Yes i understand this workaround, but it will work only on simple pattern, if i want to animate the falloff, wich is the point here, on more complex pattern it will become not really viable as we would nedd to bake the color animation each time we change one thing :/
That's why i would like to be able to use this workflow, wich is a legit one not some modo under the hood crazy unstable setup^^
So if there is no way to control color independantly on a scatter node, can we take it the other way, treat the replicator as render time with an option like don't render as instance and make the plugin create real copy, and get the color automaticaly from the shader Inputs node ?
I know for large replicator system it will increase a lot the processing time, but with simplier mesh and reasonnable number of copy it could work without filling all the gpu's memory.
I'm ok if there is a modo or octane limitation, but i just want to know it and why, i'm the kind of guy who like technical precision and not just 'it won't work'
PS: i swear i'm a nice guy don't take my posts too badly !