0
\$\begingroup\$

I'm photographing a really dark scene with my iPhone 7 Plus in RAW, and exporting it to JPEG in Lightroom 5:

enter image description here

Notice the purple cast throughout, especially in the corners. The in-camera JPEG doesn't have this problem:

enter image description here

How do I process the RAW in Lightroom 5 to remove the color cast?

I tried the following:

  • Adjusting the white balance, both temp and tint. They either don't eliminate the color cast, or change the color of the color cast (rather than eliminating it), or mess up the color balance of the entire photo, making everything look green.

  • Reducing the Blacks via the slider eliminates the color cast, but also darkens the entire image to the point it's useless.

  • Tweaking the Whites sider doesn't eliminate the color cast, either.

  • Reducing the exposure of the scene, but that makes the entire photo dark.

  • To see if Lightroom's RAW processing is the problem, I tried exporting using Preview, but that didn't help:

enter image description here

None of these work. How do I edit the RAW to eliminate the color cast, like the in-camera JPEG?

ISO and shutter speed are already at their max — 1600 and 1/3s.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Just a suggestion, but have you tried selective reducing the purple saturation? That could (maybe) give you better results... \$\endgroup\$
    – Ryan
    Commented Jul 28, 2017 at 6:07

3 Answers 3

1
\$\begingroup\$

How do I edit the RAW to eliminate the color cast, like the in-camera JPEG?

You might not be able to match it exactly. The in camera JPEG processing engine is customized to be applied to a single sensor - the one in your camera. It was loaded at at the factory with customized light response curves based on the way that specific camera responds to various levels of light at various intensities and colors. To the best of my knowledge Apple does not make such JPEG processing algorithms publically available for users to apply off camera.

To get as close as possible you're going to need to apply correction along both the color temperature axis (blue←→amber) and along the hue axis (magenta ←→green). You might need to use custom light curves for the R, G, and B components or even use a tool that lets you change the color ratios in LAB color space. There might also be some benefit to using a Hue-Saturation-Luminance, Hue-Saturation-Value/Hue-Saturation-Brightness, or Hue-Saturation-Intensity tool.

In the end it looks like your phone may be significantly reducing color saturation or even converting the very dark image to monochrome (B&W). Have you tried reducing the overall color saturation to near zero?

\$\endgroup\$
-1
\$\begingroup\$

You can try using the "remove chromatic aberration" option. Although it will not probably give a great improvement.

A simple correction you can do is reduce the saturation of the magenta(as the raw appears to have mostly magenta in the noise). I find that this may sometimes yields results that are maybe not that close to the original lighting of the scene, but are mostly good. Then you can increase the warmth as the jpeg appears to have a more yellow/orange hue.

\$\endgroup\$
-1
\$\begingroup\$

This purple "cast" is due to excess chromatic aberration noise in the scene. Because there is not enough light in the scene, iPhone pumps-up the ISO causing this. If your target is a JPEG image, in my opinion it's better to go with an image stacking app like proseezion where you can set exposure and dynamic-range to taste with a virtually noiseless image and better detail.

\$\endgroup\$
3
  • \$\begingroup\$ Do you mean colour noise or chromatic aberration. AFAIK there is no such thing as "chromatic aberration noise" \$\endgroup\$ Commented Sep 10, 2017 at 5:21
  • \$\begingroup\$ @Romeo You are right, chroma noise is the correct term. \$\endgroup\$
    – manta
    Commented Oct 24, 2017 at 20:52
  • \$\begingroup\$ chroma noise looks like varying color, not color imbalance \$\endgroup\$
    – szulat
    Commented Dec 7, 2017 at 23:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.