Blue Cast in Drum Scanned Files

[Eric James]

I recently received three Tango drum scans from WCI - all scans have, to varying degrees, a blue cast. My screen is calibrated as best I know how (spider puck). I have attached a jpg of the most troublesome image. The blue channel is blocked up in the highlights; the green and red channels are okay. In the original transparency the snow is white as I've even seen snow on film - it doesn't appear to be over-exposed beyond the films range, it's just white. Any measure I've used to remove the cast results in very jagged histograms.

I'm not sure how to proceed - any thoughts?

[Michael Gordon]

Jagged histograms are not out of the ordinary. What matters is how the print looks.

WCI has done lots of drum scans for me, although it's been a while. I don't recall one that didn't require color correction. If you can confirm that your transparency is not clipped in the highlights, then they should willingly rescan the chrome for you.

[Doug Dolde]

Try pulling up the levels tool and pick what you think is pure white with the dropper on the right side (highlight dropper). Should make it turn white.

[Michael Gordon]

Doug's method may result in clipped highlights if used at its default (L* 100). Be sure to change the L* value of the dropper to match *L value of the snow where you're clicking.

[Eric James]

I spoke with one of the scanners at WCI who said that they struggled with this image and suspected that something was amiss in the bleaching phase of E6 processing. This could be the case. He has asked that I return the tranny so that he can try a few scanning tricks. Thanks you for you input.

[Bruce Watson]

In the original transparency the snow is white as I've even seen snow on film - it doesn't appear to be over-exposed beyond the films range, it's just white. Any measure I've used to remove the cast results in very jagged histograms.

The human visual system is notoriously bad at judging color -- too many automatic adjustments going on.

I'm curious, did you correct color during exposure? That is, did you use any warming filters? The reason I ask is that the biggest light source in your picture is that big blue sky. The light you see reflected off the snow is distinctly blue, as is the light reflecting off everything else. This is especially true at altitude and during the middle of the day. It's not unusual to see color temperatures over 10000K then, and it takes a lot of warming filter to correct for it.

Charles Cramer had an article in View Camera a couple of years ago about that - 2004 maybe?

All I'm saying is that it's conceivable that what the scanner sees on the film is a faithful recording of what the film saw, which was very blue. The "traditional" correction for this in Photoshop is to lower the amount of blue in the image using the curve tool. You may need enough correction that having the image scanned at 16 bits would be useful.

It would be interesting to hear about the results from WCI's rescan.

[Eric James]

Yes, I used a Nikon A2 filter for this. I hear what you're saying about the eyes adjusting, but in other transparencies I have of snow scenes I can see the blue, but in this shot I don't see it - for once I achieved my goal with filtration. And yet the scan it very blue, not just in the snow, but everywhere. I’m curious about why only the blue channel is blown in the highlights. Is it because the predominant RGB component of the snow is blue, and there are patches/pixels of snow that are greater than 256; or is it because of the scanning technique used. I'm hoping for the later and I'm thankful that the scanner was willing and happy to try another technique or two, because the shot sucks if the snow is blown out, and in correcting the color cast the snow takes a hit.

I also was hoping that a 16-bit upgrade would help. He said that they find that 16-bit scans help with subtle blue transitions in some skies; though he was willing to try, he wasn’t too hopeful – maybe he wasn’t thinking about post-scanning wiggle room as you are. (I'm not quoting him verbatim - it's just what my novice mind heard him say.)

Thank you,
Eric

[Eric James]

The rescan did the trick: The blue highlights are no longer blocked up; the blue cast is gone; and the file is printable

[Ted Harris]

One point I found troublesome in reading through this thread ... do I correctly imply that they do not, as a matter of course, do 16 bit scans? Curious.

[Michael Mutmansky]

Ted,

File size. WCI and others charge extra to deal with 16 bit files in general, and argue that most people don't need them.

I buy that argument for general prepress work most of the time, but I think it falls a bit flat when comparing high quality photograph reproductions (inkjet or other high quality output), especially when there has been some tonal manipulation in the image.

I think, as Bruce said, the greater question here is whether the original scan is faithful to the chrome. If not, then there was clearly a scanner operator or equipment error. The improved scan doesn't really tell us where the problem was, because they probably gave the image a nudge, regardless of what the scanner saw, and whether it was operating properly.


---Michael