When Measurements and Blind Tests Go Wrong: A History Lesson

The correlation between measurements and hearing is tenuous at best. If we use history as a guide, engineers in the 1940’s were convinced people preferred narrow band music reproduction. They’d done blind listening tests to prove it!

Thanks to Harry F. Olson, we came to learn that what people were objecting to with wider-band reproduction was mostly, in a word, crappy sound reproduction.

source: The Experiment That Saved Hi-Fi

Olson was an engineer for RCA with a specialty in acoustical engineering and he set about designing a test to prove his theory. This test, unlike older tests which used the hifi’s of the day [footnote 1], placed a small orchestra out of sight from the 1,000 test participants and added an acoustic filter between orchestra and audience, consisting of metal sheets with various hole sizes which effectively cut off frequencies above and below select points. In a nutshell, 69% of listeners preferred the orchestra full-range, finding the filtered sound unnatural.

Olson’s test disproved previous test results, showing they’d reached faulty conclusions because the methodology was flawed.

Sometimes it pays to test outside the box.

  1. Eisenberg, P., & Chinn, H. A. (1945). Tonal range and Vol. level preferences of broadcast listeners. Journal of Experimental Psychology