Jump to content

nabsltd

Members
  • Posts

    16
  • Joined

  • Last visited

Posts posted by nabsltd

  1. Since there is no section specifically for issues with the forum, posting here is my only choice.

     

    After about an hour, my session with the forum logs out, even though I have "remember me" set on login.  Then, I can't log back in, as the password doesn't seem to work, and I have to use the "reset".  This has now happened 4 times, so I know that not every time was a password typo, especially since I copy/pasted into the "password recovery" page, and then the same thing into the login page.

     

    The best bet for any forum admin is to send an e-mail to my registered address, because I likely won't see any replies to this or PMs in a timely manner.

  2. Both the very good "Outline Object" plugins (BoltBait's and pyrochild's) round the corners on the outline.  In most cases, this isn't a big deal, but sometimes I want the sharp corners.

     

    In the uploaded image, the left star has BoltBait's Outline Object (v4.2) applied twice with default settings, while the right star is created using the built-in star shape three times, each one slightly larger than the last.  Notice how sharp the corners are compared to the outlined version?

     

    The issue, of course, is with arbitrary shapes.  Outline Object is the only easy way to get an outline around the edges, but on some outlines, I really want the same sharp corner look as on the right-hand star. I've also tried duplicating the shape on a new layer in the color I want and using "1px Expansion" repeatedly, but that also gives me rounding after about the 3rd application.

     

    Is there any add-on or technique I can use to get the result I want?

    Outline Object.png

  3. 10 hours ago, Rick Brewster said:

    Crop to Selection will indeed "erase" the pixels that are outside of the selection. That's 100% intentional.

     

    Why is it intentional?

     

    The "crop" function on pretty much every other paint program just crops, since it's just one more keystroke (delete) before the crop to do the same thing that Paint.NET is doing on "Crop to Selection".  By deleting before the crop, you don't give the user a choice.

     

    Also, is it only supposed to delete pixels on the selected layer, or on all layers?  Because it deletes on all layers, which isn't explained in the help file.

     
    7 hours ago, MJW said:

    A while ago, I wrote a plugin that may help, called Unselected Rectangle Keeper. For your example,  you'd select the background with the Magic Wand, then run Unselected Rectangle Keeper to delete everything outside the bounding rectangle,. You then can select the transparent region, invert the selection, and crop.

     

    Thanks...I'll check it out, although I really shouldn't need to resort to a plug-in to do what other programs do by default.

     

    Tested it, but it still uses "Crop to Selection" at the end, which means that all the pixels outside the selection on all the other layers are deleted.

     

    Since I often have layers with single color fill (to test transparency), I then have to go back and re-fill all of them if I want to do any more editing.

  4. To reproduce:

     

    1. New image

    2. Create a new layer.

    3. Draw a filled circle on "Layer 2" (keep it in the middle of the image) in a contrasting color from the "Background" layer.

    4. Choose magic wand tool, set the Tolerance to 0%, and click anywhere in "Layer 2" except on the circle.

    5. Ctrl+I to invert the selection.

    6. Image->Crop to Selection (Ctrl+Shift+X).

     

    The expected result is a square image with a colored circle on a background, but you can see that the unselected pixels in all layers have been erased before the crop is applied.  "Crop to Selection" should not erase pixels...it should only change the canvas size to match the selection.

     

  5. On 10/28/2007 at 2:36 PM, Hellfire010 said:

    Set tolerance to 0% and use the magic wand on a transparent pixel. Invert selection, then crop to selection.

    Yes, I know this is an old thread, but I also want the requested feature, and this advice doesn't work, as it gives jaggies if you have multiple layers that create the transparency you want.

     

    The problem is because the "Crop to Selection" doesn't create a rectangle of the resulting canvas size first, then crop to it, but instead deletes pixels on all layers and then does the canvas size change.

  6. On 4/27/2018 at 7:58 PM, Rick Brewster said:

    You're requesting that I remove something from the history which actually does -- by definition -- need to be there. Selections don't need to be there because, by themselves, they aren't actually part of the image. It's a workflow design choice.

     

    The trouble with layer visibility is that it is tied to the active layer, but the active layer isn't part of the undo.

     

    Here's an example:

     

    1. In a drawing with at least two layers, both visible, make the first layer active.

    2. Select an area, and hit "Delete".

    3. Make the current layer invisible...this also moves the active layer to the other layer.

    4. Undo twice...you now have the original drawing, with a selection.

     

    At this point, "Delete" or "Redo" should have the exact same effect, but they don't, because you are now on a different layer.

     

    My point is that there is already a lot of inconsistency in the redo stack.  Some are changes to the drawing, some are for "workflow" (as you put it).  So, making some changes to it that are optional and not the default that make the workflow better in some ways should at least get some consideration.

  7. 20 minutes ago, Rick Brewster said:

    The visibility flag is part of the image, it can't just be removed from the history stack.

     

    Well, it certainly could be removed if you wanted it to be, since there are a lot of things on the undo stack that aren't part of the image (like a selection that had nothing done with it).

