Tuesday, October 10, 2017

Learning while listening: RMAF 2017 part III

The best experience for me was hearing Paul Barton, chief of PSB Audio, deliver a seminar on Saturday morning. The emphasis was on his use of the acoustic labs at NRC Ottawa to test his designs on listeners. While the talk aimed at headphone design particularly, much of it was based on prior research on listening preferences with loudspeakers. The real emphasis here was on the scientific method, and use of blind tests to establish listening preferences for various response curves, with particular details provided of a simple experiment involving headphone reviewers and their reactions to curves, including the one employed in the PSB RoomFeel technology.

The seminar covered a lot of ground, not least the value of testing, the importance of preventing participants from seeing what they were listening to and from communicating their reactions to other participants. A part of the seminar covered breakthroughs in modeling and tools for measurement of the human ear  (the physical interface a headphone makes with the listener seems paramount to perceived quality) and a lively Q&A session at the end had us exploring how universal the preference for a tapered frequency curve might actually be. I learned that it has been long known that humans generally prefer a particular frequency curve when left to their own devices, which means they do not like ruler flat,  but that this preference can be distorted in early listening impressions, only to emerge  over time as listeners gain some ability to shape the sound to their preference.  Yes, there is hope your teenager will outgrow those Beats!

Naturally this had me asking, if the preference is so universal, why do we see such variance in gear at RMAF? Ah, sighted listening, a preference for a particular look, influence of others etc, all mentioned as reasons. We humans, it seems, might have a preferred response curve when all other stimuli are removed, but that's not actually the world we inhabit or make purchasing decisions within. Fascinating stuff, if only to push me harder to understand more of the science of listening preference (no, not the engineering of loudspeakers, what people actually experience).

Of course, I thought a lot about the estimable Paul Barton's words as I wondered the rooms later that day. Is it any surprise then that I found great pleasure in the Revel room where a new $10,000 pair of speakers (F228?), due to be released early next year.  Revel's rep was pretty clear in stating that their speakers are always built on the basis of scientific, blind listening tests, and a new model is only released when it clearly improves on existing models. Hence, the Ultimate series, now 10 years old, remains top of their tree as it's the best they can build currently. This new model ups the ante on the previous F series apparently and they have the data to prove it!

I visited the room more than once to check it out and have to say, it would get my vote as one of the best sounds at the show. Obviously conditions are less than ideal, and I hate the white plasticky look but the speaker had an impressive coherence and very smooth sound to go with sufficient resolution to reveal lots of detail on a range of tracks. The triumph of science?


Would say the same seems true of the Bryston gear. A pair of snazzy white floorstanding (is the color a code?) Model Ts in a a cool active system (exernal crossover before power amps feeding the various drivers). Sounded smooth, coherent and inviting, so I took some time just sitting there. One attendee behind me remarked something to effect that this was the kind of 'buy it and be done' system that you spend $25k on and then forget about gear thereafter. Of course, much as audiophiles might love that idea, I think few actually do it. But in spirit, I have to agree. Is it a coincidence that James Tanner of Bryston is another user of the NRC labs for testing speaker designs?

No comments: