Mike A

The only thing I would check: is that 1,000w power supply sufficient? I realise the A6000's pull significantly less power than the 3090's, but with all the RAM, SSD's etc - it's worth double checking. Better to have headroom than not - and it's a minor cost in a very expensive system.

MOI 3D is a nurbs based modeler. It's written by the guy that originally created 'Rhino 3D'. You might think of it as a much simpler, but still very powerful, version of Rhino or Fusion 360. It also has a great polygon mesher so you can export the nurbs based models as polys with a lot of control. I use it quite a bit for CAD / engineering type modelling / visualisation work. There are many users on the forum creating models for 3D printing very successfully. There's an active and very helpful forum at: https://moi3d.com/forum/ Looking for a quick demo of the sort of work you're doing, here is a video from 2015 - and it's been much improved since then - but it gives you an idea of the sort of thing it can do very easily - and this video is probably realtime - or close to it - BTW : ) Here's another good intro:

A modeling job like this just screams MOI3D to me TBH.

Could you post an image or two so we get an idea of the object form?

This is similar to your original source, but if wonder if it offers any further insight? https://cp-algorithms.com/geometry/tangents-to-two-circles.html

Just tagging onto this thread as i'd be interested in seeing any solution. Currently I'd do this in MOI http://moi3d.com which has some great tech drawing tools including this find common tangents feature : )

I'm really glad to see you sharing these and posting your initial thoughts. I know you've been waiting on this for a very long time, and I hope the new RW meets your needs. The results above look good to my eyes, and a massive improvement on the pre-RW Redshift version.

It's great that Side fx are putting up more training material, but I have to say I think these are poor. Robert Magee might be a great product manager - but he's not a good trainer IMHO. These are typical of a person who is an expert in their field and has forgotten that a new user knows nothing. The first one is nether a high level overview of H, nor detailed enough to be a simple tutorial on copy to points and basic RBD. He tries to do both, and succeeds in nether. Constantly changing tools without saying anything and usng vague statements. I think if you were brand new to H you wouldn't be able to follow that at all.

The issue is 'ramp tiling'. There is some technical issue that means that it isn't currently possible to tile a ramp. I'm not sure why - you might want to ask the Redshift team on their forum. The reason it works on the old Xpresso system is that the C4D shader is actually 'baked to a texture' in the node system and that texture is tiled.

Got it - my misunderstanding.

Unfortunately the c4D shader node is not available in the new C4D 'nodes' system. However, it is still available in the older Xpresso based 'shader graph' system. I don't think there is any issue mixing materials - so you could still re-use the older material?

Very clever!

I'm not 100% sure what you want, but maybe polyexpand2D would give you more control, eg: ...but this sort of stuff is C4D's specialty, and a weak area in Houdini - so I'd be doing this in C4D myself.

It's ironic that the MOUNTAIN node is probably the easiest node to use 🤣

One of the 'minor' features of 19.5 which really caught my eye was the 'Kelvin wake' deformer. Kelvin wakes being the very characteristic standing wave pattern created by ships. Something I've needed to do more than once, eg: below: This was done with displacement maps - and what a pain and effort it was. I've just had a vey quick play with the new deformer and it looks fantastic - and so much easier : )

The promise of traditional C4D ease of use plus the power of a sophisticated nodal system, all running on a much faster architecture was a very attractive one to me. But, like others, I got tired of waiting - and am currently learning Houdini. I think Maxon is in a hard place over this. When you consider the traditional strength of C4D has been 'ease of use' it's hardly surprising that the current bucketful of low level nodes isn't #1 on the average user's go-to list. I imagine Capsules were proposed as a solution to that - but they don't deliver on speed. C4D doesn't currently have the depth of nodes, the promised speed gains, or an interested user base. I expect Maxon is 'working on it', but it certainly seems very quiet out there. Having said this, Project Neutron was announced exactly two years ago - 22 July '20. That's not much time to deliver on a project of this magnitude. In many respects that's the issue. Houdini has a 20 year head start, Blender has an interested user base, but C4D hasn't yet offered anything to compete.

Yes - had a quick play with the dashbox. Looks like it will be helpful. And that new helix SOP is quite something : )

Eikon... OK. I just read your msg above, as I was about to post the attached file. I'll look into the SOPs you mention. Having spent a lifetime working with 'objects' it's taking a while for me to fully get my head around the fact that there is really no such thing in H, it's just a pile of points that might have something in common - or not : ) delete by connectivity.hiplc

Just an experiment and wondering the best way to work with this... A box, booleaned into three parts. I've used an explodedview SOP here to separate them slightly. I'd like to break this into three geo 'streams' for different processing. Currently I'm using three blast SOPs, and manually selecting each part to make a group for that part. Is there a better way?

Check you've not got duplicate points on those edges. Mesh > Optimize to weld any co-located points.

Two RFE's submitted submitted and added to the wish list : #122904 : Multi-snap priority settings to be added to global prefs. #122905 : Add 'snap to centroid' as an option in multi-snap settings. I've got to say that SideFX support are super fast. I've contacted them twice over the last couple of months and I've had an almost immediate response on both occasions.

Yes - that's a little Xpresso setup I made for camera culling. I typically use it to cull matrices - as shown, then Redshift has a workflow to clone mesh data - or whatever - onto the matrices. It's the most efficient way to scatter with C4D + Redshift. I have scattered a few million tree instances with it, but as you say C4D isn't the king when it comes to scattering large numbers of anything.

Cheers Eikon... very useful info as always. I'll put in an RFE for the snapping - and hopefully @Igorcan second it : )

In Redshift look up the OSL (Open Shading Language) shader node. You can download text based OSL shader files to plug into that node here: https://github.com/redshift3d/RedshiftOSLShaders One of them is the jiWindowBox shader. That's what you want. I know there were some issues with it last time I used it - quite some time ago, but it may have been tweaked / fixed since then. There was a discussion about it earlier last year - and I posted an example - here:

Hi guys, I'd like to set up my multi-snap (Ctrl Shift J) priorities as global defaults, so they are set every time I create a new H project. So in a wider context I'm asking how do I set up a default project similar to C4D's 'new.c4d' file?