Found 57 posts.
Search results Show results as topic list.
Technical Discussion » GPU Cache or similar in Houdini
- Eckxter
- 61 posts
- Offline
Technical Discussion » Normals flipping on extract silhoutte.
- Eckxter
- 61 posts
- Offline
If the normals always have to point upwards, you can force them by setting them manually in an attribute wrangle.
With this little line of VEX:
That way, it doesn't matter whether they're flipped or not.
With this little line of VEX:
@N = set(0, 1, 0);
PDG/TOPs » ROP Fetch /OpenGL - Frame not finishing, yet output is there
- Eckxter
- 61 posts
- Offline
I'll have a look once it happens again and let you know.
When I changed some settings such as the frame range, it didn't happen anymore.
When I changed some settings such as the frame range, it didn't happen anymore.
PDG/TOPs » ROP Fetch /OpenGL - Frame not finishing, yet output is there
- Eckxter
- 61 posts
- Offline
Hi there,
I'm creating jpg flipbook sequences with the ROP Fetch fetching an OpenGL ROP.
For some reason, sometimes it gets stuck on a single frame.
The frame is already created, as I can open it in an image viewer, but TOPs doesn't see it.
It just keeps going, thinking the frame isn't ready, stalling the tasks down-stream.
When I right-click on the task and click ‘Cook Task’, it instantly finishes.
Since this is something inside an HDA, I'd like to not have to manually cook any straggler frames.
I haven't found anything on this so far, so any tips are most welcome.
I'm on Linux Ubuntu, H18.0.416.
Eckhart
I'm creating jpg flipbook sequences with the ROP Fetch fetching an OpenGL ROP.
For some reason, sometimes it gets stuck on a single frame.
The frame is already created, as I can open it in an image viewer, but TOPs doesn't see it.
It just keeps going, thinking the frame isn't ready, stalling the tasks down-stream.
When I right-click on the task and click ‘Cook Task’, it instantly finishes.
Since this is something inside an HDA, I'd like to not have to manually cook any straggler frames.
I haven't found anything on this so far, so any tips are most welcome.
I'm on Linux Ubuntu, H18.0.416.
Eckhart
Houdini Indie and Apprentice » Bake DOP simulation for export.
- Eckxter
- 61 posts
- Offline
Export them using an Alembic ROP?
By the way, I see in the preferences of your file that you changed your Unit Length to 0.01.
I understand that you're bringing your output to Maya, which is by default in CM instead of M.
But it's best, at least in my experience, to leave Houdini unit size to 1 meter.
Otherwise a lot of things can get very confusing.
Whenever you import from Maya to Houdini, just scale down by 0.01 and when you export, scale it back up by 100.
(assuming the geometry you're using is already the correct size.)
Just beware that if you do it now, you might have to change some things in your simulation again.
By the way, I see in the preferences of your file that you changed your Unit Length to 0.01.
I understand that you're bringing your output to Maya, which is by default in CM instead of M.
But it's best, at least in my experience, to leave Houdini unit size to 1 meter.
Otherwise a lot of things can get very confusing.
Whenever you import from Maya to Houdini, just scale down by 0.01 and when you export, scale it back up by 100.
(assuming the geometry you're using is already the correct size.)
Just beware that if you do it now, you might have to change some things in your simulation again.
Houdini Indie and Apprentice » set time on pyrosolver
- Eckxter
- 61 posts
- Offline
Usually you just change the Start Frame on the DOP Net.
Then inside the DOP Net, on the smoke object, try checking on “Solve on Creation Frame”.
Make sure your source is actually there on the frame you need it.
And check if your container is where your source is.
Or if you're not sure, share you .hip file and I'll have a quick look.
Then inside the DOP Net, on the smoke object, try checking on “Solve on Creation Frame”.
Make sure your source is actually there on the frame you need it.
And check if your container is where your source is.
Or if you're not sure, share you .hip file and I'll have a quick look.
