转自：http://www.penttala.fi/blog/
THE FASTEST AND EASIEST METHOD FOR REMOVING THE ORANGE MASK IN PHOTOSHOP

实验了一下，感觉方法挺好。

继续：（去色罩步骤太繁琐，有时间再写）

上图，调整数据同上。

我根据网上找的一段关于色罩的数据来实验的：
Marvin Hoshino , Nov 01, 2007; 05:26 p.m.
There must be journal references but the only one I know in book form is Lloyd Varden's appendix to the 1960s reprint of Joseph S. Friedman's History of Color Photography. Varden is ironic and vague and he writes more about the industrial spying angle.
Here goes. As your first response pointed out, dyes are impure, failing to absorb light frequenecs they should for good reproduction. The easiest way to understand this is to imagine ideal dyes for C, M and Y, call them C*, M*, Y*.
Then the actual dyes could be represented by
C (actual) = C* + 3%M* + 1%Y*. This would be for some actual dye where there is an equivalent of 3% magenta and 1% yellow unwanted contamination. (A perfect C dye would not absorb any M or Y.)
M (actual) = M + 2%Y*
Y (actual) = is relatively good.
Now to take just the cyan dye. In an area of the image where there is only cyan (which is rare) nothing can be fixed by masking. But imagine an area when the color is supposed to be 100C* + 30M* + 10Y*. With acutal cyna and magenta dyes the color becomes 100C* (+ 3M* + 1Y*) + 30M* (.30x2Y*) + Y* which equals 100C* + 33M* + 10.6Y*. So there is now too much magenta and a little too much yellow.
Here is where masking comes in. What we want to do is to subtract from the magenta densities an amount equal to the unwanted magenta in the cyan dye. (And also, to subtract from the yellow an amount equal to the unwanted yellow contamination in both the cyan and magenta dyes. So what we want is an image that is 100C(actual) + approximately 27M(actual) + approximately 9.4M(actual).
In the dye transfer process, which uses color separation negatives, you would make the three separation negatives. Then from the magenta negative you would make a thin positive (a mask). Then you sandwich this mask with the cyan negative when you expose the cyan negative. This subtracts magenta values equal to the cyan dye impurities.
And you do the same for the magenta negative. This masking leads to all kinds of issues. Since you are sandwiching a weak positive with a negative, the negative has to be processed more contrasty than if you did not use the mask. The mask must be exposure and processed to represent the dye impurity of the dye you are trying to correct.
For color negatives, Thomas Hanson (I am not sure of his name) realized that you could do this by using a color dye that starts not clear but colored. So for the cyan dye the original dye is a pale red-pink and during development it is converted to a cyan. After development there are two images related to the cyan exposure -- a negative cyan image and in exact register a positive red-pink image. Similarly for the impure magenta dye there are two images. A negative magenta image and a positive yellow image.
The mask in color negative film only looks like it covers the whole film. It is thinner in areas of the image where the cyan and magenta densities are the strongest. So where the cyan layer goes from 0 to 100% the mask runs from (say) 30 magenta and 10 yellow (when Cyan is 0) to zero (where the cyan is 100C).
Okay, so you can completely correct for dye errors in the negative film. Using the same reasoning you can also correct for the presumed dye errors in the color print material you will be printing on. Its cyan also will have unwanted magenta and yellow so you make the negative to subtract those unwanted impurities.
This means, to answer your question, that a mask is not a filter. Its use in printing cross-processed negative would be to make the filter pack closer to what you would normally use with a regular color negative. Otherwise you would have to add red and yellow to the filter pack. Since a real color negative material has this positive superimposed onto the camera image, the contrast of color negative materials is usually lower than cross developed film.
If you are digitizing your film, the scanner expects to see a masked image. You could try scanning the images as if they were slides and flipping them in Photoshop. A guess is that the best way to go would be to do the flip in LAB color.
If you are using a camera and flash to digitize, the CN-mask should not be necessary. No harm but not gain.
To answer your other questions, the mask is in the film base before development. But the point is that some of it is destroyed in the areas where there is actual image, so that in some places there is no original dye left after development.
In cross-developed film the "bluish mask" is probably not a mask but just and overall stain on the film. Just a guess.
One thing you could do is shot a color swatch card in both regular color negative material and cross-processed film (and maybe color slide film normally processed). Then in Photoshop you could find a curve layer that converts between them (best done, again, in LAB) and save the adjustment layer. Or save the curve in CURVES and reuse it.
PS the Joseph Friedman book explains all about masking color separation negatives, but not the integral color-coupled masks of color negative film.

看不太懂，根据文中列出的公式调整的。 本帖最后由 铁笔书生 于 2016-7-1 22:18 编辑

去色罩，然后根据灰板调整RGB通道，红+1%，蓝-7%。

这张找不到可选的参考点，只去了色罩。

试试新的去色罩办法，先去色罩后调整白平衡。

Removing Orange Mask from Colour Negative Scans http://www.filmshooting.com/scripts/forum/viewtopic.php?f=1&t=24780
找到一个方法，大神可以参考一下。
http://www.brianpritchard.com/why_colour_negative_is_orange.htm 介绍色罩的

负片就是猜颜色

负片色罩难点在还原现场的明暗和对比度，所以这个个办法并不准确。

非常感谢，这个方法感觉不太行，正在实验新的方法。

斗胆调整了下色彩平衡

试试图床：（在PS里面录制了动作，快速取色罩和调整图片大小）CANON EOS 300V, 富士C200

发现用PS的自动色调不太靠谱，现在去色罩后，就手动调整一下色阶。

片子冲的脏，远眺的几张是CANON EOS 300V + 2014年到期的富士C200。

远眺

我站在高岗上远处望

挖掘机，参数同上。

自己顶一下，去色罩练习，绿叶（海鸥205，富士C200，阳光16法估测光，显影不均） 本帖最后由 铁笔书生 于 2016-6-17 17:33 编辑

谢谢分享

第二步：