Mitsuba Exporter
- TwinSnakes007
- Posts: 316
- Joined: 06 Jun 2011, 16:00
Re: Mitsuba Exporter
For instance, you could render this in Mitsuba.
Re: Mitsuba Exporter
oh interesting, ok thats how I imagined it being implemented with ICE only, but was thinking about the grid format itself on how to bring the grid volume data from another format like .Bafl or field3d or http://www.openvdb.org/ or maya fluids or houdini bgeo etc.
so for example if we managed to get the grid data into softimage from other formats and changed each cell to be a particle and reassign the density - mass etc we should be able to render it in Mitsuba!
so for example if we managed to get the grid data into softimage from other formats and changed each cell to be a particle and reassign the density - mass etc we should be able to render it in Mitsuba!
Gossip is what no one claims to like, but everybody enjoys.
- TwinSnakes007
- Posts: 316
- Joined: 06 Jun 2011, 16:00
Re: Mitsuba Exporter
Its a very simple binary format. You describe a header, and then write the data.Tekano wrote:but was thinking about the grid format itself
Here's a C# implementation of the GridVolume format.
GridVolume.cs
Re: Mitsuba Exporter
any news on this twinsnakes? looking forward to trying this out for sure
Gossip is what no one claims to like, but everybody enjoys.
- TwinSnakes007
- Posts: 316
- Joined: 06 Jun 2011, 16:00
Re: Mitsuba Exporter
I'm still pluggin away at it. I had no idea ICE RenderTree attributes would be this tough to incorporate into the code base. They have been a pain. Basically, everything can have an ICE tree!
I've got some code working that detects an ICE Attribute and validates that it has ICE data (attribute not set to "Use Default Value"), so I'm pretty close.
I've got some code working that detects an ICE Attribute and validates that it has ICE data (attribute not set to "Use Default Value"), so I'm pretty close.
Re: Mitsuba Exporter
if you are having problems implementing ICE attributes in your code you should definately post questions, maybe someone has been there before and can help you!
Gossip is what no one claims to like, but everybody enjoys.
Re: Mitsuba Exporter
...or you can look at this from another side. At least from what I heard, this issue seems to appear again in SI 2013, global score of working-not working seems to be fifty-fifty, for out of the box renderer. So, few attributes more or less, who cares
- TwinSnakes007
- Posts: 316
- Joined: 06 Jun 2011, 16:00
Re: Mitsuba Exporter
I'm not sure I understand what the bug is? Is it that the reading of ICE attributes is not consistent?Mathaeus wrote:...or you can look at this from another side. At least from what I heard, this issue seems to appear again in SI 2013, global score of working-not working seems to be fifty-fifty, for out of the box renderer. So, few attributes more or less, who cares
Re: Mitsuba Exporter
ha well in Matheus's case the bug (between scanline and rasterizer not having same values for attributes) is mute... because we will be rendering attributes with Mitsuba!
Gossip is what no one claims to like, but everybody enjoys.
Re: Mitsuba Exporter
Yeah, sometimes happens, that renderer reads ICE tree 'partially' - what you see in viewport, doesn't appear into renderer. It could be worse than just wrong reading of color attribute. Sometimes, reason is NaN in ICE (still solved by ICE, don't know how). sometimes, it's not. Form my knowledge, it's not only with Mental Ray, but integration with MR seems to be champion in errors.TwinSnakes007 wrote:
I'm not sure I understand what the bug is? Is it that the reading of ICE attributes is not consistent?
- TwinSnakes007
- Posts: 316
- Joined: 06 Jun 2011, 16:00
Re: Mitsuba Exporter
Oh yeah, I'm very familiar with NaN's, especially in reading ICE Strand data. So I do error checking when I'm reading ICE Attributes and use a default value when I encounter a NaN.Mathaeus wrote:Yeah, sometimes happens, that renderer reads ICE tree 'partially' - what you see in viewport, doesn't appear into renderer. It could be worse than just wrong reading of color attribute. Sometimes, reason is NaN in ICE (still solved by ICE, don't know how). sometimes, it's not. Form my knowledge, it's not only with Mental Ray, but integration with MR seems to be champion in errors.TwinSnakes007 wrote:
I'm not sure I understand what the bug is? Is it that the reading of ICE attributes is not consistent?
AFAIK, NaN's come from particle attributes on particles that are Born but are not evaluated yet. Basically, the data array has been created as a place holder for evaluation.
- TwinSnakes007
- Posts: 316
- Joined: 06 Jun 2011, 16:00
Re: Mitsuba Exporter
Looks like its working in my first test. Exported 602 Particles and wrote 6,000 lines of XML in 1.14 seconds.
SI Scene Mitsuba Render
SI Scene Mitsuba Render
- TwinSnakes007
- Posts: 316
- Joined: 06 Jun 2011, 16:00
Re: Mitsuba Exporter
Finally got multi-material working.
Next up is UV's and VertexColors.
Next up is UV's and VertexColors.
- TwinSnakes007
- Posts: 316
- Joined: 06 Jun 2011, 16:00
Re: Mitsuba Exporter
Since both Vertex Colors and Texture Coordinates are in the Sample space, I got UV's for free!
Re: Mitsuba Exporter
Hey great work... Mitsuba looks amazing, and your work on the exporter is great.
I had one question, how does mitsuba handle output? Im guessing aov's are not implemented yet but is there any settings for handling what file format to output and similar?
I had one question, how does mitsuba handle output? Im guessing aov's are not implemented yet but is there any settings for handling what file format to output and similar?
Who is online
Users browsing this forum: No registered users and 34 guests