FSXA MY LOD Experiment

[JohnC1]

Taxiing up to a GMAX building looks better if small 3d details are included, and using LOD type names saves CPU/GPU resources by only resolving those details when you are up close enough.

I had trouble getting understandable results so I created a standalone MDL with structures that varied in size and LOD values.

I put it 1000m above the ocean in line with a coastal runway for easy access.

I created large (APPROX 1000M) and small(APPROX 45M) structures with the same LOD values, using LOD_050, LOD_100, and LOD_150's as values.

I also added in a standard non_LOD object and 1 LOD_160. All were in the same vertical plane.

See attached GMAX object gif.

GREEN HAS A STANDARD NON-LOD NAME.
BLUE-----------LOD_050.
WHITE----------LOD_100.
RED------------LOD_150.
PINK-----------LOD_160.

Here are my slew results.

A "0" means not visible, a "1" means just became visible.
From top to bottom in this table means slewing closer and closer.

SIZE--- LOD_050- no LOD- LOD_100--LOD_150
Large ------ 0------- 0--------0--------- 0 FAR
Small ------ 0------- 0--------0--------- 0

Large ------ 1------- 0--------0--------- 0 NEARER
Small ------ 1------- 0--------0--------- 0

Large ------ 0------- 1<----->1--------- 0 NEARER
Small ------ 0------- 1<----->1--------- 0

Large ------ 0------- 0--------0--------- 1 NEARER
Small ------ 0------- 0--------0--------- 1

Predictably by now, all disappear when the single pink LOD_160 pops into view.

Here is what I glean from this

RULE 1. USE A STANDALONE MDL TO ADD DETAIL UP CLOSE.

RULE 2. USE ONLY ONE LOD VALUE IN A STANDALONE MDL.
ON APPROACHING, WHEN CONDITIONS ARE MET TO DISPLAY ANY LOD, IT BLANKS OUT ALL THOSE WITH DIFFERENT LOD NUMBERS. BLANKING OUT OBJECTS AS ONE GETS CLOSER DOESN'T MAKE SENSE.

RULE 3. DON'T USE OBJECTS IN A STANDALONE MDL WITH NO LOD VALUE.
OBJECTS WITH AN LOD VALUE NOT EQUAL TO 100 BLANK OUT ANY OBJECT WITH NO LOD VALUE.

RULE 4. ALL VALUES WITH THE SAME LOD COME INTO VIEW SIMULTANEOUSLY REGARDLESS OF SIZE. WHEN AN LOD_050 OBJECT DISPAYS BECAUSE CONDITIONS ARE MET THE SMALL ONES POP INTO AND OUT OF VIEW AT THE SAME TIME AS THE LARGER ONES.

??????

I DID A SCREEN CAPTURE AT EACH POINT WHERE A NEW SET OF LOD'S POPPED INTO VIEW. I COULD NOT SEE ANY PARTICULAR RULE OF THUMB REGARDING THE SIZE OF THE OBJECTS ON THE SCREEN, BUT MAYBE SOMEONE ELSE CAN.
COLOR VALUES ARE IN PIXELS AND THE IMAGE WAS CREATED FROM 4 COMBINED SCREENSHOTS.

Building a SIMPLE theoretical model like this gave me instant insight.

HOPE THIS HELPS SOMEONE.

 

[arno]

Hi,

Each level of detail indeed replaces the previous level of detail. You can not make additive levels of detail in mdl files. So if you only want to add more detail, you would have to duplicate the simpler geometry as well.

The distance when an object appears depends on the size of the object (bounding box of the mdl). The number in the lod is roughly the amount of pixels the objects should cover to appear.