I have been dreading being asked this question as no matter what one says, someone will tear you to pieces.
For what it's worth, these are my observations and practices. They are not absolutes as each image is slightly different. Use these as a guide, but engage your eyes and more importantly, your brain and adjust accordingly.
The flare I am discussing here always appears close to the centre of frame in the form of an out-of-focus iris-shaped area with reduced Dmax (maximum density) and a red (not magenta) colour cast. If you look closely, you will notice a second, smaller flare within the confines of the first. I have only ever seen it when using the 23mm and 28mm HR Digarons and Alpagons.
If the area and extent of the flare are not obvious, it can be helpful to dramatically steepen the curve in order to increase contrast and saturation, which makes it easier to identify the affected area. You are going to have to select it with your tool of choice, so you need to know where it is.
The next step is to select the larger flare area (we use the same method on the smaller area after the larger has been dealt with). I usually use the lasso or pen tool with a feather of approximately 77 pixels (IQ180 processed to 100%). The feather will vary according to shooting aperture and output pixel dimensions. You will have to work out your own. My figures are just to give you a starting point.
Once I have the selection, I bring up the curve tool and go to the red layer. You will always see that the histogram drops off well before Dmax (bottom left corner). Drag the bottom of the curve towards the drop off point of the histogram, but not all the way, whilst closely observing the image. You will need to stop just before the shadows start showing a Cyan cast. I usually find that the red curve will need to be dropped or 'bent' a little at the 3/4 tone as this seems to manifest the strongest red cast.
Once I've done this, I then go to the RGB curve and make much the same adjustment, moving all 3 channels together. The reason is that the flare will have scattered non-image-forming light (i.e. "flare"), into all 3 colour channels, reducing shadow density or 'clipping' the shadows. This density needs to be restored and I do this by moving the black point.
If you have done this right, you will have mostly eliminated the larger flare and will now have the smaller flare to deal with. I use the same method.
This is a global fix that works most of the time. I sometimes find I have to do a little local detailed colour management to tiny parts of the corrected area, or have to use the history brush a little because I wasn't as accurate with my selection as I should have been.
This may sound a little complex but with practice, it's pretty simple and the result is effective.
I cannot recall a single instance where I have spent more than 5 minutes on doing this.
As with any colour management work, it goes without saying that a calibrated workspace is essential as is working in 16bit colour mode.
As I've indicated at the beginning, this is how I deal with the problem. There are many other ways to skin the same cat.
Cheers,
