ark@rabbit.UUCP (Andrew Koenig) (07/05/84)
A while ago I posted a summary of an article I read in AUDIO magazine several years ago. Briefly, the article described a carefully controlled experiment to determine if people could really hear the difference between power amplifiers. Listeners were given a box with two switches. One switched the signal going to the speakers they were hearing from one power amp to another. The other did nothing at all (except to interrupt the signal for a few milliseconds so it would sound the same as the other switch). The object of the game was to determine which switch was real and which was the dummy. The two power amplifiers being tested were very different. One was a basic power amplifier that has generally been considered to be well- designed and was well-received by the audio community (pardon the vagueness; I forget the make and model). The other was the amplifier section of a Lafayette receiver. Signals were set at levels that would not clip either amplifier. More important, the input signals to both amplifiers were carefully equalized to within 0.1 dB from 20 Hz to 20 kHz. The listeners were five "golden ears." Before the listening tests started, three of the five were confident that they could hear the difference between two power amplifiers. The other two were not certain they would be able to. After the listening tests, NONE of the five listeners felt they could hear any differences between the amplifiers, and NONE of them did significantly better or worse than chance would predict at guessing which switch did which. Things that were learned during the experiment: 1. Frequency response deviations of 1/4 dB were perceived, but not as changes in coloration. Rather, they were perceived as differences in 'depth,' 'imaging,' and so on. 2. Frequency response deviations of 1/10 dB were not perceived. These deviations were also smaller than the variations between two instances of the same make and model of amplifier, and also smaller than the differences within a single amplifier from one day to another. It was thus necessary to redo the equalization measurments each time the test was tried. 3. Once differences in frequency response and level were carefully eliminated, these two very different amplifiers did indeed sound the same.
ron@brl-tgr.ARPA (Ron Natalie <ron>) (07/09/84)
Well, we sat down with three supposedly good ampifiers one day to test out a friends new Acoustat 2+2 speakers. The way we told them apart is how they handled being driven to their rated output. The combination of distortion and clipping characturistics provided a noticeable difference even between two different models of Kenwood integrated amp. -Ron