Here is another idea I had: It uses stamping of the copy operator to subdivide each primitive based on its area. The SOPs are commented, so I think the file should be pretty self explanatory. One drawback of this method is, that the resulting mesh is not watertight anymore (T-vertices are introduced)
remesh.hipnc.gz [www9.informatik.uni-erlangen.de]
Found 38 posts.
Search results Show results as topic list.
Technical Discussion » remeshing
- FrankFirsching
- 46 posts
- Offline
Technical Discussion » remeshing
- FrankFirsching
- 46 posts
- Offline
I don't know if this is really what you are thinking of: Perhaps you can use the divide-SOP with the “Bricker Polygons” option? This might generate edge strips, that run parallel with a very short distance to each other (e.g. a bricker line next to an already axis aligned edge strip within the mesh). These artifacts could be removed by a following fuse-SOP.
Hope that helps…
Hope that helps…
Technical Discussion » Exploding Car
- FrankFirsching
- 46 posts
- Offline
Perhaps I can help (it was me, who animated that lego short)
First, The otl doesn't support the crash, that you see in the video. Therefor I took the points, that describe the positions of the lego stones (I used instances to render them) and converteted them into a particle system. Then when Santa was falling down, he took a force field with him, that blasted the stones away. They collided with the floor and an interaction-POP kept them from penetrating themself. I don't know how your car does look like. Perhaps you will need a better collission detection, which would compilicate some things. Also you mention, that you have already built the car. It could help, if you describe, how you have done that. I ask only, because if you want to explode the car, you need the stones separated from each other. They need to move independently.
First, The otl doesn't support the crash, that you see in the video. Therefor I took the points, that describe the positions of the lego stones (I used instances to render them) and converteted them into a particle system. Then when Santa was falling down, he took a force field with him, that blasted the stones away. They collided with the floor and an interaction-POP kept them from penetrating themself. I don't know how your car does look like. Perhaps you will need a better collission detection, which would compilicate some things. Also you mention, that you have already built the car. It could help, if you describe, how you have done that. I ask only, because if you want to explode the car, you need the stones separated from each other. They need to move independently.
Technical Discussion » Insert Isoparm
- FrankFirsching
- 46 posts
- Offline
SuvoThere is nothing wrong with it. Maya is just using degrees to specify the smoothness NURBS. Remember: order = degree+1
…
But there are interesting thing. If you model with NURBS in Houdini using order 3, like in Maya, then you'll get smooth problems in render.
order 4 is Ok.
…
It all has to do with polynomials. The degree is the number of the highest exponent. The order is the number of addends in a full term. E.g.:
f(x) = 3x³+5x²+2x+7
Is a polynom of degree 3 and order 4.
But remember
g(x) = 9*x³+6x
is also an order 4 curve, because you can rewite it to
g(x) = 9x³+0x²+6x+0
So Houdini order 4 curves/surfaces are the same as Maya degree 3 curves/surfaces.
I hope you enjoyed this little math lectue :-)
Technical Discussion » Houdini 6.1.200 won't run under Win2K (SP4)
- FrankFirsching
- 46 posts
- Offline
Just wanted to say, that I have the same problem here. Since version 6.0 Houdini doesn't work anymore using Win2k (SP3), Geforce FX 5600. Houdini 5.5 is still working fluently (having it installed in parallel). Fortunately Houdini 6.x is running under Linux without problems, so this wasn't bugging me very much :-)
btw: In Windows these crashes are often at times, when a new window should pop up (e.g. mplay, when starting a rendering)
btw: In Windows these crashes are often at times, when a new window should pop up (e.g. mplay, when starting a rendering)
Technical Discussion » Cg fragment/vertex shaders in Houdini
- FrankFirsching
- 46 posts
- Offline
craiglhoffmanThis won't happen in the next years. The graphics-cards break in with their performance, when it comes to high-precission (32-bit-floating point) calculations. I know this, because I have implemented a fluid-solver on a GeforceFX using nVidia's Cg during my diploma-thesis. The performance of the CPU-variant was better than the optimized GPU-system. A friend of mine wrote a path-tracer. He implemented the same renderer for the GPU and for Pentium4 using SSE. He also couldn't verify, that the GPU is faster than a CPU. Even worse was the performance of a hybrid-tracer, which should utilize the best of both worlds. But the communication between the CPU and GPU was such a great bottleneck, that this version showed the worst performance of all three variants.
…
But then again, who needs render farms if the card speeds your rendering up so darn much?
…
Technical Discussion » light rays
- FrankFirsching
- 46 posts
- Offline
I have no experience with atmospheres, but your problem sounds like your shadow shader is not set to “Filter Shadow”. Perhaps it is set to “Fast Shadow”, which won't lookup any transparency information, but will terminate already, if there is somewhere between the shaded point and the light-source a piece of geometry. The “Filter Shadow” is slower, because it will have to determine all intersections of the shadow-ray with the scene and evaluate the corresponding colors.
