Why An SSB transmitter's peak envelope power is fixed by the PA and legal limit, so the only way to get more "talk power" is to raise the average power relative to that peak. A speech processor compresses or clips the audio, reducing the peak-to-average ratio of speech so more of the time the transmitter runs near full output. That raises average power and improves readability at the far end, especially for voices with a low natural average-to-peak ratio.
Watch out The choice about increasing peak power is tempting, but PEP stays the same, set by the transmitter and the 1500 W limit; only the average rises. Overdriving a processor actually adds distortion and splatter rather than reducing harmonic or intermodulation distortion.
Peak is capped, so processing fills in the valleys: same PEP, higher average = more talk power.