Convert to Curves inaccuracy

[bpedit]

Here's the mesh I'm using on a bunch of parallel lines, attempting to simulate water for an engraving project:

 

When I exit Point mode I get this, looking good for my first trial anyway:

 

But when I Convert to Curves, I get this:

 

Ooooh, ick! What's going on? Strangely, if I Export the object, with the unconverted mesh intact, as an SVG, the results are closer to the desired result but not quite as desired.

[bpedit]

After further testing I infer this has to do with AD deciding it can accommodate some of the induced curves by adjusting handles on existing points versus inserting new points. When the mesh is converted to curves the new points are respected but the adjusted handles are not.

Could a moderator please move this over to bugs.

[debraspicher]

Can you share the design you made (pre-curves conversion)? The inaccuracies are unfortunately well known and documented considerably. I wanted it to test it on the desktop version though and see if they vary at all.

[bpedit]

Thanks, disappointed to hear. I'm not sure if this file is the exact same as pictured above, I've gone through a few iterations. I tried it on desktop as well, maybe should have posted there.

waterTst.afdesign

One work around I can think of is to introduce nodes in the lines before applying the Warp. There appears to be no way to do that in AD except point-by-point. Maybe I'll try recreating the grid by duplicating a single line with the multiple points already extant.

Edit: When I remade my grid based on a line with multiple midpoints then applied the Mesh I got satisfactory results. Thanks for helping me realize I needed a work-around versus something I was doing wrong.

[debraspicher]

I downloaded and tried this in Designer 2.0.4 (Win10). You can click the images and check the results below:

SVG export seems to work fine: waterTst-original-scale.svg I checked the result Inkscape... I can open it or paste from Inkscape back into Designer. It doesn't appear to have flaws compared to the original warp group, but I haven't been looking at this as much as you might have already...

[bpedit]

Thanks. I will have to go that route. My initial test with the multi-point lines was positive but I can't repeat it. I'm thinking it has something to do with the frequency of the Mesh nodes compared to the frequency of the line nodes. I need to export to SVG anyway for use in the CNC mill so I'll just leave the Mesh unconverted. The result of this method isn't exact, in my analysis, but likely close enough. Now for some testing on the machine. Thanks again.

Edit: The exported SVG of the unconverted Mesh is very close, maybe exact, to the appearance of the Mesh before Converting to Curves. This would seem a good hint as to how to remedy the issue. 

[Dan C]

Hi bpedit,

I've confirmed this issue and completed a bug report.

Lee

[bpedit]

Still a bug! I forgot about this issue until I tried again today.