But from the perspective of sensitometry, we're assuming no light scatter, right? No non-image forming light/flare. So what I'm really thinking about is chemical fog. "Signal density" is a pretty good term for useful density. I like that. Ideally that's the density we'd want to plot against log H. In Zone System testing, Net D at any exposure level (ie gross density minus constant fb+f measured at zero exposure) is assumed to = signal density. Probably good enough, but I just find it an interesting assumption. I guess there isn't really a way to isolate signal density from whatever the chemical base fog density is at any exposure level anyway.