Gimp-Forum.net
Custom Brush Edge Transparency Artefacts - Printable Version

+- Gimp-Forum.net (https://www.gimp-forum.net)
+-- Forum: GIMP (https://www.gimp-forum.net/Forum-GIMP)
+--- Forum: General questions (https://www.gimp-forum.net/Forum-General-questions)
+--- Thread: Custom Brush Edge Transparency Artefacts (/Thread-Custom-Brush-Edge-Transparency-Artefacts)

Pages: 1 2


RE: Custom Brush Edge Transparency Artefacts - rich2005 - 03-29-2020

Quote:...Do you think this is buggy behavior worth reporting? Search function on the gimp bug database is not working for me at the moment.

Up to you. You might get a response. Make sure you are using the latest / greatest Gimp (2.10.18 at the moment) otherwise you will be told to upgrade.

Most users probably never notice the odd pixels or if they do they are lost in a 'busy' background.

I can see from your .xcf that you are using Gimp 2.10 so you will have a clipboard mask. Try that instead of the clipboard brush, which leads to making a brush. A greyscale image will make a brush that takes the FG colour in use.

Quick demo here 2 minutes https://youtu.be/XhTVmd37_uU






RE: Custom Brush Edge Transparency Artefacts - PigletPants - 04-04-2020

rich2005 thank so much for your reply.   Big Grin

That video was really helpful, and now I understand the clipboard mask too! Definitely a viable solution to this issue.

Actually I have learnt a lot about what is going on under the hood so to speak by grappling with this issue so thanks again for your all your input.

I just realised that the black colour of the artefacts is coming from the base colour of the colour channels which is black but not visible since I had my images all transparent. If I flood the canvas with red, make the image transparent, then draw the number in red and copy that as a brush the issue is effectively fixed since the artefacts now would be sampled in the same colour and practically imperceptible.

The very small niggling issue that remains is why Gimp is so random in its sampling of the brush creating the small randomisations on the edge in the first placee. I would have thought that when the brush size is set Gimp would sample the brush to memory and use exactly the pixel data from that sample every time and therefore produce identical results every time when using mholder's suggestion earlier in the thread on snapping to grid with 1x1 cell size so effectively snapping to pixel.


I'm going to have a think about it and possibly create a report about it in the future. Cheers! Cool