The absolute one is the one that produces the sound that is inherent to the speaker alone. Many variables impact the sound a person actually hears, one of them, where ir's are concerned might be room reflections. I am suggesting that a speaker has an eq curve, period. That eq curve may be altered by the cab, room, mic, listener position, ect. We can capture some amount of that information (I am assuming we currently do) in which case you are never dealing with just a pure speaker curve like your tube amp does. A tube amp reacts with the speaker, in a cab, and what you hear depends on many factors beyond these basic items. But it starts with the essential parameters of the speaker and cab. If you capture other nuances with an ir, then use that ir as your freq curve, you are not really comparing apples to apples are you? If the amp modeling is perfect, and if the speakers reproducing the sound are perfect, your ir has caused some shift in sound because it contains parameters that are different from the actual eq curve of the speaker. Now of course nothing is perfect, but the premise seems plausible to me.