Sorry for the rant…
I'm working in a scene with insanely large scales (yes, space), and the precision errors were the root cause.
Pretty sure I've got a workaround now.
Thanks!
-Jon
Found 102 posts.
Search results Show results as topic list.
Technical Discussion » CHOP lookat constraint randomly flipping
- jpparkeramnh
- 178 posts
- Offline
Technical Discussion » CHOP lookat constraint randomly flipping
- jpparkeramnh
- 178 posts
- Offline
Is anyone else having major issues getting these to work?
I have a scene with a camera transform defined in a CHOP network and am getting constant axis flipping when I play back the animation. All combinations of using an up object, etc. etc.
The old style lookat parameters work fine.
I really want to like these new constraint nodes but so far they have just been a big headache….
I have a scene with a camera transform defined in a CHOP network and am getting constant axis flipping when I play back the animation. All combinations of using an up object, etc. etc.
The old style lookat parameters work fine.
I really want to like these new constraint nodes but so far they have just been a big headache….
Technical Discussion » Lens shader ray differentials
- jpparkeramnh
- 178 posts
- Offline
Technical Discussion » mantra: [VRAY_ProcGT] procedural warning: Optimization not run on geometry (this may be ok)
- jpparkeramnh
- 178 posts
- Offline
What is this warning all about?
I'm getting it on a scene that loads just a few packed primitives.
I'm getting it on a scene that loads just a few packed primitives.
Technical Discussion » volume splice opposite?
- jpparkeramnh
- 178 posts
- Offline
Howdy,
Is there a node that will split up volumes by spatial regions… and ensure that the voxels are aligned and do not overlap?
Cheers,
Jon
Is there a node that will split up volumes by spatial regions… and ensure that the voxels are aligned and do not overlap?
Cheers,
Jon
Technical Discussion » Lens shader ray differentials
- jpparkeramnh
- 178 posts
- Offline
Technical Discussion » Lens shader ray differentials
- jpparkeramnh
- 178 posts
- Offline
I posted a bug about this, but still am waiting for a response so what the heck:
Has anyone else had the experience where using a custom lens shader seems to change the size of the ray derivatives quite a bit? We wrote a fisheye shader that we're using for planetarium productions.
One example: vm_geofilterwidth needs to be increased 2-4X to match what a standard ASAD lens looks like.
Another example: displacement dicing seems to be a lot more memory intensive, probably because the dicing width in screen space is smaller.
Has anyone else had the experience where using a custom lens shader seems to change the size of the ray derivatives quite a bit? We wrote a fisheye shader that we're using for planetarium productions.
One example: vm_geofilterwidth needs to be increased 2-4X to match what a standard ASAD lens looks like.
Another example: displacement dicing seems to be a lot more memory intensive, probably because the dicing width in screen space is smaller.
Technical Discussion » Benefits of bumping RAM up
- jpparkeramnh
- 178 posts
- Offline
The more RAM you get, the more you will notice how fast the RAM is. I've worked on a machine with cheap slower RAM (but lots of it) and definitely would notice a (not so)brief freeze whenever Houdini would clear the cache.
So RAM speed will come into play as well.
So RAM speed will come into play as well.
Technical Discussion » Non-centered vertical FOV camera to window offset calculation (trig question)
- jpparkeramnh
- 178 posts
- Offline
I think I got it, this seems to work:
sin(-35.67 + 40.72) / 2. / sin(abs(40.72) + abs(-35.67))
generalized to:
sin(neg_vfov + pos_vfov) * 0.5 / sin(abs(neg_fov) + abs(pos_vfov))
sin(-35.67 + 40.72) / 2. / sin(abs(40.72) + abs(-35.67))
generalized to:
sin(neg_vfov + pos_vfov) * 0.5 / sin(abs(neg_fov) + abs(pos_vfov))
Edited by jpparkeramnh - 2016年12月28日 21:40:27
Technical Discussion » Non-centered vertical FOV camera to window offset calculation (trig question)
- jpparkeramnh
- 178 posts
- Offline
Greetings,
I'm having a “maths” amnesia moment…
Given a camera that is 2048x1536, 0.971 pixel aspect, I need to set the screen window offset so that the FOV is -45.62 - +45.62 degrees horizontal and -35.67 - +40.72 degrees vertical.
What's the trig to calculate the screen window offset, given that the vertical FOV is off-center as described above?
Any “teach me how to fish” advice would be very helpful…
Cheers,
Jon
I'm having a “maths” amnesia moment…
Given a camera that is 2048x1536, 0.971 pixel aspect, I need to set the screen window offset so that the FOV is -45.62 - +45.62 degrees horizontal and -35.67 - +40.72 degrees vertical.
What's the trig to calculate the screen window offset, given that the vertical FOV is off-center as described above?
Any “teach me how to fish” advice would be very helpful…
Cheers,
Jon
Edited by jpparkeramnh - 2016年12月28日 01:13:00
Technical Discussion » Ability to set Viewport Up=Y or Z outside of prefs window?
