![]() Registered Member ![]()
|
Win 7, NVIDIA GT 730. Color picker's "internal Krita cursor" GETS OFFSET after 10 seconds of zooming and panning: then always erroneously picks the wrong color.
CONFIRMED "W" Tiled Mode is the cause = color picker gets off sync, but pressing "W" again and switching back to single tile mode the color picker works again. But I need seamless tiled mode of course, so please fix. Thank You! Reproduce: Open 32x32 pixel image. Zoom in, pick color and paint, zoom out to check how the seamless tiled pattern looks, zoom in to paint, PAN here and there, pick a color, paint, zoom out, repeat x3 times and after couple seconds doing this, try to zoom in and pick a color with Freehand Brush's alternate function: the ALT+LMB Canvas Input Alternate Invocation, Mouse Button, ALT + LMB, pick foreground color.... Doesn't matter which PICK option you configure there from where to pick, the "internal Krita cursor screen position probably tied to the OpenGL texture + screen surface" slide apart so they "loose focus" resulting in the color picker internal focus also SLIDES into OFFSET so you'll have to search for the correct color with the bugged color picker. While painting 32x32 pixel seamless tiles, zooming in to pick a color and paint with the Freehand brush tool then zooming out works for 10 seconds, then the color picker goes OFFSET and only gives the correct color, if I move the cursor +1 or +2 pixels in any direction - trying here and there -, mainly NorthWest, North, NorthEast AWAY FROM the EXACT pixel screen position I want to pick. I guess the opengl texture coordinates - stored in memory - erroneously SLIDE to OFFSET after zooming and panning. Cannot work fast this way, when I have to slowly guide the color picker to pick up the correct color. Its amazing, how it works for like 10 seconds, then looses accuracy after zoom + pan. Color picker works always in Photoshop CC, so it's not a driver issue. Probably how Krita handles texture read/write operations during and after zoom in / zoom out / pan: what is shown on the screen is offset from what coordinate the internal texture is stored at.
Last edited by sogun on Fri Jun 09, 2017 7:29 pm, edited 1 time in total.
|
![]() Registered Member ![]()
|
I reset the NVIDIA driver to default settings, no anisotropic filtering, no trilinear filtering, no antialiasing..
It seems, when hitting the canvas wall during zooming in, the "camera bounces off" the wall allowing zooming only within the tiled texture canvas area and the coordinates for the cursor of the color picker seem to go offset, lose sync = what pixel the picker is getting is out of sync with the texture 'pixel' shown on screen. Cause also could be when accidentally zooming out way too much so that one tile on the 32x32 pixel tiled canvas becomes really small, then maybe when zooming back in Krita to 100% and much closer, the program somehow loses its sync between what canvas image is shown on the screen and the real texture coordinates held in memory... The color picker picks a color, but that is no longer in sync with what is shown on screen. |
![]() KDE Developer ![]()
|
Can you make a short video or a screenshot and link that here? It's a bit difficult to decipher what you are saying, and that may not be your fault so much as that talking about offsets gets real complicated real quickly...
(Also, which version of Krita is this?) |
![]() Registered Member ![]()
|
Which version of Krita is this?
Krita 3.1.3 and 3.1.4 File-----> New -----> Create a 32x32 pixel image to do "seamless pixel graphics tiles" for small games. Watch the bug appearing as I pan the viewport to the edge of the workspace: the picked color goes instantly "offset". Zooming and panning away from the edge back toward the center of the workspace again seems to IN RARE CASES reset the color picker so it picks the right color again or pressing W rarely makes the bug go away . If gone, the bug always pops up again: this way its impossible to maintain a creative, efficient workflow. I'm constantly fighting the bug. Video demonstration: https://youtu.be/ch4bcIkw71o |
![]() Registered Member ![]()
|
NOT Fixed in Krita 3.2.
![]() I guess the problem requires serious thinking and a skillful person is not available to do that. ![]() |
![]() Registered Member ![]()
|
If you found a bug, please report it on the bug tracker (https://bugs.kde.org/).
|
Registered users: Bing [Bot], Google [Bot], kde-naveen, Sogou [Bot]