AS mentioned, different mics have different sensitivities meaning preamp output level will vary for the same intensity of sound source at identical preamp gain.
To make noise graphs meaningful, dB output level from calibrated sound source input at same preamp gain should be noted for each mic tested, and graph should be adjusted so relative noise comparison can be made for each mic. Also, baseline plot of noise out of preamplifier loaded with ~100 ohms input at used gain should be shown on the graph.
Without these considerations such graphs are meaningless to show which mics are indeed quieter over bandwidth.
Also, Eric Benjamin graph shows huge room 60/120 Hz electrical noise pollution occurring so at least lower frequency plots of noise over bandwidth characteristic is also meaningless. These tests should be done inside 'screen room' environment that excludes such interferences.