Found 103 posts.
Search results Show results as topic list.
Day 7. Definitely getting to understand kinefx and animation editor better.
Day 5 - Houdini was also crashing as crazy
Day 4 - Escape. We are the rabbits!
(kinefx, darn difficult to animate, but possible!)
(kinefx, darn difficult to animate, but possible!)
I guess your question is not super clear, but I will try If you need something very simple, like just have a grid, or rows and columns, and you know the size of your area, the grid is just simple, no need for two nodes. I don't think this is going away, the grid node is super useful.
Edit parameter interface - find your ramp parameter, click on the button autoscope next to the Default Points setting. If someone has more elegant solution, be my guest
Edit parameter interface and changing the max number of the select input parameter to something higher than 10 also helps.
It is about the distribution of the randomness, and you can tweak that. So if you imagine that you have object 0, 1, 2 and then you decide that you want more of the object 2, then make sure that your randomness is driven by the ramp and set your ramp which is from 0-1 to favor higher number by setting the ramp to reach the higher number sooner. is quite simple to figure this out. As you already probably know how to scatter different objects on the heighField, you should plug the attribute random node just after your scattered points, before the copy to points node, and you can tweak your randomness this way, just make sure that you create the point attribute with the same name as you defined in name attribute for your packed prims. Maybe there is also something in heightfieldscatter, but you should check the documentation for that, I don't believe seeing such a thing, however they do have this new node, attribute from pieces, which is designed for exactly that: https://www.sidefx.com/docs/houdini/nodes/sop/attribfrompieces.html [www.sidefx.com]
davide445If you would use a real street data, first of all, you are talking about some sort of 3D scan? Then photogrammetry is your friend. There are free/commercial solutions which Houdini supports to make a geometry out of photos.
Looking to use Building Generator, but using real "street view" kind of data as souce.
Defining manually different buildings style is time consuming and also will never replicate all the real buildings on a street.
Using Google Street View data is not allowed for offline usage.
As middle ground I was thinking if it's possible to extract from self-produced recorded 360 videos basic info about the buildings (prevalent color, position of the windows, number of floors) and next use these abstract data to generate the buildings.
Want to ask if someone ever faced this kind of problem.
But if you think there would be an algorithm to extract that data and magically recreate that modules based on your 360 image, technology is not there yet. We would need to train a lot of ML nodes for that to work, but I am sure there must be a paper somewhere or someone is working on it already and it will become available soon. Until then, the solutions are, hire professional 3D scanning crew and bring them in your neighborhood and then hire few more modelers to clean your geometry in a way that they remodel all the modules and texture them for the game use, for example. And this is probably too much for some indie artist pocket.
Just a note, if you plan to scan your neighborhood you better ask around if you need to acquire some permit from the local government, it is basically the same as you would shoot for the movie. So be aware of possible legal constrains as well.
RobotHeadArtThis elegance is just beautiful!
SideFX has added new tree generation tools to the latest release of SideFX labs tools (18.5.424).Image Not Found
They have quite a lot of features and options for generation and seem to be straight forward to use. Thank you for releasing these!
The best solution for me so far - save to disk in background. It prompts you to save the last version and when you save and press the button again, it fires the new instance of Houdini, so you can still load your geometry frame by frame with the file node and look in Houdini, while it is caching the geometry in the background. For example, this way I was able to stop the simulation once I found a pleasing looking volume on my disk, so instead of extra hour of waiting for the cache, I can decide if I can stop the process earlier.
I too use controllers from object context to drive my kine FX RIG. I am sure new additions will be added to KineFX and SOP context in the future, and especially to the rigpose node
If the controllers from obj level are not causing the same slow down as tons of CHOPS, that is already a step up from what we had. And adding the intuitive approach for retargeting made it all worthy, maybe not specifically for the character animation but definitely ads more freedom in procedural workflows.
Why don't you RFE? Is it not possible to reproduce?
KyleMartinVFXDo you mean like in the mplay for the rendering? Maybe try loading and refreshing the geo seq in the gplay? I am not sure if there is a built in solution for that.
Hello, while caching to disk using ROP geometry output or File Cache, is there a way to see the playback when you hit save to disk? I want to see the sim as it goes rather than only seeing the progress bar.
Congratulations on the investment, SideFX, and I hope this will help to hire more developers to deliver more excellent engineering to already outstanding tools you are delivering. It is great that the investment comes from another inventive company and the true power of synergy between passionate users, developers and management from both companies. I am very happy because of this and looking forward how this will effect the audiovisual industry in near future.
I am not sure what I am looking at here, is your camera in the ROP? Can this work?
- Quick Links