captured at the same time with my D700 and good glass. Color, tone, clarity, 3 dimensionality, all are far inferior and I simply can't imagine ever being satisfied with that when viewing a slide on a lightbox reveals it's all there in the film itself. I created a profile with Profilemaker Pro.

[Joanna Carter]

Step 1. Profile your monitor using something like the Spyder or EyeOne, etc.

Step 2. Profile your scanner, using a 4x5 target from Wolf Faust, for the film you are wanting to scan.

Scan the film into Photoshop with the colour management on the scanning software turned off. Then apply the appropriate profile to the scanned image in Photoshop.

You should now have a perfectly balanced scan.

[Peter De Smidt]

I agree with Joanna.

There are lots of ways all of this can go wrong, such as improper setting of the scan parameters, incorrect and/ or double profiles being applied....

When you scanned the target (and the pictorial slide), how did you set the white and black points? Did you disable all automatic adjustments? How did the profile get associated with the scan?

[Daniel Moore]

I use a DTP-94 (Optix XR) and Coloreyes Display pro to calibrate my EIZO CG241W, the DeltaE from it's profiles are better than EIZO's color navigator.

I did not set black and white points and all auto adjustments were turned off. The unaltered scan is brought into PS and the profile is assigned there and PS is told to ask when opening a file with no profile tag.

[Joanna Carter]

I did not set black and white points and all auto adjustments were turned off. The unaltered scan is brought into PS and the profile is assigned there and PS is told to ask when opening a file with no profile tag.

Which software are you using to scan?

Have you turned all colour management off?

Typically, a scan with no colour management will look decidedly drab. However, after you have assigned the correct profile, you should then proceed to add a Levels adjustment layer and move the black and white points, to expand the range of the image to fill the histogram. Did you know that, if you press the Alt key whilst adjusting the black and white points, you will see an ultra-high contrast version of the image, where you can see if the highlights are blowing or the shadows are blocking?

Once you have added the Levels layer, you can then add a Curves layer to further adjust the contrast.

But you do need to understand that the dynamic range of a trannie is nowhere near the capabilities of the scanner, thus the unadjusted scan will look horribly flat.

[Peter De Smidt]

You might try setting black and white points in the scanner software before making the IT8 scan. Find the values of the white and black squares, and the black and white points so that they are a bit lower for the black point than the black square and a bit higher than the white square for the white point. Make your profile. Now use those same settings and make a scan of your pictorial negative.

You can test things a bit by bringing the it8 scan into photoshop with no color managment. Now assign your profile. The target should instantly look much better, right? Now take the eye dropper and check the values of the black, middle gray and white squares. What are their readings? Make a curves adjustment layer. Use the black eye dropper in the curves dialogue box to click on the black square in the target. Use the gray eye dropper to click on a middle gray, and use the white eye dropper to click on the white square. Toggle the adjustment layer on and off. Is it better on? If so, you can save the curve file and apply it to other scans when you bring them into Photoshop. The danger with all of this is clipping detail. You can always adjust the curves slightly so that you don't lose detail.

[Joanna Carter]

You might try setting black and white points in the scanner software before making the IT8 scan

Although I, sort of, agree with you Peter, in the Epson software, this can't be done if colour management is turned off.

I have tried this approach and found the possibility of clipping was greater than the workflow I suggested, mainly due to the guidance on clipping, given by Photoshop, when holding the Alt key and adjusting the levels.

However, I think you have a valid approach, unless, like I have sometimes done, you forget to normalise everything else in the adjustments

[Peter De Smidt]

I agree with you, Joanna. I didn't know that the Epson software doesn't allow control of BW points if CM is disabled. And there is more of a chance of clipping with the method I outlined. I'd only try it if Joanna's method doesn't work out, although I have no idea why it wouldn't.

There is some good info at: http://www.hutchcolor.com/Free_stuff.html