You got me to look up what TCR (temperature coefficient of resistance) is, and found this JBL patent. So, for a 2x difference in TCR, only when the voice coil temperature exceeds 200 °C, SPL over 120 dB (1 m, free field), the difference was only over 1 dB, roughly the just detectable level. (The 1 dB level is just my guess for 120 dB SPL. It is more applicable for the healthier SPL levels.) The difference didn't exceed 2 dB when the curve stopped at ~520 °C (melting temperatures of the common aluminum alloys start at the high 500's °C, depending on what elements are in the alloy).
I doubt it "matters the most" to most people.
View attachment 556460