[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #23 from Dmitry Kazakov --- Git commit ea04b8afef1631adcff5377653c154124909c2d2 by Dmitry Kazakov. Committed on 27/07/2020 at 10:36. Pushed by dkazakov into branch 'krita/4.3'. Refactor rounded division hack in KoMixColorsOpImpl to use templates M +20 -10 libs/pigment/KoMixColorsOpImpl.h https://invent.kde.org/graphics/krita/commit/ea04b8afef1631adcff5377653c154124909c2d2 -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #22 from Dmitry Kazakov --- Git commit 3967105ce4532bd8d7955566f5144bd336c5d414 by Dmitry Kazakov. Committed on 27/07/2020 at 09:32. Pushed by dkazakov into branch 'master'. Refactor rounded division hack in KoMixColorsOpImpl to use templates M +20 -10 libs/pigment/KoMixColorsOpImpl.h https://invent.kde.org/graphics/krita/commit/3967105ce4532bd8d7955566f5144bd336c5d414 -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 wolthera changed: What|Removed |Added Latest Commit|https://invent.kde.org/grap |https://invent.kde.org/grap |hics/krita/commit/f0a9a2e70 |hics/krita/commit/e49da79a6 |51a4f5ee213413ec5e7492ed7a1 |4ab02e6cf6db92f051c408b3837 |1923|48fc --- Comment #21 from wolthera --- Git commit e49da79a64ab02e6cf6db92f051c408b383748fc by Wolthera van Hövell tot Westerflier. Committed on 25/07/2020 at 10:55. Pushed by woltherav into branch 'krita/4.3'. Fix darkening of colors using the deform brush. Based of dmitry's initial patch, this adds a hack to ensure the appropriate division function is selected. We should be using KoColorSpaceMaths and templates for this, but it is not clear to me how this should be done, and I even suspect it requires a bigger refactor. This however is fast enough and makes the brush work as expected in 8bit integer and 32bit float versions of all the available color models. M +14 -2libs/pigment/KoMixColorsOpImpl.h https://invent.kde.org/graphics/krita/commit/e49da79a64ab02e6cf6db92f051c408b383748fc -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 wolthera changed: What|Removed |Added Resolution|--- |FIXED Status|ASSIGNED|RESOLVED Latest Commit|https://invent.kde.org/kde/ |https://invent.kde.org/grap |krita/commit/04ff8ad4cc2a97 |hics/krita/commit/f0a9a2e70 |6e80c6c5d5326e09a08a36d770 |51a4f5ee213413ec5e7492ed7a1 ||1923 --- Comment #20 from wolthera --- Git commit f0a9a2e7051a4f5ee213413ec5e7492ed7a11923 by Wolthera van Hövell tot Westerflier. Committed on 25/07/2020 at 10:54. Pushed by woltherav into branch 'master'. Fix darkening of colors using the deform brush. Based of dmitry's initial patch, this adds a hack to ensure the appropriate division function is selected. We should be using KoColorSpaceMaths and templates for this, but it is not clear to me how this should be done, and I even suspect it requires a bigger refactor. This however is fast enough and makes the brush work as expected in 8bit integer and 32bit float versions of all the available color models. M +14 -2libs/pigment/KoMixColorsOpImpl.h https://invent.kde.org/graphics/krita/commit/f0a9a2e7051a4f5ee213413ec5e7492ed7a11923 -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #19 from n-pigeon --- Yup! I've read about color shifts but it is still great improvement for me! It will allow me to use this tool where I couldn't before! :D When you fix color shifts it will expand use cases even more. But this fix is huge even as it is now thank you! -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 wolthera changed: What|Removed |Added CC||griffinval...@gmail.com Assignee|ghe...@gmail.com|griffinval...@gmail.com --- Comment #18 from wolthera --- Ah, before you heap endless praise on us, as the messages state, there's still a problem within 8bit spaces, the colors avarage towards being dark (with green-red mix ending up black). The two possibilities here is either to use the precisepaintdevice wrapper as we're doing with the colorsmudge (this however requires a bit of refactor and makes the brush noticebly slower), or we're going to try and fix the kocolormixop further. -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #17 from n-pigeon --- WOW Dimitri, Wolthera, you guys are amazing, thank you! -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #16 from wolthera --- Git commit 30cb1e7106dd3cc272fc68e4478fd136e2e6e204 by Wolthera van Hövell tot Westerflier, on behalf of Wolthera van Hövell. Committed on 24/07/2020 at 21:56. Pushed by woltherav into branch 'krita/4.3'. This fixes the bug where this brush slowly outputs everything to transparent. The bug was caused by the fact that to sample the input pixels, the crossdevice color picker was used, which in turn got access to subpixel data by utilizing the randomsubaccessor, which was getting that data by mixing neighbouring pixels via the KoColorMixOp. However, the randomsubaccessor isn't able to generate perfect sums-to-255 weights every single time, meaning transparency was added. This commit fixes that by passing a sum of weights to the KoColorMixOp that only defaults to 255. Special thanks to Dmitry for figuring this out. Note: there's still an issue with colors mixing darker in 8bit, we're still trying to figure that one out. But at the least there's no transparency. M +14 -2libs/image/kis_random_sub_accessor.cpp M +8-5libs/pigment/KoMixColorsOp.h M +8-7libs/pigment/KoMixColorsOpImpl.h https://invent.kde.org/graphics/krita/commit/30cb1e7106dd3cc272fc68e4478fd136e2e6e204 -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #15 from wolthera --- Git