     

    That's kind of the problem with the undo.  For example, since setting a primary/secondary color isn't "part of the image", it isn't part of undo, but accidentally changing the color without noticing (stray click on the palette) is far more damaging than hiding/showing a layer. Likewise, having to change the "Tolerance" for the magic wand, and then backing out with undo beyond a paint bucket that used exactly the right tolerance, and then any action that goes on the undo stack will make it impossible to exactly redo that paint bucket.  And, changing the active layer isn't part of the undo stack, but I suspect most people would agree that this is far more important with respect to editing the image than changing layer visibility.

     

    Adding an undo stack for tool config would help with this, but there would still be a lot of stuff on the undo stack that has almost zero value, like most selections.  One thing I seem to want to do a lot is select an area on one layer, delete it, and then select the exact same spot on another layer and delete it.  The fact that the selections are on the undo stack makes me try to find a way to make this work, but there is no way.  Sure, I could use the select/new "template" layer/activate layer/disable anti-aliasing/paint bucket/enable anti-aliasing to get the selection, then magic wand/change layer/delete/change layer to "template" and repeat this for all the layers method, but...wow, that's painful when layer multi-select or being able to "redo" an arbitrary item on the undo stack (like the selection) would be far easier.

  8. I often turn layers on or off to change the background color so I can see how a transparent image will look with different backgrounds.  If I switch this a lot, it really clogs up the undo, and makes "undo/redo" (for "check by blink" of changes) very painful.

     

    Since changing a layer's visibility is completely non-destructive, and can never be done en masse (which might be hard to manually undo), it probably shouldn't be part of the undo stack at all.  But, it's there, and many people might like it.  I don't, so I'd like an option that can be enabled that keeps it from being added to the undo stack.

  9. Absolute awesome, except for one thing...the light source is always from the front.

     

    Mirroring and moving the shadow after it has been created can make it look OK, but it's not like the current exact render.  An addition of a control for the "ground" level (in pixels/percent from the bottom, with the default at zero, like now), and extending the angles to allow lighting from the back would be great.

  10. OK, I've figured out how to do this easily.

     

    1. Start with the object layer.

    2. Duplicate it twice.

    3. Create the dark shadow on the second duplicate using BoltBait's Object Shadow.

    4. Create a smaller white shadow on the first duplicate.  Now, the picture looks right, how a disconnected shadow would appear on a white background.

    5. Duplicate each of the shadow layers, and then merge them together, with white on top of dark.

    6. Select all on this combined shadow layer, and copy to clipboard.

    7. Move to the dark shadow layer, and use the Paste Alpha plugin with alpha source set to "Clipboard shades of gray", "Invert Calculation" checked, and paste method of "Minimum of alpha and clipboard alpha".

    8. Delete the white shadow layer and the combined shadow layer.

    9. Now, the disconnected shadow is separated by transparency, instead of white.

  11. 4 hours ago, BoltBait said:

    EDIT: This is just a quick hack, so it could probably be improved.  Feedback?

     

    Add in the ability to move the shadow and keep shadow only (I like to keep shadows on a separate layer so I can go back and change later), and it definitely solves my problem.  

     

    Perhaps it could just be merged into your original Object Shadow as an optional parameter for how fat the "disconnect" gap is.  I like the fact that you have added finer control for the movement.  I ran into some issues where a single step on the current one got rounded down, and the next step was too much.

     

    It still doesn't answer my more general question of how to do the masking, and I'd really like to know the general solution.

  12. The title isn't good, but I really don't know how to describe what I want to do.

     

    Basically, I have an object (text letters) that I want to create an "engraved" effect, using a separated drop shadow.  Here's a sample image:Engraving.png.c69997f1706bc93878c1fb57bac1209c.png  This was created by having 3 layers:

    the blue text: Text.png.fdbb50785eb25d2a864294b5f3814ebc.png

    a small gray (RGB 200,200,200) shadow (using the "Object Shadow" plugin) with Blending mode "Additive": Mask.png.b0de1a562dc5bca1942c33a9c2304554.png

    and a larger gray shadow: Shadow.png.7c889ddb7c1b379fc273f7a4e2381a5d.png

    This looks good on a white background, but I want the effect to work regardless of the background, so that the pixels that overlap between the mask and the shadow become transparent.  In other words, what's green in this picture should have no color, just alpha information: Transparent.png.fb18ae7ea90eb6069d03488e8ce98cbf.png

     

    How can I do this?

  13. I'm perfectly fine with the current default save strategy on new documents (PDN for layered, PNG for single layer), but by the time I get to the save as PNG, I almost always want the smallest possible, so I use the OptiPNG plugin.

     

    But, there is no way to set it as the default way to save PNG files.  The trick which allows setting JPEG as the default save format won't work, because opening a PNG file results in it being saved using the built-in PNG options.

     

    Any ideas on how to make it work, or is is finally time to allow people to set the default save format, and just let them shoot themselves in the foot if they want to by saving in a lossy format.

×
×
  • Create New...