One of the issues I encountered processing my photos from the Avila Cathedral is the color cast from the different lighting sources inside the cathedral. The interior is generally very dark with timed spotlights installed throughout that are activated by buttons at the stations. Most of these are incandescent with a few LEDs mixed in, providing generally very warm lighting that contrasts with the much colder sunlight coming in through the high windows. The camera tends to pick something in the middle for a white balance resulting in yellow casts on the spotlighted interior features and cyan or blue casts where sunlight is reflected. With all of the different structural elements and textures found within, it is very difficult to remove these color casts.
I processed several photos using radial filters in Lightroom to reduce or eliminate the color casts. This process can be tedious but works rather well. Greg Benz published a short tutorial video on using color-based luminosity masks (color masks?) to quickly remove color casts from interior photos, so I decided to give it a try on this image that I had already processed in Lightroom but had not attempted to remove the strong blue cast from the left side of the photo. In the end, I got very good results using this technique, but it was a bit more work than I expected.
I shot 5-exposure brackets for all of my images in the cathedral using a default exposure of 1/40-second at ISO 400. With my 12 mm manual lens, I was able to set the focus, aperture, and exposure with the brackets, then I didn’t have to worry about camera settings and could concentrate on composition and exploring the cathedral. (For an in-depth explanation of these settings, take a look at my previous post on capturing an interior panorama of the central nave of the cathedral.) For many of the scenes inside the cathedral, 12 mm was not wide enough, so I was capturing a lot of 2+ shot panoramas that necessitated using a fixed exposure setting on the camera. In addition, the light varied wildly from one shot to the next, so it was much easier to have the camera setting fixed. Any images that included the high windows that were brightly lit by the sun required multiple exposures anyway, so bracketing made sense.
Of the 5 bracketed exposures, the brightest was slightly blurry and not usable. Even shooting at 12 mm with in-body image stabilization on the Sony a6500, the 0.4-second exposure was a little bit too long. The middle exposure at 1/10-second was very sharp, so the 4 darker exposures were combined using Lightroom’s merge HDR process. After merging the exposures, the image was brightened 3 stops to produce the image below.
In capturing the images of this cathedral, I also found a very practical use for ISO invariance. The base exposure for my brackets (1/40th sec, f/5.6 at ISO 400) is underexposed by about 3 stops for this image of the choir. Since I was using a 2-stop bracket interval, I picked 1/40 sec as the base exposure to allow me to get a sharp image at 2 stops longer exposure time (1/10th sec). Any longer would risk getting a blurry image (as evidenced by the brighter exposures at 0.4 sec which I had to discard). So I should have been capturing these images using a 3-stop higher ISO (ISO 3200) which certainly would have resulted in the loss of highlight detail. However, based on my previous testing of ISO invariance of the a6500, I knew that I could simply set the exposure to a “safe” base exposure and make whatever exposure adjustments needed in Lightroom. And since HDR processing inherently reduces noise, the final image is almost free of noise, even in the deep shadows.
In addition to bringing up the exposure to account for under-exposing the base image, other edits in Lightroom included strong adjustments to highlights (-71) and shadows (+80), sharpening, slight noise reduction (+10 on the luminance slider), and liberal use of the Transform sliders to straighten the image and reduce some of the perspective distortions.
The pre-processed image from Lightroom was opened as a Smart Object in Photoshop, then the layer was duplicated using “New Smart Object via Copy” to create an independent copy. I opened the copy in Camera Raw and white balanced on one of the large organ pipes on the left side of the image. This copy was very warm and far too red, but the blue cast was removed.
Back in Photoshop, I generated a mask on the original layer using the Colour selection tool in Raya Pro and clicking on one of the darker blue organ pipes. (In his example, Greg Benz was using Lumenzia. I’m sure other Photoshop panels have similar functionality.) I had to expand the mask, then manually painted in several areas of the organ and stonework. Without these manual adjustments, the combined exposure was very flat and almost devoid of all color in the gilded woodwork around the organ. This manual editing of the mask was the most tedious part of the process, but I found it to be necessary to achieve acceptable results. I’m sure this process could have been easier with better use of the color masking tools.
After blending, the image was cropped to a 16:10 aspect ratio. I created a selection of the lower right foreground of the choir then used Transform->Warp to correct the perspective in that corner. I finished the adjustments in Photoshop with some color adjustments including a slight yellow Color Fill layer using the Color Dodge blend mode to brighter the image and a red to yellow Gradient Map to provide some additional warming and color toning. I used BlendIf on both of these color adjustment layers to protect the highlights from becoming too bright and to keep the red cast out of the shadows. (I learned these useful color adjustments from Blake Rudis at f.64 Academy.)
For comparison, here’s the same image processed completely in Lightroom. The blue cast has been reduced but not eliminated, particularly on the organ and stonework. When I review the two images, I still feel like I’ve lost something in the gilded woodwork around the organ pipes compared to the Lightroom image, but the blue cast ruins the image. And I like the overall warmer tone of the Photoshop-processed image. If you’ve used this technique for removing color casts and have any advice, I’d love to hear from you.