Harman Curve Poll (Public)

Discussion in 'Blind Testing and Psychoacoustics' started by ChaChaRealSmooth, Nov 9, 2020.

?

What do YOU think of the Harman Curve

  1. Strongly Disagree- The Harmon Curve sounds bad and it's potentially dangerous for good sound

    28.7%
  2. Somewhat Disagree - I like some things about it, but I'd rather not be the target

    38.5%
  3. Somewhat Agree- I like more things about it than not

    25.3%
  4. Strongly Agree - Now, this is how music should be heard!

    2.3%
  5. f**k me in the ass with the IE 2017 treble boost! (Harmon Curve is for kids with hearing loss)

    5.2%
  1. mrflibble

    mrflibble Friend

    Pyrate
    Joined:
    Oct 4, 2016
    Likes Received:
    358
    Trophy Points:
    63
    Location:
    United Kingdom
    Recently, I have been playing around with the AutoEQ software, creating EQ curves for my headphones. Included in the source of AutoEQ are various target curves, including Harman, diffuse field, and Oratory1990's Optimum HiFi, https://github.com/jaakkopasanen/AutoEq/tree/master/compensation

    I think it would be very useful to have an SBAF target curve for use with AutoEQ. That way people could run the AutoEQ software against the SBAF curve for their particular headphone and generate the equalization values. The target curves are in CSV format, not sure of the exact specification but they appear to be 695 rows of two columns (frequency and amplitude). What would be involved in putting this together? Could the existing measurements be used?
     
  2. Biodegraded

    Biodegraded Friend

    Pyrate Contributor
    Joined:
    May 28, 2017
    Likes Received:
    7,976
    Trophy Points:
    113
    Location:
    Vancouver BC
    You could make EQ files from the SBAF MiniDSP EARS & flat-plate compensated results here:

    https://github.com/superbestaudiofriends/headphone-measurements-frequency-response

    The SBAF target is straight line = SBAF perceptive neutral, ie the 'target curve' is flat. So to make an EQ file, make a mirror image of the SBAF compensated measurement (subtract it from zero) and then depending on whether you're going parametric or fixed-band, smooth it/fiddle with it along the lines of jaakopasanen's HD800 example.
     
  3. Psalmanazar

    Psalmanazar Most improved member; A+

    Pyrate Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    5,345
    Trophy Points:
    113
    @purr1n Speaker designers and manufacturers rightfully laugh most of the Harman research out of the building. Most of it was used to build inoffensive sounding, low detail bullshit for home consumption. Bose with science. Stamped woofers, computer modeled waveguides that wreck treble timbre, bad class d amps, because their research said it didn’t matter.

    Other manufacturers like Genelec and Kii used it to try to defeat the laws of physics. Their speakers are all no detail shit and, like Harman, religious fundamentalists, libertarians, and the modern left, they have to employ an army of apologists to tell the skeptics that what they are doing actually works, dinosaurs lived thousands of years ago, distortion doesn’t matter, these people have feelings and because they have feelings they’re not mentally ill, money is a magical number on a screen or a piece of paper so it’s fine to hand it out for free, metal tweeters are okay etc.

    The HD 600 midrange tonally sounds like a speaker with a too high up crossover and the woofer breaking up. It’s still more even than half of JBL’s speakers with their black hole dips often sounding like a notch filter at the crossover. And more detailed than all the dead inside Revel speakers and sheened over, PA system sounding modern JBL compression drivers.

    Harman also owned AKG. The AKG headphones made under Harman don’t hold a candle to Sennheiser.
     
    Last edited: Jan 28, 2021
  4. tomtrp

    tomtrp New

    Joined:
    Jan 13, 2021
    Likes Received:
    7
    Trophy Points:
    3
    Location:
    Australia
    Toole’s research results are somewhat more reliable since flat on axis and great off axis performance speakers actually will sound good (but not necessarily individuals’ favourite) since it tries to reproduce the signal at best, and those research are done in NRC not Harman. Tooled later employed by Harman so he still has conflict of interest in terms of explaining and promoting his result in Harman’s interest. Olive’s research results on headphones are much more limited and problematic and done by Harman.
     
    Last edited: Jan 29, 2021
  5. tomtrp

    tomtrp New

    Joined:
    Jan 13, 2021
    Likes Received:
    7
    Trophy Points:
    3
    Location:
    Australia
    We should accept that both Toole’s and Olive’s research results are meaningful findings but has limited generalizability, most of the results are concluded from a single experiment double blind test of a 268 people sample(and the sample has bad randomness and diversification), no repeated confirmation, not really science level correlation(but high) and precision and the statistical model has some flaws if viewed in pure statistical perspective.
    Their results are valuable and have some good explanation power but not complete, not so well tested, and not through a proper independent validating process to be accepted as truth or general rule like most of physical science results. They are loose science and both Toole and Olive have conflict of interest problems since they are employed by Harman. Just don’t put the same confidence in those research as the real hard core physical science.
    Harman and those measurement extremist only looks at the results and are so eager to put them as rule or standard while reluctant to face the incompleteness, limitations and generalizability. They highlight the findings from sample and downplay then flaws/limitations of the experiment.This is not a proper scientific attitude at all from my point of view.And it’s quite clear that Harman will benefit from the promotion and exaggeration of their own ‘’scientific’’ results as a rule maker. Bad things can happen when science is not independent and run by a profit driven company. I believe more reliable results will be available if proper research in this field can be done by independent academic entities like universities and reviewed, published by well-respected academic publications like nature.
     
    Last edited: Jan 29, 2021
  6. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Harman is too incompetent to make use of it. If anything, it's a pat on the back, confirmation, for those who subscribe to the stereotypical "Asian" consumer IEM tuning. I guess in that sense, the Harmon IEM target is valid.
     
