What is the approximate ratio of PEP-to-average power in an unprocessed single-sideband phone signal?
Why SSB voice is a varying-envelope emission: the RF envelope follows the speech waveform, so the power swings from nearly zero between syllables to full peak on vowel peaks. Averaged over normal speech, the peak envelope power runs roughly 2.5 times the average power for unprocessed audio. Speech processing (compression/clipping) raises the average toward the peak, lowering this ratio and putting more talk power on the air.
Watch out The 1-to-1 choice describes a constant-envelope signal such as key-down CW, FM, or an unmodulated carrier, where peak and average power are the same; 25 to 1 grossly overstates the peaky nature of ordinary speech.
Unprocessed SSB is about 2.5:1 PEP-to-average; add compression and it moves toward 1:1 like FM.