is there a way not to get these speckly shadows on instrument dials?

B21-soaring · 27 Feb 2023

As picture above, it's common for our planes to have quite speckly shadows on the instrument dials - I've adjusted the time/clouds/sun to make this example as extreme a I can but the speckle is usually there if you look carefully enough. Anyone else seen this? Is there an obvious correction/solution?

Gracias
B21

=rk= · 27 Feb 2023

Alpha transparency mode dither?

=rk= · 1 Mar 2023

No bites? Ok, check your transparency settings for the affected material. It looks like you are using a transparent texture with smudges, or the impression of dirt on it. This is important because it is not entirely transparent. I am guessing that the same texture that is used for solid parts, also has the transparent glass parts and setting "dither" solved a problem of ghostly transparency for the affected solid parts. You will see this a lot in models that have an integrated fuselage/canopy texture, incorrectly configured, the fuselage will be semi transparent.

Regardless, if this is a dither issue, the dither setting has a cut off level, defined by the intensity of the target color. If this color is part of the "dirt" then the dirt appears as solid specks, it has been "dithered in." Am I close?

B21-soaring · 3 Mar 2023

=rk= said:
No bites? Ok, check your transparency settings for the affected material. It looks like you are using a transparent texture with smudges, or the impression of dirt on it. This is important because it is not entirely transparent. I am guessing that the same texture that is used for solid parts, also has the transparent glass parts and setting "dither" solved a problem of ghostly transparency for the affected solid parts. You will see this a lot in models that have an integrated fuselage/canopy texture, incorrectly configured, the fuselage will be semi transparent.

Regardless, if this is a dither issue, the dither setting has a cut off level, defined by the intensity of the target color. If this color is part of the "dirt" then the dirt appears as solid specks, it has been "dithered in." Am I close?

thanks for these clues - pls excuse my slow acknowledgement - I'm currently discussing with the actual modeller who's relatively inexperienced. I'm doing all the programming/configuring and have necessarily become quite adept at reading the gltf file but I need clarify with the modeller exactly what's he's done here. In the gltf the "textures" that are listed all reference (41) dds files and none of those seem relevant for glass on a gauge. The most likely "materials" clues in the gltf are multiple entries like this I think:


        {
            "name": "Transparent.003",
            "pbrMetallicRoughness": {
                "baseColorFactor": [
                    0.800000011920929,
                    0,
                    0,
                    1
                ]
            },
            "extensions": {
                "ASOBO_material_invisible": {}
            }
        },

Generally if there's something I can change in the gltf that makes the problem appear/disappear I'll try and get confirmation there before risking a model change.
Many thanks for your advice.

=rk= · 3 Mar 2023

Ah, no problem, totally understand the issue. I was in the exact same situation a few years ago and had to interpret between modellers unaware of MS flight sim nuances and a fairly prominent developer.
I really don't recommend direct editing of the gltf file, unless absolutely necessary. There are a few ways to address the issue, but I think even the simplest one, that isn't necessarily the most efficient, of just using two instances of the same material, one instance transparent in blend mode for glass, one instance opaque for the other parts, would add a draw call, not require editing of geometry, so conceivably could be done via editing the gltf, but would probably be too complex and more reliably performed through the export settings of the 3d modelling software.

Nervegas · 4 Mar 2023

=rk= said:
Ah, no problem, totally understand the issue. I was in the exact same situation a few years ago and had to interpret between modellers unaware of MS flight sim nuances and a fairly prominent developer.
I really don't recommend direct editing of the gltf file, unless absolutely necessary. There are a few ways to address the issue, but I think even the simplest one, that isn't necessarily the most efficient, of just using two instances of the same material, one instance transparent in blend mode for glass, one instance opaque for the other parts, would add a draw call, not require editing of geometry, so conceivably could be done via editing the gltf, but would probably be too complex and more reliably performed through the export settings of the 3d modelling software.

Would the ALPHAPRESERVATION flag be of any use in this instance? I havent come across any issues with alpha dithering to this degree, so I was just curious in case I need it in my dev toolbox lol.

=rk= · 4 Mar 2023

It is such a simple thing to define transparency and lock that in with a blend mode, that I am really hesitant to modify it with a flag and we see the SDK cautions against using it's group of flags over performance drains. I doubt a flag that clamps the pixel transparency ratio in mipmaps, is going to be extremely useful to a modeller's career. What this means to accomplish, is that as a bubble canopy P-47 is moving away from you, any mipmaps that are serving as alpha masks, might lose ratio and cause the canopy/fuselage transparency to cross bleed, which the flag would prevent, within the limitations of the mipmap itself. Bear in mind, this flag is intended to address deficiencies that could be prevented, I cannot see how this would be of pressing issue to anyone beyond the most fastidious.

sal1800 · 5 May 2023

This has the appearance of the Shadow Sphere being in the incorrect location or not covering the entire cockpit. It should be in the cameras.cfg. This controls how the shadow maps are generated.