MayaMan will interpret Maya's Level Of Detail groups and convert them to renderman ones automatically... though manual control of the blend zones will almost always be required.
MayaMan makes a stab at emulating Maya's distance-to-camera LOD selection but since renderman uses a screen area based metric the conversion is a poor approximation at best and there is no control over the blend widths.... better than nothing though.
Select a lodGroup node in the Outliner and invoke the MayaManMenu entry to add attributes to a LOD group. This will add one array of floats to the extra attributes section for each child of the lod group.
The first array of floats has two entries and it controls the highest quality model. The first of these values represents the detail value (size in pixels) at which that model will start to fade out and the second value is the detail value at which it is no longer visible.
| lowerbound | the detail level at which the model starts to fade out as it moves toward minvis |
| minvis | the detail level at which the model ceases to be visible as it gets smaller (or starts to be visible as it gets larger) |
The intermediate arrays each have four entries:
| maxvis | the detail level at which the model ceases to be visible as it gets bigger (or starts to be visible as it gets smaller) |
| upperbound | the detail level at which the model starts to fade out as it moves toward maxvis |
| lowerbound | the detail level at which the model starts to fade out as it moves toward minvis |
| minvis | the detail level at which the model ceases to be visible as it gets smaller (or starts to be visible as it gets larger) |
The last array of floats also has two entries and it controls the lowest quality model. The first of these values represents the detail value at which the model will start to fade in and the second of these is the value at which the model is fully visible.
| maxvis | the detail level at which the model ceases to be visible as it gets bigger (or starts to be visible as it gets smaller) |
| upperbound | the detail level at which the model starts to fade out as it moves toward maxvis |
A well formed setup will have monotonically decreasing detail values as you read down from the top. Also, for any given value the sum of the blended contributions should be 1. The best/easiest way to ensure this is to make the first two entries of any particular group be the same as the last two from the previous group and vice versa. MayaMan does not enforce this as there may be artistic reasons to violate this principle. The following table illustrates the recommended relationship between the values in one row and the next and further explains why the first and last entries only have two values.
| maxvis | upperbound | lowerbound | minvis | |
| highrez model | assumed to be infinite | assumed to be infinite | xxxx | yyyy |
| midrez model | xxxx | yyyy | aaaa | bbbb |
| lowrez model | aaaa | bbbb | assumed to be zero | assumed to be zero |