Masks

Comments

10 comments

  • SFA

    If you have the same adjustments to apply to 2 masks  - or rather to 2 areas of the image - what stops you creating the masked area that you want.

    It doesn't sound like it is the sort of thing for which a radial mask is intended - much more like a free form creation.

     

    Edit: Typo correction.

    0
    Comment actions Permalink
  • Class A

    @SFA
    I don't understand your remark.

    The proposal makes a lot of sense.

    Imagine you have two spots in the image that require the same kind of adjustment (say brightening) and both spots are masked well by radial masks.

    Why should it not be possible to have two radial masks in one layer, so that their adjustments (e.g. brightness level) always stay linked?

    0
    Comment actions Permalink
  • SFA

    Class A,

    I could just about imagine what you have described but do you think it is a common use case?

    And if so do you think that applying and managing two (or more?) interacting radial masks make more sense than simply creating a mask as required by painting it?

    0
    Comment actions Permalink
  • Class A

    SFA,

    I'm sure I answered before, but somehow the post seems to have disappeared.

    The idea is to keep adjustments for two areas synchronised.

    Image there are two faces in an image that both need the same kind of adjustments (say brightening and a bit of colour correction). Further assume, that a radial mask is the best way to select an individual face (e.g., one can nicely control the fall-off and tweak the selection later on, if one does not 100% succeed at first).

    Now why shouldn't it be possible to have two radial masks in one layer so that whenever one tweaks the adjustments to one face (e.g., changes the colour correction), the same changes are applied to the other face?

    With the current "one radial mask per layer" restriction one has to either manually synchronise the adjustments between two layers (each containing a radial mask) or needs to brush the selections manually.

    In general, one layer should be able to feature a mix of radial masks, graduated linear masks, and brushed masks. There should be no need for "rasterisation" either. The latter is a concession to efficiency, not a necessity.

    0
    Comment actions Permalink
  • BeO

    Well, there should be the option to rasterize though. That is very useful if you start with a parametric mask, gradient or radial, and want to finetune this mask later with the brush, e.g. use a radial mask for a face as in your example and then erase it for the eyes.

    However, I also think that several parametric masks in one layer is a useful feature, just like several source points in a heal layer is a useful feature too.

    0
    Comment actions Permalink
  • Class A

    BeO,

    we agree that having several parametric masks in one layer is useful.

    On the subject of rasterisation I maintain it would be better if it could be avoided entirely:

    Fine-tuning of a parametric mask by manual brushing could be made possible by superimposing the parametric mask and the manually drawn mask. Technically, that would require the manual brush to remove a mask in place where it did not contribute before, i.e., where the only contribution to the overall mask comes from a parametric mask, but that would be solvable.

    Of course it is very possible that manual tweaks to a parametric masks would have to be redone anyhow, if the parametric mask is changed. However, that is not always the case. For instance, if the manual tweaking only occurs in areas where the parametric mask contributes with 0% or 100% then tweaking the parametric mask afterwards would often not affect the manual adjustments.

    The current approach that requires rasterisation forces one to always redo the manual tweaking, if one decides that the parametric part of the mask should be redone.

    0
    Comment actions Permalink
  • BeO

    This came to my mind too, and would be technically possible. However, it would either require another eraser type (which I know you don't like) working only in conjunction with parametric masks, or a change of the concept of the eraser to also allow for negative mask pixel values.

    Otherwise what would happen if you erase from a p.mask, then delete the p.mask and want to preserve the erased areas, e.g. for the next p.mask you like to draw, and how do you show this with the greyscale Alt+m? Blacker than black? :-)

    For my taste, even if you find answers to this, it's probably over-engineered, not intuitive, prone to future issues when maintaining the product, and too costly to implement in the first place. 

    Imo, not worth it. I'd rather like to have a parametric color range rectriction similar to luma range, or a better softproof with bpc checkbox, or something "fancy" like a out of gamut warning. 

     

    0
    Comment actions Permalink
  • Class A

    What is not intuitive about preserving manually painted mask modifications?
    Whether you add to a mask or take away from it, your modifications should just live on.

    In contrast, nothing is intuitive about "rasterisation". A non-technical person would rightfully question what that is about and whether it is really necessary to show the guts to the user.

    I understand why they chose the "rasterisation" route but I was disappointed with that design because it lacks elegance and negates the flexibility of parametric masks.

    0
    Comment actions Permalink
  • BeO

    You brush an area, then define a radial mask, then erase from the lower opacity area of that radial, then delete that radial again, then you end up with negative masked area from your eraser, which I do not find intuitive, or it's back to zero which means you lost your eraser action, not intuitive either. Give or take, there are advantages and disadvantages. Your milage may vary, and maybe I just don't have enough imagination for what you mean.

    regards

    0
    Comment actions Permalink
  • SFA

    Class A,

    In your example of using 2 (or more?) Radial masks (or another type of pre-shaped mask) on separate people in an image - why is it necessary to use a specifically shaped RADIAL mask?

    Using one for a whole image makes sense from an artistic control point of view. For the whole image area.

    I could see is for a portrait type image as used to be common a couple of generations ago when formal portraits were often matted for an oval visible image  - usually to hide some unpleasant out of focus edges and lighting issues prevalent in the technology of the time.

    I don't get it in the Instagram age when one can in  any case quite quickly and easily just create the dual or multi-purpose mask anyway.

    It seems like a push for unnecessary over complexity and too many ways of doing the same thing.

    Sometimes less is more.

    My first impression of C1, way back in V5, when I compared it to the 2 editors I was using at the time, one being LR, was that all I had to do was open a file to see better results than I could see doing the same thing with the others.

    I don't really want a tool that has a billion an done ways of giving me a result. I would much prefer a tool that I can simple use to open a batch of files and know it would give me, for most of them, a decent level of perfection with no further input from me.

    I appreciate that is unlikely to happen ... however the distraction of adding more and more tools for distinct purposes should not, on my opinion, be an objective; rather it should be something to be avoided if at all possible. As a matter of design principle.

     

    Still, those are just my thoughts. Actually probably closer to dreams as I doubt they will happen.

    The problem with simple things is that people rarely perceive their true and effective value. Unless that expensive  - but that's a different market motivation.

    0
    Comment actions Permalink

Please sign in to leave a comment.