Posted: Fri Nov 20, 2009 6:02 am Post subject: Particle Volume Cloud: Lookup Table issues with Satellites ?
Hi,
I 'm having a serious problem here. In a particle cloud volume, the Lookup Table goes nuts when rendering with satellites. Every refresh of a render region, for example, gives a different Lookup Table, and consequently different cloud "looks" (at certain cloud positions). When switching back to local rendering only, order is restored
This is more evident when I disable "Cell Interpolation" - then the wacky Table is visible in all its glory ! Certain Cells appear wildly stretched, others lighter/darker, etc.
Switching the Table Interpolation setting to "Full Raytracing" eliminates the problem, but that, naturally, has a huge impact in render times.
BA Volume is installed in Workgroup path, satellites are properly configured for 3d party shaders, and there are no license file access problems across.
The setup is very simple, a grid emitter, some ICE particles (rate 250), some forces, and a typical Particle Volume / Density / Fractal Scalar shader, set to produce some pretty dense particles (overall density 50). Acceptable Cell Size for this is 0.2, Marching settings set to Auto, High quality.
Softimage 2010 SP1, Vista64, latest BA Essentials / BA Volume versions installed.
So, is this a known issue, or am I doing something wrong ? Assuming the former, is there maybe a workaround ?
This is the first report about a broken lookup table for satellite rendering.
Are you using the XSI particle compound or seperate shader nodes?
Just to check: all satellite machines have sp1 installed for rendering? _________________ Holger Schönberger
Binary Alchemy - digital materialization
I 'm using separate nodes, no XSI compounds anywhere. And yes, all satellites have SP1 installed (actually, I 'm trying it with just one satellite so that makes it even simpler).
I should also mention that the problem is not very consistent. Over rendering of a sequence of 250 frames, it only comes up after frame 100 or something - sometimes it goes off at around 90, sometimes around 150 or more. I noticed that a full reboot (Windows, not just XSI) of the host machine makes the problem appear even later (after frame 160-170 in the same sequence). But once it does appear, then it doesn't go away, for any frame, and just keeps getting worse, until the next reboot.
The particles are pretty dense and they are set to "pop" in, in full scale (no fade-in or gradual sizing-up from the emitter).
I found another thing that looks like it could be the cause of the problem :
Looking into the plugin manager, I can see the BA Volume entry. There, under the "\Addons\BA_Shader_Collection_volume\Application\Plugins\" path, there is a "BA_ICE_cache.64.dll" file, which is reported as "Invalid" in the XSI plugin manager.
Could this be the problem ? If yes, why is it "invalid" and how can I fix it ? (I 've already tried uninstalling/reinstalling BA Volume, with XSI restarts in between).
I even tried completely uninstalling BA Volume, and using only the version included with XSI. Still the same problem.
By the way - in XSI 7.x and 2010, BA Volume was more or less included. I assume that if you install BA Volume v3.x anyway, it "takes over" (so you 're always rendering using the latest version), is that right ?
For example, I can't see "fractal_4d_scl" anymore, but of course I see XSI's "Fractal_Scalar" instead (which looks like it's the same). Same with "BA_volume_cloud". Is this behavior correct ? Am I using the latest BA Volume version (and is there any way to verify that) ?
Apologies for all the questions, I just seem to be a bit confused here. I really hope we can sort this thing out. Thanks again !
No, that file is not used.
The BA volume package has additional shaders to the volume cloud shader.
It does not replace the XSI shaders.
As you said that restarting the machine helps, can you take a look at free memory of the satellite?
Any problem there?
All machines are x64? And have the same ammount of memory? _________________ Holger Schönberger
Binary Alchemy - digital materialization
No, that file is not used.
The BA volume package has additional shaders to the volume cloud shader.
It does not replace the XSI shaders.
Then something must be wrong there too, because I only see the "Particle volume cloud" shader (the one included with XSI) available, and not the "BA Volume Cloud" anywhere. I have reinstalled the BA Volume Shaders and it's still the same. You mean I should be seeing both the BA_Volume_Cloud and XSI's Particle_Volume_Cloud ?
Quote:
As you said that restarting the machine helps, can you take a look at free memory of the satellite?
Any problem there?
All machines are x64? And have the same ammount of memory?
Yes, all machines are identical, all x64 with 8 GB ram.
What about tha "invalid" .dll I mentioned, could that be relevant to my problem ?
You can remove the dll with that problem. It is not used and not a part of the shading network. _________________ Holger Schönberger
Binary Alchemy - digital materialization
So, when installed in 7.x/2010, your addon does not touch the XSI-included "Particle Volume" shader ? It doesn't replace it with a newer version ? In other words, in XSI versions after 7.x, the included "Particle Volume" shader is exactly the same, whether you have the BA_Volume package installed or not ?
Apologies for insisting on this, but I need to understand what's happening here; because I found that if I uninstall BA_Volume completely, the same scene renders perfectly with satellites (obviously using the included Particle_Volume shader).
But if the installation does not mess with the included particle volume shader, then why doesn't the scene render correctly (with satellites) when the addon is installed ?
There is only the volume shader in XSI.
The 7.5/2010 addon does not have have a dll/so file for the volume shader.
You have the same volume shader if you install the addon or not.
The addon extent the functionality with extra shaders like the mesh distance, PT cache loader...
So if there is a problem with the lookup table, then it is part of XSI, not part of the BA addon. _________________ Holger Schönberger
Binary Alchemy - digital materialization
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum