Resolve
Sponsor: Headphones.com
Not exactly. While there's always the potential for very minor variation, two GRAS 43AG-7s will be compatible with one another when operated in a similar manner and the measured frequency response results for the same headphone will basically look the same - within positional variation tolerances. Most reasonably quiet environments aren't going to have an impact there, but they are relevant for readability of Harmonic distortion products. Among different measurement rigs, like a GRAS 43AG vs the older 45CA, vs a 4128 or a 5128, or even among some of the clone couplers, yeah then you can expect different results.
The measurement itself is like math, whether a sweep or an FFT, but once again its result is particular to the condition of being on the test 'head'. But, what I think you're getting at here is also something covered Hammershøi and Møller et al's Design Criteria for Headphones, and is also further discussed in Thiele's work that identifies correctness of 'tone color' (timbre) in DF HRTF.
Harman isn't really in need of much defending here since the measurement posted was not in relation to Harman (it doesn't exist for the 5128). But I seem to have to do this on a regular basis.
There are limitations to the statistical model that predicts listener preference, particularly that it doesn't factor in the ultra low or ultra high frequencies and also isn't able to accommodate fine-grained information to a sufficient degree. This means there are bound to be outliers. But for the over-ear listening experiments that were done, it is the most comprehensive public research we have on this topic to date. It would be great to have a larger study, sure... but this is the best indication we have right now of what the various segments of the population prefer (note that it's not singular to the curve that's commonly used, that's just what the largest group preferred).
Is the suggestion here that the headphone hobby is like homeopathy? You know you may be right there :|
It is a uniquely subjective endeavor, in that I agree... but so is shopping for shoes. Everyone has a different shape to their foot, different arch requirements, different size requirements and so on. But it still makes sense for shoes to be generally foot shaped in order for them to have a reasonable fit for people. Most people want their shoes to be generally foot shaped, despite also having individual fit requirements that also need to be accommodated.
So it is with headphones. People have all kinds of different preferences, and indeed different ears. But at the same time it makes sense for headphone's to have a frequency response at the drum reference point that generally takes the shape of a human HRTF in some respect.
The current limitations of headphone measurements is that they don't take into consideration the various preference windows around the reference curve, or other segments of the population that may have a different preference, so you get all kinds of confusion where people think these graphs are meant to be pure statements of goodness or badness... which they aren't, and this is something we're in the process of trying to remedy.
Yup. I always have to bring up Frank Jackson's What Mary Didn't Know, which is a fantastic paper about the importance of experiential information to what counts as knowledge.
5) That being said, if all reviews and all or most customer impressions align, you can assume with a decent probability that the product is good. (Still no guarantee you like it though...)
I think this is a very reasonable approach to making purchase decisions, along with an important caveat to remember. Just because others love/hate a thing, doesn't mean you will.
Last edited:
