Impulse responses do not react because they are not dynamic. A speaker IR is a static snapshot of a speaker and mic at a given amplitude. A speaker's behavior changes as the amplitude goes up and especially so for speakers that provide distortion characteristics. IR's are a terrible way to present amps because they do not react with amps, where speakers do react via a dynamic impedance relationship as well as via speaker distortion (which IR's don't do!). It's the same reason why room IR's don't sound like convincing dynamic rooms that react differently according to volume in the room. They are static reverberation snapshots, and they sound like it. IR's can still be used as tools for given purposes, but they are not adequate replacements for real physical reactive spaces, speakers, and microphones. In case it isn't clear what I'm saying, here is an example. If you could heavily attenuate an amp to get essentially the same output curves for any practical given input signals as would happen at high volume, that amp will still sound different* give different output curves through the same greenback at low volume vs. at high volume, because the speaker will compress and distort at high volume. And when someone plays dynamically, there are essentially volume changes taking place across the audible spectrum. Replacing the greenback in this scenario with a greenback IR, the IR only has one curve by it's very nature, reacting in the exact same way no matter any volume changes via attenuation or dynamics. This is why IR's sound less dynamic, because they are not dynamic at all! In other words, soft picked notes get exactly the same speaker filter curve as hard picked notes, because an IR is only a single curve. Dynamic response will eventually be brought to the speaker IR world, likely via multiple curves from low volume to high volume with interpolations filling the inbetweens and some sort of distortion algorithms. But I haven't heard of anyone doing it to date, likely because it is a complicated thing to pull off. This is also why manufacturer speaker curves don't tell us much. Because we are only seeing a single curve at a given amplitude! That single curve can't show how a greenback's response will change and distort at high volume vs. stay clean at low volume, compared to a high power speaker which will react differently at high volume vs. low volume and stay pretty clean throughout the amplitude range. * Human hearing perceives sound differently according to volume. See: Fletcher Munson curves for example. I should also say here that for super distorted djent type amps and players, IR's might well be perfectly acceptable, because they aren't using anything near as much dynamic range as say an edge of breakup classic rock player.