question

This is all due to a breaking change in Blender as you have found out. The new comp node is an image node with R G B output. The ASOBO exporter and the Vitus legacy exporter are looking for the old Separate Red Green Blue node. So you must stay at blender 3.1.0. Once you go to blender 3.3 this breaking change affects all MSFS exporters. If you have saved your blend file from 3.3 you will have to delete and redo all your materials. ( I believe most of the values stay for factors and values etc.) but you will need to check them all. It's very difficult to go back from 3.3 to 3.1. Not even sure file Append will work.

A note to all stay on blender 3.1.0 and the legacy exporter.

It seems you built this in blender 3.3 and are now going back to 3.1? correct?

I just tested by creating a new material in Asset Library and using it from there VS creating a new material inside the Blender file. It only works within the Blender file! Every time a material is used from Asset Library or possibly via copy-paste (not tested yet with the new Exporter) the ORM is ignored despite being present in the settings. The same bug was there with Blender2MSFS Exporter but it was occasionally fixed by reloading images... Something must be differen in how materials are exported based on whether material is new or pasted/linked from Asset Library or other file.

Sometimes when copying between files the gltf settings node is duplicated which causes the sim to not use the comp texture. Setting it back to the original makes the material work again. Could be worth checking.

I think I now have a workaround! For anyone stuck this could be valuable - I wasted 2 days trying to overcome this problem. Turns out there are 2 separate issues:

1. The new Exporter presents a button "Migrate Material" when the material was originally done in a legacy exporter. The material that's created after that looks OK on the first glance, but if the same material is created from scratch you can see a difference in the RGB splitter node.

This is the migrated MSFS material, it has "Color" as input in the splitter node:

And this is the newly created material, it has "image" input in a splitter node:

I can copy this node from a fresh material and paste into the broken material, and reconnect the nodes in the same way. This fixes how the material looks inside blender. However this still doesn't fix the export, comp texture is still missing in MSFS.

2. After dropping the material from Asset Browser into your current project file, I have to revisit the shader tree and check that the gTLF settings has not created a new setting, and if it did to revert it to the first original settings:

This is the object dropped from the Asset Library with its materials with the first step done, but not the second one:

After those 2 steps are both done, the exported result has comp texture working correctly:

Thanks @RomanDesign for sharing your findings on this. I had complained to Blender about this change. But... Its good to know the Node still exists in their system and can still be used. I am going to try this on my project and see if it exports.

Bill
LHC

@lyonhaart001 @WombiiActual @WombiiActual

I have spent ~4 hours with ChatGPT and finally got it to write a "fixer" script for this issue. I literally spent hours on doing it manually already, and then my whole material library file got the red nodes again somehow (from opening and saving in another Blender version? Or just for no reason?). Those were a couple of hundred materials and objects, so I decided to try ChatGPT instead of spending hours on fixing it and then possibly losing it again. Also, I had this happen in several other files, so this needed a more automated workaround.

Execute this in the file that has those nodes turned red with "Color" input and it will convert them to proper "Separate RGB" nodes with "Image" input and will reconnect it to the same places as the old ones.

It will also automatically remap extra "gITF Settings.00x" node groups to the default one and deletes unused ones.

I have to say ChatGPT is something else... It outputted a reasonable code based on my description of the problem and fixed and improved it based on my feedback until it was working fine. Took some time, but not more than explaining it to a live programmer. I don't have sufficient Python and Blender knowledge to write such a script myself. Main functionality was written in minutes, but the bulk of the time went into trying to fix the lack of connections of the new node to the same places as the old node. The ways it mapped it didn't work again and again. Until I instructed it to target specific nodes for connections, by names.

Basically, I can run this script once when my file is otherwise ready for export, and it will fix all the materials and gITF settings that bugged out during the work with the file, in one click. Now this makes it very easy. It would be still nice if this workaround wasn't needed in the first place, of course.

import bpy


# Initialize a counter for the number of instances found and replaced
count = 0


# Loop through all objects in the file
for obj in bpy.data.objects:
    # Loop through all materials in the object
    for slot in obj.material_slots:
        if slot.material:
            # Check if the material has a node tree
            if slot.material.node_tree:
                # Get the material node tree
                tree = slot.material.node_tree
                # Loop through all nodes in the tree
                for node in tree.nodes:
                    # Check if the node name contains "SplitOcclMetalRough"
                    if "SplitOcclMetalRough" in node.name:
                        # Create a new Separate RGB node
                        sep_rgb = tree.nodes.new(type="ShaderNodeSeparateRGB")
                        # Set the location of the new node
                        sep_rgb.location = node.location
                        # Copy the input connections from the old node to the new node
                        for input in node.inputs:
                            if input.name in sep_rgb.inputs:
                                for link in input.links:
                                    tree.links.new(link.from_node.outputs[0], sep_rgb.inputs[input.name])
                            else:
                                if input == node.inputs[0]:
                                    if input.links:
                                        tree.links.new(input.links[0].from_node.outputs[0], sep_rgb.inputs[0])
                                else:
                                    if input.links:
                                        tree.links.new(input.links[0].from_node.outputs[0], node.inputs[input.name])
                        # Create new output connections
                        for output in sep_rgb.outputs:
                            if output.name == "R":
                                occlusion_node = tree.nodes.get("Occlusion Multiplier")
                                if occlusion_node:
                                    tree.links.new(output, occlusion_node.inputs[1])
                            elif output.name == "G":
                                roughness_node = tree.nodes.get("Roughness Multiplier")
                                if roughness_node:
                                    tree.links.new(output, roughness_node.inputs[1])
                            elif output.name == "B":
                                metallic_node = tree.nodes.get("Metallic Multiplier")
                                if metallic_node:
                                    tree.links.new(output, metallic_node.inputs[1])
                        # Remove the old node from the tree
                        tree.nodes.remove(node)
                        # Increment the counter
                        count += 1


# Output the number of instances found and replaced in the console
print("Found and replaced", count, "instances of SplitOcclMetalRough nodes.")


# Find all node groups with "glTF Settings.00" in their name
matching_groups = [group for group in bpy.data.node_groups if "glTF Settings.00" in group.name]


# Find the "glTF Settings" node group
new_group = bpy.data.node_groups.get("glTF Settings")


# Remap users of the matching node groups to the new group
count = 0
for obj in bpy.context.scene.objects:
    if obj.type != 'MESH':
        continue
    for group in matching_groups:
        if group.users:
            group.user_remap(new_group)
            count += 1


# Print the number of node groups renamed
print("Renamed {} node groups".format(count))


# Delete unused data blocks
bpy.ops.outliner.orphans_purge()

@RomanDesign There is also a breaking change in blender 3.3/3.4
There are breaking changes to the node for ShaderNodeSeparateRGB has changed to ShaderNodeSeparateColor. Blender 3.4 has breaking change with ShaderNodeMixRGB to ShaderNodeMix.

Edit: the Code also changes its input and output numbering. What was Color1, Color2 is A, B inputs, but that will not work you must change input (1 and 2) to (6 and 7), and then output "result" was 0 and will now be 2

Enjoy the Blender fun.