The noise measurements I have been doing utilize a Tektronix or Boonton distortion analyzer -- the Boonton is "in cal" and for sine waves the TEK and Boonton mate perfectly up to 100kHz -- both use the Analog Devices AD536 log converter chips. The Boonton uses a 2.2uF timing (averaging) capacitor and the TEK 4.7uF -- so there are some obvious differences when measuring low levels of random noise.
The noise measurement articles (noise from regulators, power supplies, batteries, diodes, references) are bandwidth limited -- sometimes 100 milliHz to 10 Hz, sometimes across the audio band, sometimes to 100kHz -- the question is "which horse for which course?" -- and is it reliable to measure noise when a log converter is used (or the new delta-sigma converters from Linear Tech) or should I rely instead on the thermopile of a defunct HP 3403C?
FWIW, Linear no longer manufactures the LT1088.