ABX Console Listening Test

This website documents BaiLin He's senior research project for City Tech's Entertainment Technology Department, which was done with Faculty Adviser John Huntington.

Introduction

Sound consoles, or "mixers", combine and process audio signals and are important building blocks of professional sound systems. Within the industry, there is a lot of speculation on and opinions about the impact of mixers on sound quality. So to explore this issue, this research project compared the basic sound of two professional digital live sound consoles at very different price points: a Behringer X32 and a Yamaha CL5.

The project used a double-blind testing method called "ABX", where listeners were first given a recording of a musical performance through a known console designated A, and then a recording through a known console designated B. Listeners were then asked to identify a clip recorded through console X, the identity of which was unknown to the listener but is is either A or B, selected randomly. A professional guitarist, Howard Rappaport, played as consistently as possible each time, and the same mic and cable was connected to a randomly selected input on either console. Howard did not know which console was being recorded at any given time.

We did trials for two different mics, an AKG C414 set to cardioid pickup pattern, and an omni-directional Earthworks M30, both positioned as close as physically possible to each other. Apart from gain matching between the consoles for each mic (done at the input gain stage), no processing (EQ, compression, reverb, etc) was engaged in either console (SMAART measurements below verify this). The digital output signal was routed directly out over a Dante network to a Dante-enabled audio recorder. So, all this means that what we were comparing in the ABX test was only the consoles' input stages, A/D converters and internal signal flow.

Results

Using ABX testing, it is generally accepted that to achieve a 95% confidence level in the existence of an audible difference between two devices under test, a listener should be able to clearly identify the unknown X as either A or B at least 9 out of 10 times. 58 respondents took our survey, and listening to the AKG C414, only 1 respondent (less than 2%) correctly guessed 9 or more unknown X's. This person reported that they were listening on "2018 Macbook Pro Speakers", so it's most likely that this person just hit the correct guesses by chance. No one (0%) guessed 9 or more unknown X's when listening to the Earthworks M30.

Given the minuscule number of respondents who guessed 9 or more X's correctly, we feel--given our test setup--that there is no audible difference between these two consoles up through the A/D converter and internal bus structure. And this agrees with our measurements (below) which show a very flat frequency and phase response on both consoles up through about 3k, where the Behringer exhibited a slight difference in phase response. It's conceivable that different source material with more high frequency content could have made audible the this slight phase difference.

So is all this to say there is no difference between the $2,200 Behringer X32 and the $25,000 Yamaha CL5 Of course not. There is a dramatic difference in build quality between the two consoles, and the Yamaha is more likely to have more robust input circuitry. And there are likely dramatic and audible differences in the sound quality in EQ, reverb, dynamic control, summing, etc. To prove those differences, however, would require additional and different testing.

Last updated: April 20, 2019