Hello weekendwally,
Tests of endurance, sensitivity and maximum output levels in dB SPL of domestic loudspeakers are not performed with single-frequency test tones, but with what's known as a "broadband noise signal" with weighting applied. It's a simulated program signal (much like pink noise with equal energy octave to octave) but intended to simulate the bulk of energy content of a complex music signal.
One example used for years during tests of loudspeakers at the Canadian National Research Council had a uniform output between 200 Hz and 1 kHz, rolling off above and below those points at rates that resulted in -10 dB points at 25 Hz and 9 kHz.
If you wanted to try using a fixed-tone 1-kHz signal at extremely high SPLs like 120 dB, I'd suggest you use a big horn-loaded system with very high sensitivity (over 100 dB SPL, 1 watt/ 1 meter). Such speakers exist for concert sound reinforcement. They do not have flat frequency responses (unless carefully equalized) but typically distortion is very low and they are designed to be indestructible at such high SPLs.
Using a fixed-tone test signal at 120 dB with domestic Hi-fi speakers would risk damaging individual drivers. I'll check with Ian, the owner and designer of Axiom speakers, on the broadband noise signal used for our tests of the M80. The specification for the broadband noise signal used at the NRC used weighting according to IEC 268-1 (International Electro-technical Commission).
Regards,
