My guess, and it's only that, is the large amount of blue in the image causes any averaging process used by an auto white balance process to push toward yellow. How good are the results using the white point dropper on a neutral area?
This yellow tint appears in every photo I try to convert in ACR, not just this one.
But I had tried using the white point dropper method on the snow in this picture and the result is pretty close to the first picture, without showing any trace of tint.
I had never used ACR auto wb before, so I tried it on some of my raw files. On an image similar to yours, i.e. large expanse of blue sky, it did go yellowish. However on another image with a dominant red object the auto wb pushed it toward cyan. So I would conclude the ACR auto wb process is apparently not as reliable as the camera determination, and certainly inferior to direct measurement on a neutral object.
So I would conclude the ACR auto wb process is apparently not as reliable as the camera determination, and certainly inferior to direct measurement on a neutral object.
Oh, Auto WB is reliable for what it was designed to do...take the image colors into consideration when determining the auto white balance. It's doing exactly what it's supposted to do...the fact that it's often not what you want shouldn't be a surprise. The way a camera determines WB is by using a WB sensor (either the sensor itself or an external sensor) which is much more accurate than measuring image colors after capture. I've seen Auto WB help in certain cases where there simply is not white neutral in which to measure but I've found manually adjusting WB the best solution in most cases.
Everyone works differently, but as a workaround to the incompatibility between Camera Raw's Auto white balance and your expectations, I suggest this:
Set your defaults up to use the "As Shot" parameter - that would be the white balance determined by the camera.
If it's not right, i.e., you have a color cast you don't like, use the white balance dropper tool and set it yourself.
That's what I do.
Jeff Schewe wrote:
The way a camera determines WB is by using a WB sensor (either the sensor itself or an external sensor) which is much more accurate than measuring image colors after capture.
Yes, some more expensive cameras have external WB sensors. Other don't and are using the same raw data from the same main sensor as ACR, just algorithms for determining WB are better than in ACR
I don't use WB from the camera or ACR anyway. I always set the same manual WB for the whole batch of photos taken under similar lighting conditions, then correct it manually for some photos if needed (for instance if taken in shadows etc). It's giving more consistent results for me