I really like the new local adjustments but they don’t follow rotation or perspective corrections, and that’s more than annoying, especially, if you don’t know about this.
Attached you will see what I mean. I masked the big chain in the foreground but upon rotating the image (just as an example, it’s the same for other geometric changes) the mask is not congruent with the formerly masked object any longer.
This behavior limits your own work flow a lot. In addition, I would like to change the color of the masks, sometimes the contrast to the picture background is not very good with the blue one.
I also ran into this problem. After having done all (read: ALL) corrections and having spent some time on drawing the perfect mask for an object, I noticed that the whole image was not exactly horizontal. It was below 1°, but the effect on the masked object was horrible. After that, I always check H/V aligment first.
As this happens way too easy, I support this request.
(Even if I acknowledge your current approach and its limitations.)
I can’t speak for anyone else, but I always perform perspective changes very early on in my process flow and well before I apply any local adjustments. While I appreciate the frustration this may have caused for some people, making sure that perspective, distortion, and image rotation modifications are applied before applying local masks seems like a fairly minor change to one’s workflow.
I agree one has not to change one’s workflow a lot, and normally, I also correct perspective and so on early in the process, but sometimes you want to optimize things, that’s fairly common. And in these cases even slight perspective changes make the set masks useless and deliver strange effects in the photo.
I just don’t want to be forced by DxO PhotoLab to do the whole masking process again when I suddenly feel I could tweak the perspective just a bit.
Is it possible to select a repaired spot and move it away or change its size, like it is possible with local adjustments?
The difficulty I see is, that if you apply global transformations on local adjustment vertices, they could be moved into each other, so that one local adjustment point covers the other and it is not possible to select the covered one. To make this work, one would need new functionality to select and move these covered adjustment points.
In addition, if you apply global transformations to local adjustments, the local adjustments change their shape. So a U-point circle is projected as tilted oval and must be adjusted in ‘3D’ with the global transformation in mind, while others remain 2D, because they were created after the last global transformation was applied.
Yes, but problems have to be solved in general. The worst case is the important one, because no one will write code to handle special cases, where the tool behaves differently in dependency of the currently applied global transformation. If you would buy ViewPoint next month, you would expect it to work with every perspective. This includes the rotation around all three axes and also for tilt and flip.
i can’t vote but yes, it would be a good thing that the mask is pinned on the spot and that this layer is also adjusted in shape and size when the image is shifted in shape and size.
it is easy to forget something and a readjust for final view can ruïn you local correction efforts.