1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

Testing audiophile claims and myths

Discussion in 'Sound Science' started by prog rock man, May 3, 2010.
879 880 881 882 883 884 885 886 887 888
890 891 892 893 894 895 896 897 898
  1. castleofargh Contributor
    they did many things, it's too bad that the papers were only available for free a short period of time after each was published. but you can probably find Sean Olive talking about some of it in videos.
    they tested age and found no significant difference, they went to several continents to see if there could be cultural or genetic differences and apparently didn't find any(with the caveat that they didn't go test isolated population deep in the amazonian forest or that kind of stuff). I had many critics(and praises) about their job when they presented the first papers, because I thought that was it and it was clearly incomplete. but as the years went by and they kept testing new stuff depending on the results of previous tests and questions they came up with(you know, real science). so I can say that it's probably the most extensive and serious study about subjective preference of headphone signature. I doubt any internal study by any big firm went as deep as they did with as much rigor.

    I don't remember specifically when the listeners could pick the songs or if the choice was limited in genre, so I don't want to tell you something wrong. but I would assume that they have answered that type of problem at least a decade or 2 ago with their extensive research on listening tests with speakers. my guess would be that there is a matter of music genre involved but that it falls under the vast mystery of audio's circle of confusion, where at some point people master a certain type of music because they expect it to be played mostly on certain types of systems, and soon enough people start thinking that they need a system with a specific signature to best enjoy that genre, close the loop and enjoy ^_^.
    GearMe, WoodyLuvr and gargani like this.
  2. gargani
    Yes, I saw that. Thanks to yourself, castleofargh, and bigshot for your inputs.
    I now understand the theory and process used to derive the harmon curve.
    WoodyLuvr likes this.
  3. bigshot
    No, the opposite is true. Music is mixed on calibrated speaker systems in a room. The Harman Curve is an attempt to duplicate that experience as closely as possible with headphones.

    The thing people are missing is that every set of headphones you buy already has a response curve. If you apply a Harman Curve to that, you're stacking "mountains on top of mountains". The curve of the headphones is being added to the Harman Curve. In order to apply the Harman Curve, you first have to EQ the cans flat. Then you know you aren't adding or subtracting anything because of stacking two curves on top of each other. No headphones are flat out of the box. It's a two step EQ process... 1) achieving flat response in the headphones 2) applying the Harman Curve to that.

    Another misconception is that a response curve can be "perfect". Any target response is just a starting point. It isn't a destination. Every person's ear is different, every set of transducers is different, every room is different. A studio environment can be calibrated pretty tightly because the room is designed and built from the ground up for that specific purpose. A normal living room can't because it's designed for a completely different purpose. So for a home, you make compromises and adjust the target curve to fit your particular circumstances. A little more salt here... a little more pepper there... until you the listener is satisfied.

    There are some things that can't be strictly objective because everyone's hearing and situation is different. But if you start with a target curve, you are beginning from a firm calibration. If you wander off too far, you can always snap back to calibration and try something else. Eventually you arrive at a response curve that is perfect for you, and that becomes your own target response.

    I find that the Harman Curve is almost identical to my own personal target response. A lot of other people feel that way about it too. That makes it a good place to start.

    I'm going to post this another time and point out something I didn't mention before...


    If you notice, that green line is the response curve of a calibrated speaker system in a room from the main listening position. Notice how closely it follows the target. The difference is only in the bass. I think that is because in a room, bass flows across the floor like water and comes at the listener from all directions. With headphones, the bass is contained, so you have to boost the level a bit to make it feel as enveloping as it would in a room. That is just a guess though. I might be wrong.
    Last edited: Jul 14, 2019
    WoodyLuvr likes this.
  4. TheSonicTruth
    Sadly, one of the headphones whose raw response closely matches Harman - NAD's VISO HP50 - is discontinued after only five years on the market(it debuted in 2013). At less than $150msrp, it seemed like good competition to the venerable MDR-7506.