Houdini Lounge » Random Seed Value in a Fractal SOP
- FrankFirsching
- 46 posts
- Offline
Hello!
You have to activate the stamping of parameters. This is done in the Stamp-tab of the copy-SOP. The check-box “Stamp inputs”, needs to be active for this, else your copies won't vary. In Var1 you can name the new parameter (e.g. BoardNr in your case). In Param1 you define the value of the new parameter (in your case $PT, the point-number of the template). In the upstream-graph of the copy-SOP you can acces the new parameter by param(“BoardNr”,1). As you might expect, the first argument specifies the parameter, which value should be read. The second is a default value for the parameter. Simply put this expression into the seed of the fractal-SOP, then you should get what you want.
Frank
You have to activate the stamping of parameters. This is done in the Stamp-tab of the copy-SOP. The check-box “Stamp inputs”, needs to be active for this, else your copies won't vary. In Var1 you can name the new parameter (e.g. BoardNr in your case). In Param1 you define the value of the new parameter (in your case $PT, the point-number of the template). In the upstream-graph of the copy-SOP you can acces the new parameter by param(“BoardNr”,1). As you might expect, the first argument specifies the parameter, which value should be read. The second is a default value for the parameter. Simply put this expression into the seed of the fractal-SOP, then you should get what you want.
Frank
Technical Discussion » TK TCL
- FrankFirsching
- 46 posts
- Offline
Hello!
tk is the ToolKit for tcl and yes it is “only” a bunch of UI extensions. If you type tcl in the textport (normally with an argument specifying a script), Houdini assumes, that you want to execute a script and halts its execution and waits til the end of the process. This is also called modal execution. In the case of the tk command houdini assumes, that the script you give him builds up a UI and the script is not executed in an atomic way. This is non-modal execution, where the script and Houdini execute in parallel. For example the script builds up a window with a button and the user clicks on it. Now the script sends a message to Houdini to execute some commands.
You can compare the tcl and tk commands with starting programs in a Unix-shell with or without ‘&’. To sum up, the tcl command waits until the end of the executed script, while the tk command continues execution of Houdini.
Hope this was not too confusing.
Frank
tk is the ToolKit for tcl and yes it is “only” a bunch of UI extensions. If you type tcl in the textport (normally with an argument specifying a script), Houdini assumes, that you want to execute a script and halts its execution and waits til the end of the process. This is also called modal execution. In the case of the tk command houdini assumes, that the script you give him builds up a UI and the script is not executed in an atomic way. This is non-modal execution, where the script and Houdini execute in parallel. For example the script builds up a window with a button and the user clicks on it. Now the script sends a message to Houdini to execute some commands.
You can compare the tcl and tk commands with starting programs in a Unix-shell with or without ‘&’. To sum up, the tcl command waits until the end of the executed script, while the tk command continues execution of Houdini.
Hope this was not too confusing.
Frank
Technical Discussion » error in meta-boundry test
- FrankFirsching
- 46 posts
- Offline
arctorYes you're right. These settings definively don't exceed the condition of the 30 nested branches. So theoretically it should prune the correct way. Now I don't know any further, why this happens. Perhaps there's really a bug in the L-System-SOP?
… but if I set the generations to 5 and increase the step size to 0.2 and middle click again the Lsystem String is much smaller… >30 for sure…..?? but it still prunes incorrectly…there is something I'm not getting here…
Technical Discussion » error in meta-boundry test
- FrankFirsching
- 46 posts
- Offline
Your L-system has another problem:
Houdini allows only 30 nested branches. Your rules create a much more nested string in relation to the square brackets .
Here is a scene, which demonstrates this:
http://wwwcip.informatik.uni-erlangen.de/~sifrfirs/PruneGrowingTotallyOut.hip [wwwcip.informatik.uni-erlangen.de]
Houdini allows only 30 nested branches. Your rules create a much more nested string in relation to the square brackets .
Here is a scene, which demonstrates this:
http://wwwcip.informatik.uni-erlangen.de/~sifrfirs/PruneGrowingTotallyOut.hip [wwwcip.informatik.uni-erlangen.de]
Technical Discussion » error in meta-boundry test
- FrankFirsching
- 46 posts
- Offline
Hi!
With the L-System-SOP is everything all right. It's your rules, which make it branch outside of the control-volume. In your rule1 simlpy erase the first ‘F’ and double your step-size. The reason for branching outside, is, that the point, which is tested to lie inside the volume, really lies inside it, but with your rule1, you step outside and then you branch. As noted above, if you delete the first ‘F’, the test is made at the branching point and all will be ok.
Frank
P.S.: There exists at least a second person, who likes and uses L-Systems (me!)
With the L-System-SOP is everything all right. It's your rules, which make it branch outside of the control-volume. In your rule1 simlpy erase the first ‘F’ and double your step-size. The reason for branching outside, is, that the point, which is tested to lie inside the volume, really lies inside it, but with your rule1, you step outside and then you branch. As noted above, if you delete the first ‘F’, the test is made at the branching point and all will be ok.
