Everyone,

Just tried the HDA method and it works like a dream!

The steps used were:

1. Wrapped the “global_parameters” null into a subnet.
   
2. Converted the subnet to an HDA with some meaningful label and saved the HDA to the scanned asset library (e.g. in C:\Users\<username>\houdini18.0\otls). N.B. no promotion of parameters is required.
   
3. In the existing projects, installed the new HDA and then replaced the old “global_parameters” null with the above HDA.
   
4. Used the opchange Hscript command to convert all references to “global_parameters” to (e.g.) “my_HDA/global_parameters”.
   

So now, whatever project I am working on, if I decide to change something globally, I just need to unlock the HDA, dive inside and change one or more settings, jump back out and perform “Save Node Type” and “Match Current Definition” on the HDA. The changes will be reflected across all the other projects.

Thanks to Tim and tamte specifically for the HDA suggestions/details and, once again, to everyone else for their excellent ideas!

All the best and keep safe,
Peter

Hi Tim,

Thanks for answering my questions. Agree that maybe a digital asset might be the way to go.

To everyone who has contributed to this thread so far - many thanks for your input and time - much appreciated.

Cheers,
Peter

Thanks Tim, that's another very interesting idea!

Apologies in advance, but being a complete novice when it come to Python and its use within Houdini, I have a few follow-up questions:

1. Using your method, would I still be able to retain the user-friendly functionality of the global parameters null UI to set/update values (there are a total of ~60 parameters with a ~50:50 split between colour information and other numeric values)?
   
2. Assuming that I can retain the UI functionality, then how would I export these parameters to the JSON file - either to begin with, or if some parameters subsequently get changed?
   
3. Where exactly/how would your Python code snippets be incorporated into the UI to replace the hard-coded values?
   

Cheers,
Peter

Many thanks for the quick responses - much appreciated. Will give them a try out.

Still interested in any other suggested solutions to the problem.

Cheers,
Peter

Hi there,

I am working on a video for a client comprising many separate, but inter-related, scenes. Each of these scenes, for workflow management reasons, are generated by discrete Houdini projects with frame sequences rendered out using Redshift. However, most of these projects use the same set of objects (many of which are instanced onto surfaces) that have a common set of parameters across all projects (e.g. colour, number of instances, object size, etc).

The issue that I have is how to efficiently update globally this common set of parameters across these multiple projects.

At the moment I have set up a null object with a parameter interface that defines these global parameters, which are then referenced by the various components comprising the given scene. I then copy this “global parameters” null from one Houdini project to another (each of which likewise have the same set of components reference the correct parameters). This obviously is very labour intensive and prone to error when dealing with a large number of Houdini projects.

Is there a means of defining these parameters in a project independent way, and then importing these parameter settings into each project automatically? I am sure there must be some sort of mechanism for doing this, but haven't hit upon it yet. Any suggestions would be very much appreciated.

FYI. I am using Houdini 18.0.566 with RS 3.0.28 on a Windows 10 workstation.

Thanks in advance.

Peter

Hi there,

Thanks for the link - much appreciated - some useful stuff there!

The volume gradient method almost works, except that the points on the sphere slide towards its upper pole as it descends. This is very similar to the results I was getting using the minpos in a popnet method (ref. https://www.youtube.com/watchv=Jgtzb3VBJTM).

Still open to further suggestions from people.

Cheers,
Peter

Hi there,

I've been trying to solve this problem (i.e. for a more complex scene showing a coronavirus merging with a host cell) for some time now, but with no success. Can any of you provide some help?

In its simplest form we have a sphere and a box (see attached hip file). The sphere is animated to pass into the top surface of the box. Both objects are converted to VDBs, VDB smoothed, then VDB converted back to surface(s). This results in the sphere descending and smoothly merging with the top surface of the box to create a single surface.

Points are then scattered onto the geometry (e.g. to be used later for instancing spike proteins onto the virus and other proteins on the host cell surface). The aim is to have all the points remain at fixed locations on the surfaces of the sphere and the box as they move towards each other and eventually merge.

