Level of Detail


Using Level of Detail 

LOD 

Level of detail, LOD, techniques can be a useful tool to maximize scene efficiency and provide maximum throughput of geometry, while also providing other benefits. The idea behind LOD is to provide multiple versions of a model to the renderer.  Now, it can decide to use one or more of the models based on quality and efficiency considerations. This can reduce the memory footprint and time required to render the object. Level of detail is especially relevant for very large, complicated, models and in situations where many characters or objects are running around. In a shot of a crowd, characters up close to the camera will have sufficient intricacies, while on a distant ridge a much simpler model can be substituted for the character. 

  • Models of different complexities can be substituted during certain ranges, measured in pixels.
  • Versions of a model must be blended together, to provide seamless transitions.
  • LOD increases scene efficiency and decreases render times, by minimizing memory footprints. 
  • Reduce aliasing artifacts by using models with appropriate detail for their size.

With MTOR you can attach level of detail attributes to groups of objects.  Here's the basic workflow:

  1. construct several (3 or 4) versions of your model
  2. group them under a single grouping node
  3. attach MTOR LOD Control attributes to the grouping node
  4. attach MTOR LOD Detail Range attributes to the top node of each model version
  5. assign a Slim Archive appearance to the top node of each model version
 

LOD Attributes Menu

You'll want to group the different variations of a model together (simple, medium, and complex versions). To this grouped node, parent to the model versions, you'll want to attach a "LOD Control" 

Attributes> Level of Detail> Attach LOD Control

LOD Detail Range Menu

The "Detail Range" attribute must be attached to each version of your LOD model. Each version of a model must have an Archiver attached to it.

Attributes> Level of Detail> Attach LOD Control

The detail range is entered in pixel area. A 10x10 pixel limit would be entered as "100," for instance. For more explanation of what these boxes mean, read on . . .

 

Detail Range

The detail range not only allows the specification of the visible range of a particular model version, it also lets you specify how different complexities of models will blend together. The "start" value defines at what resolution (in pixel area) that the image will come into view, appearing first as nearly transparent. The opacity increases as the resolution approaches the "low" value, at which point the object becomes completely opaque. Up to the "high" value, the object is opaque. During the range specified by "high" to "end," the model version will fade out, linearly.  

Start Low High End
0 0 100 144

This would be a detail range for the simplest model. At zero pixels the simple model is used. Here we can see that once the model reaches a pixel size of 10 pixels (100 in pixel area) it will fade out until it reaches its maximum "end" limit, 12 pixels across, (144 in pixel area). 

100 144 5184 6400

This would be a detail range for a medium-res model. from the width of 10 to 12 pixels the model will fade in, from 12 to 72 the model is opaque, and from 72 to 80 the model fades out, presumably into the high-res model. (The pixel area for a width of 5184 is 72 and the pixel area of a 80x80 pixel square is 6400.) 

 

Attach Archivers

Remember that Archivers must be attached to each of the different resolution models.

 


 

Pixar Animation Studios
(510) 752-3000 (voice)   (510) 752-3151 (fax)
Copyright © 1996- Pixar. All rights reserved.
RenderMan® is a registered trademark of Pixar.