    Wish I'd heard about this can sooner. Coulda-woulda-shoulda!
    Last edited: Jul 14, 2019
  5. bigshot
    my cans that are very close have been discontinued too. Good reason to EQ
  6. castleofargh Contributor
    I guess it's a matter of defining "close". if by that you mean the general shape beside the low end, then sure it's very similar. but then we had that with the diffuse field compensation that's been used for headphone graph for so many years before. it also "closely follows" that general shape.
    the problem being that almost everybody I talked to over the years(myself included because I talk to myself a lot ^_^), disliked something tuned to DF target and thought it had clearly too much amplitude in the upper freqs. so if that small of a change(if we pretend not to see the massive bass boost) can move opinions from "meh, and too bright", to apparently the preferred response on average for people, maybe it shouldn't be treated as a small change? I know I'm clearly nitpicking here, but I feel like I should.
  7. bigshot
    Any target curve is a generalization. Different people will want to adjust to their own taste. It's allowed. Modifying target curves isn't a mortal sin, thank goodness!
  8. SilentNote
    It seems to me the diffuse field curve successfully characterizes the amplification done naturally by our pinna, but failed to take into account of the amplification of the listening environment. What Harman is doing is accounting for the bypassed listening environment as well. This obviously has a lot of subjectivity in it as there is no objectively correct listening environment, and thus is preference based and a statistical average is taken.

    Still, everyone's pinna and "expected" listening environment varies so nothing beats auditioning with your own ears if you are looking for that magical pair of IEM for yourself.
  9. castleofargh Contributor
    yup, the big issue with a purely objective approach for headphones or for that matter IEMs, is that we have to assume that the tactile bass from sound in natural environments or in speakers isn't a factor. because if we say it is, then the solution is to actually provide tactile vibrations to the body. and we have to figure out, how, how much, if only a part of the body is enough and if not if a boost will compensate for missing the rest of the body, etc. well we already know that basically any physical vibration even at the wrong frequency does give an impression of more bass presence. but figuring out the rest isn't simple, and would require ultimately to accept a sort of standard for shakers' implementation. or just for the sake of determining a curve for headphones, use a woofer with the headphone, properly figure out the upper frequency and then go crazy over how to simulate the low end better? IDK
    the other elephant in the room IMO, is how we listen to stereo albums mastered on/for speakers. so even if we bother with some ideal room and just the right amount of reflection, and sources at 30° when using a nice dummy head, in the end if the music we listen to on headphones, doesn't send the left channel to the right with the proper delay and correction as speakers do, then that reference loses a good deal of its purpose. because the subjective difference is not trivial. if we had the all thing identical to speaker playback, then of course I would expect the preferred sound on speakers to also be the preferred sound on headphones. but that's not what we're working with in practice, so the objective approach has to work on many assumptions and IMO turns out to be fairly incomplete. which is why I appreciate how Harman went for listening preference instead of trying to go pure objective approach on that one.
    but of course at the end of the day I also EQ to my own taste and don't really care what other people like or not. that's a headphone user's privilege.
    WoodyLuvr and analogsurviver like this.
  10. analogsurviver
    One of your best posts so far.

    Things along the above lines started LONG TIME AGO - still firmly in the pure analog domain, with CD launch still at least a decade in the future.

    All of the issues mentioned have been tackled by various manufacturers from the past - partly, one, maybe two issues at the time - but never all together. The times were not yet ripe, the technology - or better said, profitability of such an endavour - has been at anything but reasonable investment return level. Nor has been the headphone market at anything like the magnitude it is today - where it, effectively, holds the lion's share of R & D in all of audio today.

    Harman Kardons approach with curve is the easiest way out possible - and does NOT address enough issues if we are finally to experience the sound of reproduction of recorded sound via both loudspeakers and headphones to converge towards similar>same impression - that of the sound heard in a live performance.

    Issues not covered by HK:

    1.) Type of headphones. Open air, semi closed, closed back.
    BIG DIFFERENCE - as it governs human ability to use - or not to use - intraaural time/frequency information and the use of our outer ears (pinna).

    An open air ( ear speaker ) "headphone" like AKG K-1000 ( or its successor, MySphere 3.x ) does allow for completely undisturbed natural use of our ears and head - whereas closed back (and IEMs ) completely prevent our natural ability to hear. With semi closed somewhere in between. Open air type require (almost) no individual compensation, closed backs and IEMs require crossfeed and frequency compensation because of the lost frequency filtering of our outer ear ( pinna ) - which gets completely avoided by IEMs. The more the headphone or earphone is removed from the open air design, the more and more complex and individually adjusted processing it requires. While hit/miss/chase/almost got it approach with curve(s) can bring an improvement, the total lack of crossfeed ( if and when required ) leaves LOTS to be desired.

    2.) Type of loudspeaker. Point source, Line source or Dipole.

    The primary difference among the three basic types ( or what actual loudspeakers approximate to in real life ) is their interface with room; the most affected being the point source ( omnidirectional, exciting room resonances in all directions/planes equally ), the least affected being the dipole ( does not radiate in its plane at all - its polar characteristics is figure of eight ) - with line source being in between the two extremes.

    To lump such diverse interaction of the speaker basic types into a single curve for listening with headphones is a gross approximation - at best.

