rake doesn't properly align the strokes #44604

New Issue

Nahuel Belich · 2015-05-04T19:36:58+02:00

Nahuel Belich commented

2015-05-04 19:36:58 +02:00

System Information
win 7 x64

Blender Version
Hash: 9715d4c

Short description of error

rake strokes doesn`t properly align one after an other this with normal space stroke but its really noticeable using curves to draw curve paths, as you can see in the image at the bottom follows a nice line but on the curves the rake stroke doesn't fully align with the path

Exact steps for others to reproduce the error
1-rakeissue.blend shoud be alredy on texture paint mode
2-aim on the plane align the curve on the plane
3-hit enter to draw

**System Information** win 7 x64 **Blender Version** Hash: 9715d4c **Short description of error** rake strokes doesn`t properly align one after an other ![Blender_2.jpg](https://archive.blender.org/developer/F170063/Blender_2.jpg) this with normal space stroke but its really noticeable using curves to draw curve paths, as you can see in the image at the bottom follows a nice line but on the curves the rake stroke doesn't fully align with the path **Exact steps for others to reproduce the error** 1-[rakeissue.blend](https://archive.blender.org/developer/F170066/rakeissue.blend) shoud be alredy on texture paint mode 2-aim on the plane align the curve on the plane 3-hit enter to draw

Nahuel Belich commented

2015-05-04 19:36:58 +02:00

Changed status to: 'Open'

Nahuel Belich commented

2015-05-04 19:36:58 +02:00

Added subscriber: @NahuelBelich

Eugenio Pignataro commented

2015-05-04 19:49:56 +02:00

Added subscriber: @EugenioPignataro

Eugenio Pignataro commented

2015-05-04 19:49:56 +02:00

it would be good to fix this not only for the curve stroke, also the other stroke modes(space, dot etc) will benefit of a fidelity control on the rake of the stroke direction

Bastien Montagne self-assigned this 2015-05-06 12:35:58 +02:00

Bastien Montagne commented

2015-05-06 12:35:58 +02:00

I’ll check this too…

Bastien Montagne commented

2015-05-06 13:16:02 +02:00

Added subscriber: @Psy-Fi

Bastien Montagne commented

2015-05-06 13:16:03 +02:00

Ok, so basically, issue here is we do not have access to tangent at the draw point, only to the line made by current point and a previous one. This is a reasonable approximation when draw points are close enough, and there is no singularity (or even sharp changes) in slope, but breaks in any other case.

Antony, I do not think we consider this a bug? sounds more like limitation of current design…

However, there is one thing that makes things even worse here - previous rake reference point is generated by averaging itself and current draw point (paint.c, line 552-553), this sounds very bad to me? Replacing

  interp_v2_v2v2(ups->last_rake, ups->last_rake, mouse_pos, u);  /* u == 0.5f */

by

  copy_v2_v2(ups->last_rake, mouse_pos);

Alleviates the issue, at least.

Ok, so basically, issue here is we do not have access to tangent at the draw point, only to the line made by current point and a previous one. This is a reasonable approximation when draw points are close enough, and there is no singularity (or even sharp changes) in slope, but breaks in any other case. Antony, I do not think we consider this a bug? sounds more like limitation of current design… However, there is one thing that makes things even worse here - previous rake reference point is generated by averaging itself and current draw point (paint.c, line 552-553), this sounds very bad to me? Replacing ``` interp_v2_v2v2(ups->last_rake, ups->last_rake, mouse_pos, u); /* u == 0.5f */ ``` by ``` copy_v2_v2(ups->last_rake, mouse_pos); ``` Alleviates the issue, at least.

blender-admin commented

2015-05-07 18:10:48 +02:00

This issue was referenced by a5dead2e8c

This issue was referenced by a5dead2e8cee27c33771f7d74020ef9e15beceee

Antonis Ryakiotakis commented

2015-05-07 18:11:20 +02:00

Changed status from 'Open' to: 'Resolved'

Antonis Ryakiotakis closed this issue

2015-05-07 18:11:20 +02:00

Antonis Ryakiotakis commented

2015-05-07 18:11:21 +02:00

Closed by commit a5dead2e8c.