Frank
P.S.: There exists at least a second person, who likes and uses L-Systems (me!)
Houdini Lounge » Cutting a hole in a nurbs circle that has been extruded.
- FrankFirsching
- 46 posts
- Offline
I have tried it one more time, and I see what your problem is. I have created a small scene, which you can look at. It is a rather unconventional approach, but I tried to comment all I have done. You can middle-click on all SOPs, which are displayed in blue and Houdini will show a comment, I have written, at the bottom of the pop-up. Here is the scene file:
http://wwwcip.informatik.uni-erlangen.de/~sifrfirs/TrimIntoIrregular.hip [wwwcip.informatik.uni-erlangen.de]
http://wwwcip.informatik.uni-erlangen.de/~sifrfirs/TrimIntoIrregular.hip [wwwcip.informatik.uni-erlangen.de]
Houdini Lounge » Cutting a hole in a nurbs circle that has been extruded.
- FrankFirsching
- 46 posts
- Offline
Hi!
The caps, that are placed at the ends of the extruded object are simple Bezier-curves. Houdini only shows them as surfaces. If you place a convert-SOP between the extrude and the project-SOP, with the following settings:
From type: Bezier curves
Convert to: NURBS surface
it should work.
Frank
The caps, that are placed at the ends of the extruded object are simple Bezier-curves. Houdini only shows them as surfaces. If you place a convert-SOP between the extrude and the project-SOP, with the following settings:
From type: Bezier curves
Convert to: NURBS surface
it should work.
Frank
Technical Discussion » Documentation of the image3d-format
- FrankFirsching
- 46 posts
- Offline
countv
Out of curiosity, what are CFD's? and Navier-Stokes?
CFD stands for Computational Fluid Dynamics and Navier-Stokes is the name of the mathematical formulas, that describe sich a flow. They are used to simulate the movement of water or gases.
Technical Discussion » Documentation of the image3d-format
- FrankFirsching
- 46 posts
- Offline
Hello!
I have written out some constant volumes using a vex shader and looked at the files they produced. I already understood, that the i3d-format is tile-based. But there are passages in the files, where one finds hundreds of zeros, which I think serve the only purpose to align the data to a specific memory-boundary. At the end of the file and also at some positions in the header, there are some numbers, which I don't know what they stand for. I already managed it to convert one of my own datasets into the i3d-format, but this isn't a general-purpose-converter. I hardcoded some numbers I found in the VEX-generated files. So it is tied to a special resolution like 64x64x64 voxels.
I think, the best would be, if SideFX would provide a pdf, which documents the file-format like they do it with the .geo-format.
It would also be interresting, if the file-format supports colors and not only density-values. Currently I use 4 different files for colored smoke, one for red, green, blue and alpha. If I could pack these informations into one i3d-file, it would also be better.
thanks
Frank
I have written out some constant volumes using a vex shader and looked at the files they produced. I already understood, that the i3d-format is tile-based. But there are passages in the files, where one finds hundreds of zeros, which I think serve the only purpose to align the data to a specific memory-boundary. At the end of the file and also at some positions in the header, there are some numbers, which I don't know what they stand for. I already managed it to convert one of my own datasets into the i3d-format, but this isn't a general-purpose-converter. I hardcoded some numbers I found in the VEX-generated files. So it is tied to a special resolution like 64x64x64 voxels.
I think, the best would be, if SideFX would provide a pdf, which documents the file-format like they do it with the .geo-format.
It would also be interresting, if the file-format supports colors and not only density-values. Currently I use 4 different files for colored smoke, one for red, green, blue and alpha. If I could pack these informations into one i3d-file, it would also be better.
thanks
Frank
Technical Discussion » Documentation of the image3d-format
- FrankFirsching
- 46 posts
- Offline
No, I don't want to write an image3d-shader using vex. What I have done is writing a stand-alone application, which solves a CFD problem using the Navier-Stokes equations. Now I'm having some volumetric datasets in my own file-format. To render them out using Houdini I need to convert them to the i3d-format. I looked at the files produced by the image3d-output and already noticed, that the files are tile-based. But there are some things, I don't understand what they mean.
Frank
Frank
Technical Discussion » Documentation of the image3d-format
- FrankFirsching
- 46 posts
- Offline
Hello!
During a project at university I have written a program, that simulates the flow of smoke. Now I would like to render these volumetric datasets in Houdini using i3d-textures. Looking through the PDFs I haven't found any format-description of i3d. Has anybody experiences in writing out own i3d-files?
Bye
Frank
During a project at university I have written a program, that simulates the flow of smoke. Now I would like to render these volumetric datasets in Houdini using i3d-textures. Looking through the PDFs I haven't found any format-description of i3d. Has anybody experiences in writing out own i3d-files?
Bye
Frank
-
- Quick Links