CopyToPoints variation with changing point count

[scorpes]

Hi, I would like to know how I can make variation (different objects) using the copytopoints node and changing point counts. I now use the foreach loop with metadata and I use the iteration value to add randomness to a switch node.

Like this tutorial:

But when point counts change, this method doesn't work anymore and all objects are constantly changing over time. 

Any idea how to fix this?

 

copytopoint_var_changingPointcount.hip

 

[Atom]

That is normal, if you deform or change the surface you are scattering the points onto, it's similar to changing the seed on the scatter. You will get a different result every time the surface changes. Read the help card on the PointDeform node to lock down the point set before you CopyToPoints. There are a lot of example on this forum, try the search bar.

Edited January 22, 2020 by Atom

[scorpes]

But there is not something like make the copytopoint randomizations based on @id, instead of @ptnum? My initial point source is a pop network with changing point counts

[Jason]

Look into using the CopyStamp methodology where you can control the variation with a Switch SOP containing the randomization method you choose.   You could choose something based on @id, for sure.

[scorpes]

CopyStamp is the old method..., not compile-able , there is no way to do this with the new method?

Edited January 22, 2020 by scorpes

[markinglevfx]

You can use this code in your switch. Currently you're using the loop iteration in the switch, but the point/loop order changes as new points come in so it's different each frame. By referencing each point's id attribute when selecting the object it instances you can keep them constant throughout. Hope it this makes sense.

Adjust the max fit value (currently 1) to account for more instance objects

fit01(rand(point("../foreach_begin1/", 0, "id", 0)), 0, 1)

 

Edited January 22, 2020 by markingleukc

[scorpes]

yes this is what I was looking for, thanks!

1 hour ago, markingleukc said:

You can use this code in your switch. Currently you're using the loop iteration in the switch, but the point/loop order changes as new points come in so it's different each frame. By referencing each point's id attribute when selecting the object it instances you can keep them constant throughout. Hope it this makes sense.

Adjust the max fit value (currently 1) to account for more instance objects

fit01(rand(point("../foreach_begin1/", 0, "id", 0)), 0, 1)

 

 

[anim]

1 hour ago, scorpes said:

CopyStamp is the old method..., not compile-able , there is no way to do this with the new method?

whatever Mark said since you are using 17.5

however technically the new method since H18 is using Piece Attribute to copy variants to points and avoiding for loop altogether

[Iskander]

2 hours ago, anim said:

whatever Mark said since you are using 17.5

however technically the new method since H18 is using Piece Attribute to copy variants to points and avoiding for loop altogether

Yep and i think it is even faster than compiled for-loops.

[Iskander]

4 hours ago, scorpes said:

CopyStamp is the old method..., not compile-able , there is no way to do this with the new method?

Check this.
+now it is rather simple

 

copy2points_newpipeline.hipnc

Edited January 22, 2020 by Iskander