Monday, January 6, 2014

Do Listeners Agree on What Makes a Headphone Sound Good?

This past weekend, I attended the ALMA 2014 Winter Symposium in Las Vegas where I gave a talk entitled, "The Perception and Measurement of Headphone Sound Quality: Do Listeners Agree on What Makes a Headphone Sound Good?" The presentation gives a summary of some key findings of our headphone research conducted over the past 18 months. It also includes some unpublished preliminary findings from a current study on headphone preferences of trained and untrained listeners both young and old from China, Canada, USA, and Germany. The focus of this study is to determine whether listeners from different cultures and age groups prefer the same headphone sound quality as trained listeners when the influence of headphone brand, fashion and celebrity endorsement are removed from the test. 



The abstract for my talk is reproduced below. A PDF of the slide presentation can be downloaded here:

The popularity of headphones has now exploded to produce annual worldwide sales of almost $10 billion. Premium headphones ($100+) now account for 90% of the annual revenue growth, as consumers’ audio experiences are becoming a primarily mobile one. Market research indicates sound quality is a driving factor in headphone purchases with brand and fashion also being important factors among younger consumers. Yet, ironically the science behind what makes a headphone sound good and how to measure it is poorly understood. This combined with the lack of perceptually meaningful headphone standards may explain why purchasing a headphone today is like playing Russian Roulette with your ears. The magic bullet to achieving more consistent headphone sound quality is science.


We recently conducted a series of controlled double-blind listening tests on popular headphones (both real and virtualized models) to better understand the relationship between their perceived sound quality and acoustic performance [1,3,5]  A second set of experiments measured listener preferences of different headphones equalized to different target curves responses including the recommended diffuse and free-field target curves [2].  A third set of experiments used a method of adjustment where listeners directly adjusted their preferred bass and treble levels of a headphone and loudspeaker equalized to the same in-room target response [4]. In this way, we could measure the variation in individual listeners’ taste in headphone spectral balance, and determine the extent to which the preferred headphone target response should simulate the response of an accurate loudspeaker in a reference listening room.


Together, the results of this research show that when the influence of brand, fashion and celebrity endorsement are removed from headphone tests, both trained and untrained listeners regardless of age and culture, generally agree on which headphones sound best and this correlates to their acoustical performance.


References
  1. Sean E. Olive and Todd Welti, "The Relationship between Perception and Measurement of Headphone Sound Quality", presented at the 133rd Audio Eng. Soc. Convention, San Francisco, USA, (October 2012).
  2. Sean E. Olive, Todd Welti and Elisabeth McMullin, "Listener Preferences For Different Headphone Target Response Curves",  presented at the 134th Audio Eng. Soc. Convention, Budapest, Hungary, (May 2013).
  3. Sean E. Olive, Todd Welti and Elisabeth McMullin, "A Virtual Headphone Listening Test Methodology", presented at the 51st Audio Eng. Soc. International Conference, Helsinki, Finland, (August 2013).
  4. Sean E. Olive, Todd Welti and Elisabeth McMullin, "Listener Preferences for In-Room Loudspeaker and Headphone Target Responses"  presented at the 135th Audio Eng. Soc. Convention, New York, USA, (October 2013).
  5. Sean E. Olive, "Do college students prefer the same headphone sound quality as trained listeners?", presented at the 4th ISEAT, Shenzhen, China, (November 2013).

19 comments:

  1. Great presentation. Thank you & happy new year!

    ReplyDelete
  2. I just looked through the slide deck. Even after all this time of following (and enjoying the results of) your work, I'm still impressed at how much money and scientific rigor Harman devotes to making sure I'll actually like the product you produce. Amazing!

    ReplyDelete
  3. Thanks for publishing this, Sean! I really appreciate that Harman sees this kind of public publishing as a worthy activity, in addition to the research itself. Really top notch!

    ReplyDelete
    Replies
    1. Thanks Kevin. I'm lucky that Harman allows me to publish research as most audio companies don't. That's why I've been here 21 years in march.

      Delete
  4. Keep marching forward. Great work Sean! Behind you all the way!

    ReplyDelete
    Replies
    1. Thanks Tyll. I have a great iPad app I'd like to show you soon.

      Delete
  5. Which is the replicator headphone? Senn 515 or other?
    If it's 515, why did you choose such a poorly performing set(distortion, air pressure leakage) to emulate top quality sets? Shouldn't you use something like Stax?

    ReplyDelete
    Replies
    1. In this study, the replicator phone was a Sennheiser HD518. We chose it because it has good repeatable fit among listeners and it is relatively smooth and easy to equalize. Stax does have lower distortion but I can't say whether it's fitment is better or worse.

      Delete
  6. This is the type of work that only a large and diverse audio company could support, and then only with a gifted and experienced scientist at the helm. And then allowing that scientist (or engineer, if you prefer) to publish findings that seem obvious to a few and probably heretical to others is extremely rare. Typically there are too many administrative and legal barriers between the data and the outside world. Sean, in particular, and HK in being supportive, have done the audio world, consumer and otherwise, a great service over the years and especially in terms of the results reported here. Oh, and please keep up the blog as well.

    ReplyDelete
  7. Very helpful post - as usual. Thanks so much! I hope this one goes viral. Market research indicates sound quality is a driving factor in headphone purchases with brand and fashion also being important factors among younger consumers.Headphonezone.in offers best headphones for superior listening experience. It has different range of color that suits your mood.

    ReplyDelete
  8. This info you provided in the blog that was really unique I love it!!! headphones for tv

    ReplyDelete
  9. These blogs are valuable because these are providing such informative information for all the people.

    headphones for tv

    ReplyDelete
  10. Thats lot of research & information! The kind of headphone/speakers always affect the music's impact! its like first impression is the last! I found some great Headphones Online

    ReplyDelete
  11. Sean,

    Could you explain why these exact EQ types were used for bass and high frequencies? Bass - a shelving eq to ~200Hz; High - a roll off from ~400Hz.
    Why not use a Baxandall EQ (as suggested by Bob Katz for mastering applications)?

    ReplyDelete
  12. You tested listeners' prefered equalization for the Sennheiser HD 518.

    I think you should try to replicate this result with other headphones.

    And control for previous exposure to loudspeakers.

    But I cant really believe in blind tests of headphones. Loudspeakers maybe but not headphones.

    ReplyDelete
  13. Great Article, very informative and very unique! Thanks Sean

    ReplyDelete
  14. What makes headphones to sound good ? In my personal case I found it relatively easy to answer. When I decided to get as good headphones as possible I did not walk on the 'keep buying until you find what you like' road. I carefully thought about the audio characteristics of headphones I would like to have, in other words I constructed a pre-conceived notions of what my headphones have to sound like. Following this I simply read a lot of reviews to find out which headphones would generally fit my 'headphones profile' and then I tried them to find out which ones have the best sound signature and sound quality fit. [It helps greatly if one has an opportunity to hear and evaluate headphones before buying them]. Beyerdynamic T1 won the evaluation contest as the best over-all open headphones and I ended up buying them. For my second pair of headphones, in this case the closed ones and complementary to open T1, there was practically no contest because there are very few closed headphones of top quality on the market. I found Audio Technica W3000ANV to be the best closed headphones.
    A very interesting and most influential factor in deciding what kind of sound my perfect headphones have to have were my religious and political beliefs.

    ReplyDelete
  15. Hi Sean

    I am still waiting for the first commercial headphones that implement the preferred response coming from your research. Presumably it will be a Harman Group product. Is it the new AKG N90Q?

    ReplyDelete