Technical Discussion » Retimed temperature values fade/pop
- Eckxter
- 61 posts
- Offline
I've put an explosion sim through the Retime SOP in H17 and noticed the values of the temperature field fade and pop.
The pop happens every 20 frames, with which I assume is when the Retime reads an actual frame so it's not interpolating anymore.
Anything in between decreases in value slightly over time and then back up a bit, up until the pop.
I've been changing the parameters on the Retime node, but nothing seems to help.
Converting the VDBs back to Houdin Volumes also didn't change anything.
If anybody has any ideas, I'd be happy to hear them, as I've currently ran out.
Eckhart
The pop happens every 20 frames, with which I assume is when the Retime reads an actual frame so it's not interpolating anymore.
Anything in between decreases in value slightly over time and then back up a bit, up until the pop.
I've been changing the parameters on the Retime node, but nothing seems to help.
Converting the VDBs back to Houdin Volumes also didn't change anything.
If anybody has any ideas, I'd be happy to hear them, as I've currently ran out.
Eckhart
Edited by Eckxter - Oct. 12, 2018 10:38:36
Technical Discussion » Houdini to Deadline - Output path error
- Eckxter
- 61 posts
- Offline
This post is kind of a follow-up question on my other post asking about the wedge setup not producing multiple videos.
This one. [www.sidefx.com]
But as this post is about a different issue, I thought it'd be best to keep them apart.
I think I've found a thread asking about the same or a similar issue, but unfortunately nobody answered to it.
This is the tread. [www.sidefx.com]
I'm trying to send a Null ROP with other ROPs connected to it to Deadline.
When I click ‘Submit’ it fails and shows in the console the message “Output path for ROP is not specified”.
Then I tried with a Geometry ROP and it worked fine.
Then with an OpenGL ROP and again I get that error, even though its output is set to an accessible folder.
The eventual network I want to be able to send to Deadline is the following:
A Geometry ROP caching out the sim,
connected to an OpenGL or Mantra/Arnold ROP flipbooking or rendering the scene,
connected to a Shell ROP in which I run a pre-render script converting the generated image sequence to a video.
Locally it's not a problem. Only when I try to submit it with the Deadline submission tool, it gives me the error.
My understanding is that when submitting a .hip to a renderfarm, it just copies your scene to the farm and executes it there.
So I'm not sure why it cares so much whether or not it finds an output on the submitted ROP at the time of submitting.
This one. [www.sidefx.com]
But as this post is about a different issue, I thought it'd be best to keep them apart.
I think I've found a thread asking about the same or a similar issue, but unfortunately nobody answered to it.
This is the tread. [www.sidefx.com]
I'm trying to send a Null ROP with other ROPs connected to it to Deadline.
When I click ‘Submit’ it fails and shows in the console the message “Output path for ROP is not specified”.
Then I tried with a Geometry ROP and it worked fine.
Then with an OpenGL ROP and again I get that error, even though its output is set to an accessible folder.
The eventual network I want to be able to send to Deadline is the following:
A Geometry ROP caching out the sim,
connected to an OpenGL or Mantra/Arnold ROP flipbooking or rendering the scene,
connected to a Shell ROP in which I run a pre-render script converting the generated image sequence to a video.
Locally it's not a problem. Only when I try to submit it with the Deadline submission tool, it gives me the error.
My understanding is that when submitting a .hip to a renderfarm, it just copies your scene to the farm and executes it there.
So I'm not sure why it cares so much whether or not it finds an output on the submitted ROP at the time of submitting.
Technical Discussion » ffmpeg render script in wedge
- Eckxter
- 61 posts
- Offline
For the moment I sort of worked around it by flipbooking all wedges before converting them to a video.
After the wedge, the image sequence gets converted to an .mp4 with all versions in 1 file.
Even though this is fine for me, I'd still like to be able to create a video per wedge.
So if anybody has some input in the meantime, it'd be much appreciated.
I also found that the Shell ROP that was giving errors is inside the Wedge ROP.
