IF the disk continues to rotate at the same frequency, then you get an I/O increase exactly proportionate to the increase in linear (not areal) density. A new technique which puts twice as many bits per inch of track should read twice as many bits per second.

Write techniques may or may not need to be adjusted (preheat a track? a sector? realtime per bit-domain?) but should eventually converge to the same speed-up as reading.