commit 5101cddfdd340d47503801944c6dfe7906bf088e by Wolthera van Hövell. Committed on 24/07/2020 at 21:54. Pushed by woltherav into branch 'master'. This fixes the bug where this brush slowly outputs everything to transparent. The bug was caused by the fact that to sample the input pixels, the crossdevice color picker was used, which in turn got access to subpixel data by utilizing the randomsubaccessor, which was getting that data by mixing neighbouring pixels via the KoColorMixOp. However, the randomsubaccessor isn't able to generate perfect sums-to-255 weights every single time, meaning transparency was added. This commit fixes that by passing a sum of weights to the KoColorMixOp that only defaults to 255. Special thanks to Dmitry for figuring this out. Note: there's still an issue with colors mixing darker in 8bit, we're still trying to figure that one out. But at the least there's no transparency. M +14 -2libs/image/kis_random_sub_accessor.cpp M +8-5libs/pigment/KoMixColorsOp.h M +8-7libs/pigment/KoMixColorsOpImpl.h https://invent.kde.org/graphics/krita/commit/5101cddfdd340d47503801944c6dfe7906bf088e -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #14 from wolthera --- Git commit 15b1ebf076f476b399161d5928419b7be106c568 by Wolthera van Hövell tot Westerflier. Committed on 24/07/2020 at 09:06. Pushed by woltherav into branch 'wolthera/BUG-290383-colormixopfix'. Fix deform brush by making KoColorMixOp handle non-255 weights. This fixes the bug where this brush slowly outputs everything to transparent. The bug was caused by the fact that to sample the input pixels, the crossdevice color picker was used, which in turn got access to subpixel data by utilizing the randomsubaccessor, which was getting that data by mixing neighbouring pixels via the KoColorMixOp. However, the randomsubaccessor isn't able to generate perfect sums-to-255 weights every single time, meaning transparency was added. This commit fixes that by making the weights in KoColorMixOp always use the sum of the weights as the normalization factor. Special thanks to Dmitry for figuring this out. M +8-4libs/pigment/KoMixColorsOpImpl.h https://invent.kde.org/graphics/krita/commit/15b1ebf076f476b399161d5928419b7be106c568 -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 Tymond changed: What|Removed |Added Status|REOPENED|ASSIGNED -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 vanyossi changed: What|Removed |Added CC||gbird...@gmail.com --- Comment #13 from vanyossi --- *** Bug 418271 has been marked as a duplicate of this bug. *** -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #12 from Boudewijn Rempt --- At this point in time, I don't think anyone knows. You could try to go through the git history and see when that was added, but you might need the calligra-history git repo to do that. If all else fails, we can copy your fix into the deform brush and make that the only user of this code. -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 --- Comment #11 from vanyossi --- The issue with deform is that it mixes the alpha channel of the layer and the dab, using the subpixel information as weight data. as far as I understand, when the mix is done in the deform brush case there is alpha at 100 in one channel which in repeated strokes causes the paint to go invisible. Is this intentional? -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 vanyossi changed: What|Removed |Added Status|RESOLVED|REOPENED Resolution|FIXED |--- --- Comment #10 from vanyossi --- had to revert the fix as it broke many other transformations. -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 Boudewijn Rempt changed: What|Removed |Added Latest Commit|https://invent.kde.org/kde/ |https://invent.kde.org/kde/ |krita/commit/bd9602f5229bad |krita/commit/04ff8ad4cc2a97 |d80cde58a89198d21e903fa388 |6e80c6c5d5326e09a08a36d770 --- Comment #9 from Boudewijn Rempt --- Git commit 04ff8ad4cc2a976e80c6c5d5326e09a08a36d770 by Boudewijn Rempt, on behalf of Ivan Yossi. Committed on 24/06/2019 at 08:21. Pushed by rempt into branch 'krita/4.2'. Avoid rounding coords for hue and value on bilinear sampleData M +33 -16 libs/image/kis_random_sub_accessor.cpp M +2-1libs/image/kis_random_sub_accessor.h https://invent.kde.org/kde/krita/commit/04ff8ad4cc2a976e80c6c5d5326e09a08a36d770 -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 vanyossi changed: What|Removed |Added Latest Commit||https://invent.kde.org/kde/ ||krita/commit/bd9602f5229bad ||d80cde58a89198d21e903fa388 Status|ASSIGNED|RESOLVED Resolution|--- |FIXED --- Comment #8 from vanyossi --- Git commit bd9602f5229badd80cde58a89198d21e903fa388 by Ivan Yossi. Committed on 24/06/2019 at 03:14. Pushed by ivany into branch 'master'. Avoid rounding coords for hue and value on bilinear sampleData M +33 -16 libs/image/kis_random_sub_accessor.cpp M +2-1libs/image/kis_random_sub_accessor.h https://invent.kde.org/kde/krita/commit/bd9602f5229badd80cde58a89198d21e903fa388 -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 Boudewijn Rempt changed: What|Removed |Added Status|CONFIRMED |ASSIGNED -- You are receiving this mail because: You are watching all bug changes.
[krita] [Bug 290383] Deform Brush leaves whitish or darkish traces in 16-bit, if scrub long enough on the gray field
https://bugs.kde.org/show_bug.cgi?id=290383 Boudewijn Rempt changed: What|Removed |Added Assignee|krita-bugs-n...@kde.org |ghe...@gmail.com -- You are receiving this mail because: You are watching all bug changes.