Not sure how to fix that since all it seems to be doing is run “hou.node('..').hdaModule().applywedge(hou.node('..'))” as a Pre-Frame Script and I'm not too familiar yet with these scripts.
After the wedge, the image sequence gets converted to an .mp4 with all versions in 1 file.
Even though this is fine for me, I'd still like to be able to create a video per wedge.
So if anybody has some input in the meantime, it'd be much appreciated.
I also found that the Shell ROP that was giving errors is inside the Wedge ROP.
Not sure how to fix that since all it seems to be doing is run “hou.node('..').hdaModule().applywedge(hou.node('..'))” as a Pre-Frame Script and I'm not too familiar yet with these scripts.
Technical Discussion » ffmpeg render script in wedge
- Eckxter
- 61 posts
- Offline
I'm trying to set up a Wedge ROP network so that I can cache to disk, make flipbooks and then have the image sequences converted to separate video files with a render script using ffmpeg that I put in the Post-Render Script on the OpenGl ROP or in the Pre-Render Script on a Shell ROP that comes after the OpenGL ROP.
At the moment, it all works fine for the 1st wedge iteration.
I get the caches, I get the image sequence and ffmpeg gets called and successfully converts the images to a video file.
Then the wedge errors out and all it says is “Error rendering child: /out/wedge1/shell1”, with no further information.
The odd thing there is that even when I don't use the Shell ROP in my network, it still says that one is the problem.
But it only shows the error on the Wedge ROP.
When I don't call the render script and wedge for 3 iterations, I get 3 caches and 3 image sequences.
So I'm guessing something is going wrong after the script when the wedge is supposed to go to the second iteration.
At the moment, it all works fine for the 1st wedge iteration.
I get the caches, I get the image sequence and ffmpeg gets called and successfully converts the images to a video file.
Then the wedge errors out and all it says is “Error rendering child: /out/wedge1/shell1”, with no further information.
The odd thing there is that even when I don't use the Shell ROP in my network, it still says that one is the problem.
But it only shows the error on the Wedge ROP.
When I don't call the render script and wedge for 3 iterations, I get 3 caches and 3 image sequences.
So I'm guessing something is going wrong after the script when the wedge is supposed to go to the second iteration.
Edited by Eckxter - Jan. 11, 2018 08:26:10
Technical Discussion » Frame number in flipbook export
- Eckxter
- 61 posts
- Offline
Hi guys,
I was wondering if there is a way to keep the info that's shown in the top left corner of MPlay.
Specifically when exporting a flipbook.
In MPlay itself, it shows the creation time, the resolution, the frame number and the current color info you're viewing.
But when exporting either 1 frame or the entire flipbook, all this info disappears.
I'd love to keep all of this in that corner.
Thanks in advance!
I was wondering if there is a way to keep the info that's shown in the top left corner of MPlay.
Specifically when exporting a flipbook.
In MPlay itself, it shows the creation time, the resolution, the frame number and the current color info you're viewing.
But when exporting either 1 frame or the entire flipbook, all this info disappears.
I'd love to keep all of this in that corner.
Thanks in advance!
Houdini Engine for Maya » Copy Arnold StandIn
- Eckxter
- 61 posts
- Offline
Houdini Engine for Maya » Copy Arnold StandIn
- Eckxter
- 61 posts
- Offline
Sadly enough, I've already tried that to no avail.
As well as reloading the HDA and changing other parameters that do not change the output type and then doing a sync.
As well as reloading the HDA and changing other parameters that do not change the output type and then doing a sync.
Houdini Engine for Maya » Copy Arnold StandIn
- Eckxter
- 61 posts
- Offline
So far everything seems to be working. Except I'm running into a troublesome issue.
Normally I can see the points that get scattered, but sometimes they disappear, never to be seen again.
The thing is that I know they are still there, because I added the option to copy a box from within the HDA to the points.
When activating that option, the boxes get copied to where the points are.
But when deactivating the copying again, the points are still not visible.