    3.) There has been at least one commercially available loudspeaker line using "crossfeed in reverse" - Polk Audio. Again taking in consideration human head dimensions and spacing between the ears - something all other loudspeaker manufacturers "gloss" over. https://en.polkaudio.com/discover/sda-technology

    Above is only the tip of the iceberg, visible from the surface. We all know the vast bulk of the iceberg is hidden below the waves ... - but today's technology AND market penetration by headphones should yield a commercially viable solution in reasonable future.

    Just don't delude yourself it can be done - at least not well - with a "curve" alone.
    Last edited: Jul 15, 2019
  11. TheSonicTruth
    Personal example: A lot of people point out the high-freq spike(around 9.5kHz) on the Sony MDR-7506. My hearing is already down 4dB at 8kHz, and I can barely hear above 12kHz anyway. And the rest of their signature is neutral enough to my ears.
    taffy2207 likes this.
  12. taffy2207
    I get the same thing with my phones. They're known for having peaky treble but I barely notice as I'm 48. My wife is 29 and has the same set up as I have and it bothers her more than me, but not that much. They sound like bliss 99% of the time to my old ears. I wouldn't recommend them to anyone though, they're far too eccentric, but, thankfully so am I.
    Last edited: Jul 15, 2019
    TheSonicTruth likes this.
  13. gregorio
    They can, individually, sometimes. For example, headphones that "a lot of end listeners like" might be good as a reference when mastering, headphones which are flatter and exaggerate detail would be good when editing, isolation and a weak bass are priorities for monitoring a cue mix (by the musicians), etc. In other words, there are different requirements for different jobs and different job roles. There is no single headphone that can do all the required jobs well and there can't be, because the requirements are sometimes contradictory/mutually exclusive. It would be like trying to find one car that fulfils the roles of; small city car, large family cruiser, pick-up truck and race car. It would necessarily be compromised in at least one, if not all of these different roles.

    1. That their "sole purpose is to offer maximum sound quality" is why they are NOT pro headphones! Firstly, "maximum sound quality" from headphones would only be useful under certain conditions/situations and even then, they still have to be practical, (durable, easy to carry around, quick to interface with studio equipment, etc.).

    2. Firstly, the only headphones studios tend to own are those supplied to the musicians for listening to the cue mix. The different headphones used by the different engineers are generally considered "personal items", owned by the engineers themselves rather than the studios. I'm sure some engineers use some esoteric audiophile headphones but very few and the occasions where they would be the appropriate headphones for the job are very rare. Secondly, we've already ascertained (several times) that you have very little/limited experience of commercial recording studios and certainly not world class ones, so you should be asking yourself that question!
    2a. That's what you're going on, staged publicity photos of recording sessions? Oh dear!

    Again, you have no actual pro training/experience and little/no idea of why/how real professionals work, you just make-up false "facts" to suit your agenda and have convinced yourself you know all about professional/commercial recording and production from your purely amateur experience!

    1. I don't believe that's entirely the case. We may not have absolute proof but there's very strong circumstantial evidence.

    2. Your figures aren't really accurate but even if they were, it would be largely irrelevant, as fundamental frequency gives us relatively little information as far as human speech is concerned. The fundamental frequency of some sounds common in speech can be as high as 1kHz and extremely rarely is there any content at 80Hz. Also, the harmonics in speech are vital, they provide much of the information regarding the speaker's mood, intentions and meaning. Human speech can cover a range of 80Hz - 12kHz but the most vital range is about 200Hz - 6kHz, while the area around 2.5kHz - 3.5kHz is particularly important for clarity/intelligibility. As far as I'm aware, our sensitivity around 3kHz is due to that being the resonant frequency of the ear canal and it seems more than a little coincidental that this is also a particularly important frequency area for speech. Regarding evolution, complex spoken language appears to be a defining characteristic that's unique to our particular sub-species and possibly even to our sub-sub-species! Early Homo Sapiens (c. 200k-100k years ago) had a significantly different vocal tract physiology compared to Homo Sapiens of c. 100k-50k years ago and they wouldn't have been capable of all the vocalisations required by complex spoken languages.

    SilentNote likes this.
  14. TheSonicTruth

    That's what I was after with that question.
  15. johnn29
    From Tyll Hertens presentation on the Harmen target curve he mentioned that the bass boost likely reflects the boundary gain when a flat in an anechoic chamber loud speaker is placed in a real room.

    That video is really great for anyone interested in this stuff.
    castleofargh likes this.
879 880 881 882 883 884 885 886 887 888
890 891 892 893 894 895 896 897 898

Share This Page