Another way of calling loudspeaker efficiency/ sensitivity is: dynamic compression.
When the loudspeaker has no dynamic compression, it is capable of giving all the amplifier put out, then it translates to 100% efficiency which is 112dB/Wm.
An ultra-efficient 102dB/Wm is 10dB down from that, that is: it has a 10dB compression already. It's not just peak SPL drop! It's the dynamic range being compressed.
Thus, using 102dB/Wm loudspeaker is like adding a 10dB dynamic compressor to the music. The 92dB/Wm speaker adds a 20dB compression, and a 82dB/Wm speaker adds a whopping 30dB dynamic compression! Now, combine that with the compression that is applied to the recordings, and no wonder the playback sounds so lifeless!
The dynamic compression of the loudspeaker is the primary reason we can tell it's recorded music, not live.
Going for higher power does not restore compression, so a 82dB/Wm at 100W is not equal to 102dB/Wm at 1W, although both produce 102dB/m peak volume: yet the range is 20dB diminished for the 82dB loudspeaker vs the 102dB one.
If you feel that you are hitting a brick wall with your system development, then you can try going for more efficient speakers. Be warned - the higher the sensitivity, the more the system weak links are exposed. Go above 100dB/Wm, and the exposure of weak links is absolutely ruthless. The clothes of the emperor are revealed....