Any clue on what could be happening? Maybe a viewport bug?
I've added a Maya2016 file and HDA in attachment. I hope it works if you want to test it.
Thanks!
Normally I can see the points that get scattered, but sometimes they disappear, never to be seen again.
The thing is that I know they are still there, because I added the option to copy a box from within the HDA to the points.
When activating that option, the boxes get copied to where the points are.
But when deactivating the copying again, the points are still not visible.
Any clue on what could be happening? Maybe a viewport bug?
I've added a Maya2016 file and HDA in attachment. I hope it works if you want to test it.
Thanks!
Houdini Engine for Maya » Copy Arnold StandIn
- Eckxter
- 61 posts
- Offline
Even though Maya would be able to render the instances, it seems the best option is the 2nd one.
It should work with outputting just the points and then instancing the stand-ins onto those.
Thanks!
It should work with outputting just the points and then instancing the stand-ins onto those.
Thanks!
Houdini Engine for Maya » Copy Arnold StandIn
- Eckxter
- 61 posts
- Offline
I'm struggling with copying a mesh to A LOT of points.
Since it gets very heavy in Maya to copy a mesh of 20 polygons to something like 5 million points,
I was thinking of creating an Arnold StandIn and copy that instead.
Unfortunately, when selecting the StandIn, I get this error:
“Selection has no supported object that can be used for input.”
Does anybody have an idea if this is at all possible or maybe a tip for another way to do this?
I'm currently working with Houdini 15.5 Indie licenses and in Maya 2016.
Thanks!
Since it gets very heavy in Maya to copy a mesh of 20 polygons to something like 5 million points,
I was thinking of creating an Arnold StandIn and copy that instead.
Unfortunately, when selecting the StandIn, I get this error:
“Selection has no supported object that can be used for input.”
Does anybody have an idea if this is at all possible or maybe a tip for another way to do this?
I'm currently working with Houdini 15.5 Indie licenses and in Maya 2016.
Thanks!
Houdini Indie and Apprentice » Rising wax from candle.
- Eckxter
- 61 posts
- Offline
I tried to do something with density with GAS solvers, but I couldn't get it to work.
I'm probably just doing something wrong, but I don't know as I do not have a lot of experience with and understand of GAS solvers.
So just to try and make it understandable;
The thing I would like to get is calculate the amount of particles in a separated bunch of particles.
Imagine a large body of fluid particles and now and then a bunch of particles rise up and detach from it, creating smaller bodies of fluid.
What I'd like to put together is something that would detect each of these bunch of particles and tell me the amount of particles that each body of fluid has.
So I suppose basically the mass of each of them.
Then each particle of its fluid body would get the mass assigned to it, which I would use to control its velocity.
With that same attribute I could drive a color value so that each fluid body has a different color.
I hope that's a bit clear.
Any help or tips on how to accomplish something like this would be highly appreciated!
Thanks!
I'm probably just doing something wrong, but I don't know as I do not have a lot of experience with and understand of GAS solvers.
So just to try and make it understandable;
The thing I would like to get is calculate the amount of particles in a separated bunch of particles.
Imagine a large body of fluid particles and now and then a bunch of particles rise up and detach from it, creating smaller bodies of fluid.
What I'd like to put together is something that would detect each of these bunch of particles and tell me the amount of particles that each body of fluid has.
So I suppose basically the mass of each of them.
Then each particle of its fluid body would get the mass assigned to it, which I would use to control its velocity.
With that same attribute I could drive a color value so that each fluid body has a different color.
I hope that's a bit clear.
Any help or tips on how to accomplish something like this would be highly appreciated!
Thanks!
Houdini Indie and Apprentice » Rising wax from candle.
- Eckxter
- 61 posts
- Offline
I took a different approach and got rid of the need to kill particles.
Now I would like to know if there is a way to calculate the amount of particles in a group af seperated fluid particles.
I'd like to make it so that drops of fluid with less particles travel faster than drops with more particles.
