behaviour of "squished bakery"

Hello everyone,

I'm trying to flatten a cutting pattern, that I want to generate with GH. The tricky part are the dynamic seam holes, which I believe are destroying my "clean" srf.

In the attached images you can see a comparison of the original (clean, left) surface without stitch holes. I then process it in GH to get the holes (right). Unfortunately, the resulting (baked) surface does not behave the same way as the original does. As you can see below, it's all wrinkled up (red arrow).

1. I'm pretty new to GH and Rhino, so creating the holes in this way might be a non-standard approach to start with? Or does it seem legit this way?
(I went through some tutorials [THANKS!!], but I'm not yet familiar with any best practice)

2. could someone please give me a hint, how i could tackle this? Obviously I need the same clean squish (green arrow), only WITH the holes...

3. Currently I'm working on a MBair i5 (2core) inside Parallels and the computational drag for this little problem feels already quite significant (~10-20 secs to compute or so). Would you consider this normal for GH, or have I gone down a resource intensive path?

here's a screenshot of my GH-file:

Any hint is very well appreciated,

thanks,

vincent

Attachments:

Screen Shot 2016-06-19 at 18.07.50.png, 123 KB

Replies to This Discussion

Permalink Reply by Armin Seltz on August 5, 2016 at 5:31am

Could you not "cut" the holes into the flattened surface?

It almost looks like its converting it into a mesh when flattening with the holes.

Can you attach the .gh and rhino file?

Regarding performance: This is quite normal, especially if you are using Parallels. If its not worth using Bootcamp, then you will just have to live with the waiting time. But yeah, its totally normal for some stuff to take long, going up to a few minutes for heavy stuff.

Permalink Reply by vincent on August 8, 2016 at 4:40am

Hey Armin,

thanks for the reply. well, do you mean i should flatten the srf BEFORE i cut it? i'll have to work with that idea in the evening, at work right now. is there a theoretical reason for doing so?

the files are attached

thanks!

ps: no bootcamp if possible. waiting for the Mac WIP to properly handle GH ("explicit history")

Attachments:

grasshopperCUT-02.1.gh, 12 KB
grasshopperCUT-02.3dm, 3 MB

Permalink Reply by Armin Seltz on August 8, 2016 at 6:28am

There is absolutely no reason for doing that other than trying to get it to work. If the original surface flattens without problem, but once the holes are cut into it does not, then I wonder if it needs to be cut in the unflattened state in the first place. You can cut the holes in the flattened surface just as well.

Let me check the file and see if I can figure out whats going wrong.

As for the Mac WIP, sure, I am hoping for that too. As their is no plugin support for GH on the Mac yet and probably wont be till Rhino 6/GH 2, there is no way round GH on Windows right now though (at least for me, as I use a lot of plugins for GH). I have Bootcamp solely for Rhino/GH, but then I have an iMac with a large HD. I can understand not wanting it on a Macbook Air with limited HD space.

Permalink Reply by Armin Seltz on August 8, 2016 at 7:24am

Ok, so I checked your file. I am not sure what command you used to unroll your surface, because it is obviously not done in GH, but in Rhino.

I tried using UnrollSrf, but it told me that the surface is doubly curved und therefore not unrollable (developable). I then used UnrollSrfUV, which worked and also lets you unroll any curve/edge you select. I did that with the surface and the bottom edge and then used those in your GH file, which actually worked straight away and is much faster (by the way, your original GH file took 14 seconds to calculate on a 2013 iMac with and i7 and Windows 7 using Bootcamp - so your times using Parallels are not so bad).

I have made a few other improvements to your GH script. I have internalised the data in the curve and surface input, which means they are not referenced from Rhino any more. But you can set them again the usual way by right-clicking and setting one geometry.

If you bake the output right at the end of the GH file, you get a clean surface in Rhino.

Attachments:

grasshopperCUT02.1_improved01.gh, 28 KB

Permalink Reply by vincent on August 8, 2016 at 2:54pm

whoo-hoo, i believe you must have solved the next question, that i didn't even ask yet: "how can i make GH squish the damn thing for me"

won't be able to look through it for the next 2-3 days though, but that looks supergreat at first glance! thank you so much!

laater,

vincent

Permalink Reply by Armin Seltz on August 9, 2016 at 5:47am

I wish that was the case. But as I mentioned I did the flattening in Rhino and then simply used the unrolled geometry and unrolled curve in GH. But I am still wondering how you got that flattened piece, because UnrollSrf doesn't work (because of double curvature) and UnrollSrfUV results in a different shape.

I have never done any unrolling in GH, I am sure there are plugins for it.

Permalink Reply by vincent on August 10, 2016 at 10:12am

I used the squish command, hence 'squished bakery' ;)

the squish has some neat parameters to ensure the material will only be compressed or stretched, which fit my usecase very well.

I found out that one can get (some?) rhino-commands to work in GH via a plugin called lunchbox, but I'm not there yet. first the seams need to work properly.

I haven't checked this, but "they say", this might work: https://tinyurl.com/zcc9cz2

Permalink Reply by Armin Seltz on August 10, 2016 at 10:22am

Ok, yeah that looks pretty good. It seems like they had figured out how to do the squish in GH. So in that case all this shouldnt be a problem, because you can just use the squish to get the nice flattened surface and put the holes in those the way I have done. Then you can also use the current version where you put the holes in the unsquished model to use that as well (for rendering for example).

I will try it out tomorrow.

RSS