- jpparkeramnh
- 178 posts
- Offline
Hi,
Is there a way to adjust this preference other than through user settings? Perhaps an environment variable or HOM / hscript command?
Cheers,
Jon
Is there a way to adjust this preference other than through user settings? Perhaps an environment variable or HOM / hscript command?
Cheers,
Jon
Houdini Lounge » Packed primitives information
- jpparkeramnh
- 178 posts
- Offline
Thanks for the explanation, it's an eye opener.
Since mantra doesn't yet support nested instancing, what about RenderMan support for this? For RenderMan, would it be technically possible to generate a delayed read archive or instance procedural at each packed primitive level?
There was a paper by Pixar describing some of the tradeoffs between many instances vs. fewer larger instances in RenderMan, which sounds similar to what is happening with mantra.
I may be misunderstanding RenderMan's use of instances as I have only used them briefly on a show, and they were generated by in-house software.
Since mantra doesn't yet support nested instancing, what about RenderMan support for this? For RenderMan, would it be technically possible to generate a delayed read archive or instance procedural at each packed primitive level?
There was a paper by Pixar describing some of the tradeoffs between many instances vs. fewer larger instances in RenderMan, which sounds similar to what is happening with mantra.
I may be misunderstanding RenderMan's use of instances as I have only used them briefly on a show, and they were generated by in-house software.
Technical Discussion » Reading specific channels from Alembic Mesh
- jpparkeramnh
- 178 posts
- Offline
Check out the Python alembic extensions:
http://www.sidefx.com/docs/houdini12.5/hom/abc_extensions [sidefx.com]
In particular you want the function alembicArbGeometry.
What I've done is use an attribCreate node with python expressions to grab the abc attributes I want to apply them to the geometry.
http://www.sidefx.com/docs/houdini12.5/hom/abc_extensions [sidefx.com]
In particular you want the function alembicArbGeometry.
What I've done is use an attribCreate node with python expressions to grab the abc attributes I want to apply them to the geometry.
Technical Discussion » Rendering artifacts with Volume procedural octree divisions
- jpparkeramnh
- 178 posts
- Offline
I had the same problem. Ended up using a coarse (1024) vdb grid as the acceleration structure and doing the density falloff in the surface shader instead.
Technical Discussion » Volume FFT convolve
- jpparkeramnh
- 178 posts
- Offline
So now I'm getting the result back into the original position, convolved, but the value scale is way off. I suspect it's because I need complex multiplication and not just a volume mix/multiply? Is there a way to do this while not using python/C++?
EDIT: found a way here… should have been obvious.
http://odforce.net/wiki/doku.php?id=wikicomplex.h&s [odforce.net]=complex
Also, I'd forgotten that blur kernel volumes need to be normalized to sum to 1.
EDIT: found a way here… should have been obvious.
http://odforce.net/wiki/doku.php?id=wikicomplex.h&s [odforce.net]=complex
Also, I'd forgotten that blur kernel volumes need to be normalized to sum to 1.
Technical Discussion » Volume FFT convolve
- jpparkeramnh
- 178 posts
- Offline
I'm trying to convolve one set of color volumes with another and can't seem to get the settings right to give me sane results… I can't seem to get the inverse FFT results to not be transformed to the corners. Any tips?
Technical Discussion » note on memory usage when rendering lots of particles
- jpparkeramnh
- 178 posts
- Offline
You should modify opacity with another VOP, like a constant or noise.
Generally points don't look good large and up close, and the answer is usually to generate many many more of them so that the individual points are no larger than a few pixels on-screen. This way they all average together to create a more complicated look. The sum is greater than its parts, as they say.
Generally points don't look good large and up close, and the answer is usually to generate many many more of them so that the individual points are no larger than a few pixels on-screen. This way they all average together to create a more complicated look. The sum is greater than its parts, as they say.
Technical Discussion » note on memory usage when rendering lots of particles
- jpparkeramnh
- 178 posts
- Offline
You should modify opacity with another VOP, like a constant or noise.
Generally points don't look good large and up close, and the answer is usually to generate many many more of them so that the individual points are no larger than a few pixels on-screen. This way they all average together to create a more complicated look. The sum is greater than its parts, as they say.
Generally points don't look good large and up close, and the answer is usually to generate many many more of them so that the individual points are no larger than a few pixels on-screen. This way they all average together to create a more complicated look. The sum is greater than its parts, as they say.
Technical Discussion » note on memory usage when rendering lots of particles
- jpparkeramnh
- 178 posts
- Offline
In a VOP material you can use the Lighting Model node, set to ‘isotropic’, or use the ‘physically based phase lighting’ node. Modifying this and the opacity should give you a more volumetric look to your particles.
Technical Discussion » note on memory usage when rendering lots of particles
- jpparkeramnh
- 178 posts
- Offline
In a VOP material you can use the Lighting Model node, set to ‘isotropic’, or use the ‘physically based phase lighting’ node. Modifying this and the opacity should give you a more volumetric look to your particles.
-
- Quick Links