On the origins of the STC curve, I did a little hunting in the JASA archives. One of the better papers I found was:
Sound Insulation and the Apartment Dweller by T.D. Northwood in JASA Vol 36, No. 4, April 1964.
It gives a little background, plenty of opinion, and even delves into NC curves, which - like STC curves - were relatively new back then, too.
It would appear that other references point back to the committee that developed ISO 717. It seems the intent, at the time, was that the ASTM committee was basically following the ISO committee's lead on sound isolation rating. It wouldn't appear that it fully happened that way. When/why STC diverged from Rw might be something we never know. (Or do we know? Maybe it's just me? :? )
if you read the standard that teaches how to calculate STC - ASTM E413 - you find some itneresting things. It states the intended utility of the single number rating
"These single-number ratings correlate in a general way with subjective impressions of sound transmission for speech, radio, television, and similar sources of noise in offices and buildings. This classification method is not appropriate for sound sources with spectra significantly different from those sources listed above"
I guess that basic commentary and the very-tempting nature of the strong correlation between these:
-Aweighted noise reduction, in dB, for the frequency spectra from 125-4000hz for a noise source that is flat in all bands
lead to my quasi-conclusion.
In no way am i a history buff, and i hope my statement above was properly disclaimed as my own hunch.
I would be curious to know exactly how that atrocity came into being.
Alot of the reason why STC has such a high frequency of cutoff (125hz) may relate to the general assumption in ancient times that low-frequency data would deviate much more greatly from lab to lab than would high frequency data.
I would strongly suspect that isn't so, based on personal experiences and an analysis of NRC, Orfield, Riverbank, USG and other labs datas over the years.
There are IMMENSE mid/high freq variations amongst the labs, even when flanking limits cannot necessarily be given as the cause, and the low-freq variatiosn aren't necessarily larger.
Now, for the reference panel, a non-resonant simple mass-law panel, i would guess that mid/high frequency variatiosn from lab to lab would be minimal, while low-freq deviatiosn might be large.
.... a typically tested double leaf wall is no mass law panel, and atl ow frequencies the radiation behavior is completely different - typically pistonic resonance/decoupling behavior - and i believet hat it is the difference in radiation behavior at low-freqs that would cause typical double leaf walls to be MUCH CLOSER to each other among different labs than would simple panels. At least until below the MSM region, where wild deviatiosn would again occur.
thist opic - testing for TL - gets me a little hot and bothered, so i better shush up now.