I select background mask after finishing foreground AI mask. Then choose gradient. Define it. Then note it also involves the foreground mask. Not what I wanted. How to prevent this?
PL 9. Win 11. GeForce RX 3080. Latest driver.
If I’m reading this correctly, try starting a new mask for the gradient instead of applying the gradient in the first mask.
The first mask might be an animal. then I create a new mask (2nd) for background, select the gradient symbol, and draw it across the image. It covers everything in its path - including the first mask animal.
I think its easier to work the other way, start with the background and work forward:
-
First I use a Gradient mask (Mask1) which by default will include the balloon:
-
Then hit the Add Sub Mask option to add an AI mask for the subject (in this case the flying balloon):
- Then select the Balloon mask and hit the Invert Shape button, this will remove the balloon from the Gradient mask:
-
Create a New Mask by pressing the Add New Mask button:
-
Select the Balloon mask again and hit the Duplicate Mask button:
-
Select the Balloon Copy mask and drag it under the Mask2 heading:
-
Select the Balloon Copy mask and hit the Invert Shape button to revert it to a normal mask:
You now have TWO masks, Mask1 is the gradient (excluding the Balloon) and Mask2 is the Balloon.
Changes to Mask1 will not impact the Balloon and changes to Mask2 won’t impact the gradient.
15 Likes
@LVS has explained the process very well.
Don’t feel at all like you’re alone in having to figure this out. I think we all did.
The one thing I will add is that I frequently see a bug where clicking on a mask in the list does not always make that mask active. I don’t think I’ve ever had the “Add sub-mask” button go onto the wrong mask, but I have certainly found that clicking on a mask and then making adjustments often ends up adjusting the one I had previously selected. Not the newly selected one.
I have to say this is a very clear demonstration. Thank you so much.
A lot of us are stumbling in the dark with this stuff and you have shone a little more light.
1 Like
Much gratitude for the effort you put in to showing the forum users how to accomplish this edit. I imagine many DXO users would want to do this, but there is no method (that I have found) put out by DXO to do it.
Have to say though, if this is the easiest way, can’t imagine how tedious the hard way is!
It does seem a little tedious, but their methodology does provide for a lot of flexibility in how editing can be done.
The good news is that if the main subject is recognized by the AI Subject detection then it is possible to create a Preset (or Copy/Paste Adjustments) from the first edit that can then be applied with one click to multiple images.
This allows the rough edits to applied quickly and then the user can fine-tune the adjustments as required.