Strong Color Cast Removal


Today’s Question: I remember some time ago seeing you demonstrate a technique for quickly removing a strong color cast from an old family photo using Photoshop. I now have some images that could use this correction, but I can’t find any information about it. Could you refresh my memory of the technique?

Tim’s Quick Answer: Yes! The technique you’re referring to involves using an inverted version of the image with the Average blur filter applied, combined with the Color blend mode and a reduced Opacity setting.

More Detail: I’ve often found that especially with faded old family photos, it can be very difficult to get good color using normal adjustment techniques. Fortunately, a quick technique can provide an excellent solution to balance the color for images with a very strong color cast.

Open the image in Photoshop, and create a copy of the Background image layer by dragging the thumbnail for that layer to the “Create a New Layer” button (the plus in a square icon) at the bottom of the Layers panel.

Next, choose Filter > Blur > Average from the menu, which will apply a blur to the Background Copy layer that is so strong that all pixels will have the value of the average color of all pixels. In other words, this layer will now represent the color cast for the image.

We want the opposite of the color cast in order to compensate for that color cast, so from the menu choose Image > Adjustments > Invert.

On the Layers panel click the popup at the top-left that shows a value of “Normal” by default, and choose “Color” from the popup. This is the blend mode, which affects how the pixels on the current layer interact with the layers below. In this case that will cause the underlying image to have a stronger color cast than it started with, but of the opposite color.

The final step is to reduce the strength of the correction. To do so reduce the value for the Opacity control at the top-right of the Layers panel. A value of around 50% should work well, though the optimal value will vary based on the original image. Note, by the way, that in many cases you will want to increase the saturation of the image, and of course possibly apply other adjustments to fine-tune the overall result.