The problem is that the scattered points change position from frame to frame. I believe this is because the resulting converted mesh(es) are constantly changing, which is forcing the scatter to re-assign the point positions.

I have tried various methods: timeshift + attributeinterpolate, or using a popnet to do effectively the same (as per this video https://www.youtube.com/watch?v=Jgtzb3VBJTM), but these methods don't seem to work as intended and the points still jump around on the surface(s).

Any assistance with this problem would be very much appreciated.

Thanks in advance.

Peter

Hi there,

First of all, forgive the rather lengthy explanation, but I want to make sure that the problem space is well defined, so please bear with me.

I need to build a 3D surface plot of some spectra data for a client. The spectra data will be supplied as float values in a CSV file and I will be (presumably) using the TableImport node to access this information.

The CSV file will be formatted as follows:

• The first column of data will represent points along the X-axis (i.e. energy).
  
• Many subsequent columns will each represent a set of Y-axis values (i.e. absorption) at a given temperature step. For example, there could be a 100 of these columns and I am assuming for the moment that the temperature steps will all be the same.
  

So, to summarise, the 3D plot will comprise energy along the X-axis, absorption up the Y-axis, and temperature along the Z-axis.

Using an example CSV data-set to experiment with, I have setup a TableImport node to read the X-axis values into one float attribute (energy), and the columns of Y-axis values into a single float attribute (absorption) which has an attribute length set equal to the number of temperature columns.

This data shows up nicely in the Geometry Spreadsheet and I can see the following columns listed and with the expected float values beneath:
“energy, absorption#0, absorption#1, absorption#2, absorption#3, absorption#4, …” where #-n represents the number in square brackets.

So far so good.

I then pipe the output of the TableImport node into a PointWrangle node with VEX code that should generate the 3D plot. For example, @P = set(@energy, @absorption, zoffset); .

However, when I try to treat the multiple @absorption values as an array of float values (e.g. using @absorption, where i = 0, 1, 2, …), the VEX compiler complains bitterly (e.g. “Ambiguous call to array index operator”). It seems that I can only access and plot the first column of absorption data against the energy (i.e. it will only recognise @absorption as a legitimate variable, and nothing else).

So there are two questions that I hope you might be able to answer for me:

• How do you access values of an attribute that is in the form of a float array?
  
• Is there a better method (Python?) that I could use to generate this 3D plot?
  

Once again, apologies for the lengthy explanation.

Thanks in advance,
Peter

Hi Michael,

Once again, many thanks for answering all my questions so quickly.

Regards,
Peter

Hi Michael,

Many thanks for the quick response and for the example - very much appreciated!

A few basic questions that I hope you could address:

1. What were the key steps that you went through to rig that geometry?
   
   
2. Whenever I have tried to use bones they ended up distorting the geometry - what did you do to prevent that happening?
   
   
3. Are there any gotchas that I should be aware of?
   
   
4. In your example rig, when the chain_goal is moved around, geo1 should only rotate around the vertical axis - instead it appears to still be rotating on all axes - what would need to be done to prevent this?
   

I agree with you that a tutorial on this subject is badly needed and would be a very useful resource for those of us who are not yet Houdini experts.

Thanks in advance,
Peter

Dear all,

I am currently working on a project for a major client that involves animating a multi-axis robot arm (see attached image) - the arm comprises several rigid parts with horizontal axis of rotation and a section attached to the base that has a vertical rotational axis.

I have now spent a few days hunting for information or tutorials that detail how to generate a setup to enable driving this rig from the tool-end of the arm using IK. Alas, I have had no success so far and have ended up more confused than at the start of the exercise.

As this is a time-critical project and seeming to have hit the proverbial brick wall, I am turning you all in the hope that you can advise me with either some clear step-by-step instructions or point me to some tutorial(s) on this topic.

Many thanks in advance.

Peter