Can someone give me an explaination re: power amp.


So, I own a simaudio i-5 driving paradigm studio 60 v.3, and I am not happy with the sound at low volume. turn it up everything is great, but low volume has muffled sound. Everyone says the paradigms are efficient, but i go out looking for more efficient speakers, and happen to come home with a loaner power amp- a simaudio w-3 (125 wpc), and the sound at low volume sounds markedly better with the power amp. this is contrary to everything i have convinced myself to believe. i have convinced myself that a good integrated is all i need, but I cannot deny the sound. so instead of asking the age old separates vs. integrated question, i'm just going to ask- why does my system sound so much better at LOW volume with the power amp. Maybe the speakers aren't as efficient as I thought? Maybe the 70 wpc simaudio i-5 isn't as great a machine as i thought? maybe there's a technical explaination? Please enlighten me! and thanks, and merry christmas to all you audiophiles out there! (oh, and the reason I didn't come home with speakers is that i really love the full range of sound from the towers, and i couldn't find any highly efficient speaker that had the range I enjoy. not that it dosen't exist, but i just couldn't find it. I live in portland, oregon).
blazerfan
Bob - I looked at two latest Stereophile reviews of tube amps and I don't see "u" shaped curve.

First: http://stereophile.com/solidpoweramps/moscode_402au_power_amplifier/index4.html

Second: http://stereophile.com/tubepoweramps/hyperion_ht-88_monoblock_power_amplifier/index4.html

In case of the second (Hyperion) it is straight line with lowest distortion at min power. That would imply that u shaped characteristic of SS distortions is caused most likely by nonlinearities (possibly output transistors) around zero. What do you think?
Post removed 
Bob - It might be THD +S/N or some nonlinearities of output transistors in spite of feedback (there is no perfect regulator). I don't know if THD can be measured without noise. Traditional THD meter deliveres pure sinewave to the input and subtracts same sinewave from the (scaled) output measuring what left - harmonics + noise. THD + noise levels are really small - 1/100 of percent = 1/10000 (-60dB) of very small signal (0.1W). It is equivalent to about 1uV at the input. Tube amps, I looked at, had higher THD to start with, so this effect is probably hidden.
If you're after efficient speakers, try the Auditorium series speakers from Living Voice. These speakers are an easy load for amplifiers to drive; sound great at low volumes as well as at higher volumes.
Kijanki, some tube amps will make higher distortion at very low power levels just like transistors. Others (SETs for example) the distortion becomes unmeasurable.

Feedback is appearing more and more to be a poor means of ridding audio circuits of distortion. An amplifier that might not have had much beyond the 5th or 6th harmonic will suddenly have harmonics beyond the 81st when feedback is applied.

Much depends on propagation delay and how effective the circuit was as an amplifier before the feedback was applied.