Hello,

I am new to Houdini, coming from 3Ds.

I love Houdini but at times I struggle to do very simple things. One key difference I found is that Lights and Cameras cannot exist in the Geometry context, I call it Houdini Apartheid. In Max you work with lights and cameras in the same way as objects there is no separation.

As an example I have an undulating terrain, I have a point projected onto it using a Ray sop. I would like to copy its Y position to the Camera. I cannot figure a way to parent the camera to the point.

I assume the Houdini way is get the Y position of the point to the Camera. How to do this?

I cannot copy the parameter as a relative reference as its getting its position not from a transform but from the ray projection… I know this is super simple for any seasoned user but thats not me, I have searched for tutorials regarding how to reference an attribute from one object to another but I have not found anything. The documentation also does not clarify it enough for me.

In general I cant see why its not possible to create cameras inside the Geometry node, is there a good reason? Maybe I am simply not understanding the correct way to think about it but considering Housinis flexible approach this constraint seems out of character. Am I missing something?

Another example/question: If I wanted to make a lamp asset the light would need to be a separate node in the upper level linking to a point or a mesh, while the object itself could be deep inside geometry subnetworks, this separation is inconvenient. How could I make an easily reusable lamp asset? Would I have to make a reference parameter from the light to the node…? Are there any examples of this?

Any advice or thoughts are appreciated.

I feel like there is a moment where it will “click” for me, hopefully its soon.

Use a rivet object

have a look at the the attached file

Thank you Michael & Goat

Michael,

The camera rig solves the problem exactly, on the other hand the Lamp asset appears to reveal the limitation of the “apartheid”. Lets say I want to place this lamp and light on an apartment floor and instance the floor, it does not allow me to object merge the digital asset as it is a “invalid SOP”.


Now my workaround is either to create points with the lamp geo and pull them out after scattering the lamps and instance lights on OBJ level or pull out objects to make mesh lights. This just adds an extra laborious phase of work to the process that appears unnecessary.

I may be still in the 3DS MAx mindset and not getting it but what I would expect to be able to do would be to copy the asset onto a grid and the light goes along with it…

Sure - it's not as good as 3dsMax for this particular thing. Other weaknesses include; bones not at the geo level or bones aren't joints, or, modelling tools aren't as intuitive or featured, or Mantra is not GPU accelerated compared to redshift etc. There are many things that other software packages are better at.

Houdini gets easier when you think of it as part of a pipeline of tools. The good part is that Houdini keeps getting developed.. i.e. H17.5 will bring in Tops, which should be utilising your hardware better than ever, and H18 will bring in Lops, which could be a Hydra viewport and lighting improvements.

OK so its really a limitation.

It seems like quite a simple and fundamental thing so I thought perhaps I was missing something. Considering the awesome power of Houdini this is a bit baffling, there is so much to love.

Sure - I guess the first mistake people make is thinking Houdini is ‘all other apps++’ instead of a unique set of advantages and disadvantages.

goldfarb
have a look at the the attached file

thank you, so easy

The context separation is over, check out Solaris.