Now I would like to know if there is a way to calculate the amount of particles in a group af seperated fluid particles.
I'd like to make it so that drops of fluid with less particles travel faster than drops with more particles.
Houdini Indie and Apprentice » Rising wax from candle.
- Eckxter
- 61 posts
- Offline
I have now removed the group nodes and just put in VEX rules to set particles' dead value on 1.
I have used all of the following, but none seemed to kill any particle.
if (@age >= 1 && length(@v) < 1)
{
dead = 1;
}
———————————
if (@age > 1)
{
dead = 1;
}
———————————
dead = @age > 1 && length(@v) < 1;
———————————
dead = @age > 1;
———————————
dead = length(@v) > 0;
———————————
dead = @age > 0.5;
I made sure that I activated Age Particles on the Solver.
But even then, killing them purely on age for the sake of testing doesn't kill any particle.
And killing them when velocity length is larger than 0 also does nothing.
I really don't understand why it's not working.
And I have the feeling it's going to be something very minute.
I have used all of the following, but none seemed to kill any particle.
if (@age >= 1 && length(@v) < 1)
{
dead = 1;
}
———————————
if (@age > 1)
{
dead = 1;
}
———————————
dead = @age > 1 && length(@v) < 1;
———————————
dead = @age > 1;
———————————
dead = length(@v) > 0;
———————————
dead = @age > 0.5;
I made sure that I activated Age Particles on the Solver.
But even then, killing them purely on age for the sake of testing doesn't kill any particle.
And killing them when velocity length is larger than 0 also does nothing.
I really don't understand why it's not working.
And I have the feeling it's going to be something very minute.
Houdini Indie and Apprentice » Rising wax from candle.
- Eckxter
- 61 posts
- Offline
Hi guys,
I have a couple of issues with my project.
What I want to accomplish is wax rising from a candle in interrupted streams and it has to be able to loop every 10 seconds.
The first issue that I'm having is that the particles that I don't want anymore, are not being killed by the POP Kill.
I want to kill particles that are older than 3 seconds and barely move anymore.
As you will see in the .hiplc, I am first grouping particles that are older than 3 seconds.
Then from that group, I group the particles that have a velocity length of lower than 0.5.
This is to prevent getting these very long and thin streams of wax.
I would like to have shorter streams as those that I get at the start of each stream and that they then just cut loose from the candle instead of these constant streams.
I had found this thread [sidefx.com] in which they mention to put an * in one of the group nodes.
I did that on my initial group node, as I source from other groups down the line.
But that didn't help.
My second issue is that I would like to get different velocities for the different streams of wax.
I'm not quite sure how I could assign multiple velocities to groups of particles within the same simulation.
I currently set the gravity in the positive direction, but I'm guessing I'm going to need something else eventually.
Or should I just create particle simulation for each stream of wax?
Thanks in advance!
Eckxter
I have a couple of issues with my project.
What I want to accomplish is wax rising from a candle in interrupted streams and it has to be able to loop every 10 seconds.
The first issue that I'm having is that the particles that I don't want anymore, are not being killed by the POP Kill.
I want to kill particles that are older than 3 seconds and barely move anymore.
As you will see in the .hiplc, I am first grouping particles that are older than 3 seconds.
Then from that group, I group the particles that have a velocity length of lower than 0.5.
This is to prevent getting these very long and thin streams of wax.
I would like to have shorter streams as those that I get at the start of each stream and that they then just cut loose from the candle instead of these constant streams.
I had found this thread [sidefx.com] in which they mention to put an * in one of the group nodes.
I did that on my initial group node, as I source from other groups down the line.
But that didn't help.
My second issue is that I would like to get different velocities for the different streams of wax.
I'm not quite sure how I could assign multiple velocities to groups of particles within the same simulation.
I currently set the gravity in the positive direction, but I'm guessing I'm going to need something else eventually.
Or should I just create particle simulation for each stream of wax?
Thanks in advance!
Eckxter
-
- Quick Links