I’d be surprised if the vertex color is fully redundant, just because game devs don’t do that-- they optimize the hell out of everything they make, and anything that’s left, is for a reason. I wonder if the vertex color is intended as occlusion for dynamic lights.
I also think the choice of “100” as a necessary emission strength is probably not correct, because 100 is a weird number that blows everything out and leaves you with poor precision from 8 bit-per-channel textures. And hue is probably not something that should need correction if everything else is right. Although I can’t imagine what would possibly affect it very strongly. Color management (filmic etc) can affect it, but only by a little bit.
Your general philosophy here sounds pretty wise: lightmap * diffuse -> emission (leave vertex color out of it), with light sources given plain emission meshes, but there are a lot of dimensions to tuning it beyond the strength of the emission. If you have a method that works well for most bits but not this one bit, it could be that this one bit is different, or, it could be that the method you had wasn’t right. In the case of emission, the other dimensions of how emission is going to look are going to come from bloom settings and color management.
However, the fact that this mesh has no textures does suggest that it has something special going on in the shader-- something weird that’s going to require a different paradigm for it. You said that it has a vertex color layer but no textures, but in your image above, you have a light.tga for it-- does it have a lightmap?
I don’t know if it’s the case, but it’s not impossible it’s intended to use some kind of multiplicative blend (instead of add or alpha blending). Multiplicative blends tend to be more saturated than additive or alpha blends. Here’s a quick multiplicative blending demo: