thanks Dave

@pbacot said:

I have another. For working by sections. This requires placing cross sections (all from resizing one shape) and skinning the result. You can reshape the cross sections before skinning to account for shallows etc. Actual placement: might be good to draw lines across first to decide where sections go. Question is what gives a model that can look like real river conditions. This one comes out looking like a skate park.

Nice! curviloft seems to do a better job than sandbox. I think one of the key criteria for a river is to have shallow inside bends and deeper outside bends, and of course some natural variability! So my original example trying to extrude a standard cross-section probably wasn't the right way to start! Cheers - Mick

Magic mate!!! Drop that autocad and you'll never look back.....!

Gilles thanks for the tip.

TIG thanks for yet another great plug-in - it worked exactly as advertised.

In my situation it works as a great cleanup tool. If the smooving is done with appropriate care to get it approximately right, when Drop Vertices is applied, the smooth curves created by smoove aren't disrupted to0 much and it comes out nice.

thankyou

@pbacot said:

You can do this. Before pasting in the closed shape to the grid, soften the edges in the grid by selecting all of them and clicking soften coplanar in the smooth/soften window. After you paste in the shape, you'll be able to select the outside and erase it.

Ahhh.. yes of course, I have seen that before - my brain just didn't want to recognise the solution.

@pbacot said:

I don't know that you can lock the outside edges in this process. I guess you just have to avoid them. Anyone?...

Maybe... and probably not too hard to do but it would be good to know if there is a technique to lock the shoreline in place.

Thanks

thankyou TIG/mitcorb,

I turned all layers on and unhid all then zoomed to extents. I had a left behind a line at the original coordinates about 7000km away from the origin. I investigated one of my previous models that had the same issue and it was the same thing (a hidden component that I found in the outliner). Once I'd dealt with those, all was good.

Thanks!

thanks Mac1, it certainly is good practice to take a disk image of a working system - something I used to do, but complacency has stepped in over the last year or 2. One utility I've read about recently is "sandboxie" which I think may really suit my situation where I'm often trying out all sorts of software/downloads, etc. Cheers - Mick

Gilles thanks heaps - I knew they'd be a way !

and Dave R thanks for pointing out the subtleties that are needed to make it work.

@ Pilou - this is definitely a trippy plugin, I couldn't get it to do what I needed, but I got some interesting results with randomising the components.

Regards

thankyou Giilles - that worked for me. Cheers - Mick

Hi Jean, the example file is overly simplified, but demonstrated my issue, but maybe I over-simplified it. Simply moving the face is possible, but in my situation I had an irregular shape with faces aligned on many different XY angles and I wanted to be able to selectively push-pull them out on the horizontal. As it turned out I needed to use a combination of your technique and smart-push pull to get the shape I wanted. Thanks for your reply - Mick

Gaieus/Bjornkn, I'd forgotten about flip edges. Thanks heaps for your time and patience!

OK, I've got my mesh. 😍 The original cad contours had 1.3 million edges.

Thanks guys for your suggestions... like I said - it's been a great learning experience.

TIG - I experimented more with the simply contour plugin - most of the problem were with my contour set - some were ludicrously convoluted whereas most were quite simple. There was no one setting that did a good job with all of them, but with some selective processing of different sets of contours based on complexity, it worked well.

Ogan - I tried tgi3d amorph. Haven't had time to explore it all properly (hardly at all actually - there seems a lot too it), but I ran through your suggested method. I had a similar experience here to that described above. Setting a 20m segment length worked for some of the contours, whereas other lengths were required to get anything acceptable for other segments. In other words I had to do a lot of selective processing using different settings to get a good overall result.

In the end I used "Edge Tools" to simplify all the curves to within 2m of the original, then went and cleaned up the overlapping contours with reference to the original dwg. At the same time I deleted any extraneous noise. This all took about 1hr, but wasn't too onerous. Then I just used the sandbox tools to create a mesh from contours. The result is shown in the attached. This was definitely the quickest way to process the entire contour set in 1 go and came up with an acceptable result. I think my method posted at the beginning of this thread gave a really nice result, but was too labour intensive for anyone dealing with a lot of lines, but otherwise I'd recommend it.

mesh.jpg

thanks again for your help with this...

Regards

Jean - thankyou that's a nice way to go about it.

Zapwizard - that's a good suggestion using TIG's drape control points plug in - I'll experiment a bit with this. I had a very quick look at Artisan, but need to research it a bit more... but from what I can see, it is possible to manually add or simply detail in the mesh. It would be good however if there was a tool that could analyse the vertical complexity in the xyz data so that points that were similar to their neighbours would be tagged as redundant, so that an optimal irregular mesh could developed. Does such a plug in exist? Does Artisan do this?

Thanks for your replies so far

@sketchymick said:

I'm pretty confident that you'll be spot on with this, and I'm expecting that there will be 11.7 degrees between the north/south shadow and my compass north. One thing I can say is that my prediction is probably more accurate than my ability to measure it. 😉

Well I measured it today and I got about 12.5 degrees - which is good enough for me given the error is about as much as the thickness of the graduations on my compass. So Jim, I reckon your method works a treat. 👍