  7. CrankyRat

    CrankyRat New

    Joined:
    Mar 13, 2021
    Likes Received:
    0
    Trophy Points:
    1
    Location:
    Reno, NV
    Jesus, what a load of ... Your just f'ing with us right? If not you have just revealed some sort of odd Synesthesia here concerning politics, music, and the world. But thank you, you are an active member but I will skip your silly stream of consciousness BS. Your ears must be shaped like waffle irons.
     
  8. luckybaer

    luckybaer Friend

    Pyrate Contributor
    Joined:
    Mar 19, 2018
    Likes Received:
    2,191
    Trophy Points:
    93
    Location:
    Missouri
    Lurk moar. Lol...
    Take some time to get a feel for this place before you shit all over a long-time (dare I say “semi-beloved” ;-)) member, ffs.
     
  9. Friday

    Friday Friend

    Pyrate
    Joined:
    Sep 29, 2015
    Likes Received:
    520
    Trophy Points:
    93
    The second coming of RobS?
     
  10. Ox Cart

    Ox Cart Facebook Friend

    Joined:
    Mar 13, 2018
    Likes Received:
    202
    Trophy Points:
    33
    Location:
    Austin, TX
    Audeze LCD-series tuning is anti-harman tuning. I've always kind of loved it, and now I think I even understand it. Contrary to what seems to be a popular opinion in other circles these days, Audeze's approach to tuning is actually kind of genius. Let me explain why ...

    I used to own an LCD2C and loved it for a while, then sold it after I started nitpicking and noticed that it sounded muffled on at lower volumes, especially on things like acoustic folk music (not to mention that it was a tad heavy). Eventually I grew to miss it and I bought a Verum One. With ZMF Ori pads, the Verum does a fair approximation of the LCD2C, with flat and amazingly well extended bass along with somewhat laid back upper mids and generally low distortion/good technicalities, but it never quite gave me the tingles the way that the Audeze could, especially on baroque choral music.

    Here's a raw MiniDSP E.A.R.S. frequency response comparison between the two for reference.

    [​IMG]

    Today I played around with EQ to see if I could recapture some of that Audeze magic. I boosted between 1-2KHz by 2dB, which ads clarity and punch and also contributes to that kind of wet/reverby sound that works well with choral, music. I also added a wide 3dB boost at 15KHz for air which helps get a more spacious feeling and increases apparent detail. Those changes by themselves aren't enough, however, to make the Verum sound like an Audeze. The real magic lies in the upper mids.

    Against all that Harman holds holy, I cut 2dB at 4KHz and 6KHz and boom! there's that Audeze sound. Why does this work? Harman says you need a massive sub-bass boost, but this thing sounds plenty bassy, impactful and rumbly. Harman also says that you need lower treble for clarity and realism, but the timbre on this thing is actually pretty decent (not Sennheiser decent, but not bad). Everything sounds very real, albeit a little warm, while the main effect of the pulled back upper mids is to make instruments sound a bit more distant, creating a good sense of space and an experience similar to what I'm used to hearing at live concerts back when those were a thing.

    One thing I've noticed is that with an Audeze style tuning, I tend to listen at somewhat louder volumes than with a Harman or Sennheiser type tuning. My hypothesis is that this all comes down to equal loudness. Our ears' relative sensitivity to different frequencies varies by listening level. At lower SPL, bass and treble are relatively quiet. It's not surprising that people would enjoy a healthy bass boost at lower listening levels. However, as you turn up the volume, perceived bass and treble start to catch up with the upper mids and you don't need as much of a boost.

    From Wikipedia

    [​IMG]

    If you look at equal loudness contours, you'll see that our hearing is most sensitive right at 4 KHz. I have no citation for this, but I suspect that due to our sensitivity there, 4 KHz ends up playing a large role in our perception of loudness. Audeze's "trick" is to keep the area around 4 KHz relatively subdued. Assuming that people do judge loudness by that frequency, and assuming that people set their volume based on how "loud" the music sounds, for every dB you cut at 4 KHz, people turn up the volume by a dB to compensate. Thanks to equal loudness, the perceived effect on bass and treble will be even greater than however much you turned up the volume. That could explain the perception of Audeze as somewhat bassy despite having very linear measuring bass, people simply listen to Audeze louder because the relaxed upper mids invite it. Another other happy consequence of this trick is that, as volume gets louder, perceived detail, clarity and dynamic range increase because you're pushing further above the noise floor of your listening environment. That could potentially account for people's perception of Audeze as generally quite detailed despite their somewhat dark sounding tuning.

    A good confirmation of this is that when I turn down the volume, the Audeze style tuning does start to sound muffled and dull after a point. I should note that even with this, I don't listen super loud - 75dB SPL(A) is enough to let me enjoy this tuning. For comparison, I can enjoy my HD58X at 65dB (assuming a fairly quiet listening environment).

    In hindsight, I don't feel too bad about selling the LCD2C. I think it went way too far in cutting 4 KHz and I often ended up listening at around 85dB to enjoy it. So while the Audeze 4 KHz trick has a lot of merit, in implementation they sometimes take it too far.

    P.S. The observant of you might be wondering about that 6 KHz cut, why is it necessary? Well, to my ear, after cutting at 4 KHz, timbre on hi hats and such sounds a bit off. Adding a 6 KHz cut leads to a more gradual rise from the recessed upper mids into the treble, better balancing out the harmonics and giving a more natural sound.

    P.P.S If the "trick" to increased enjoyment is just turning up the volume, why not just turn up the volume on other headphones? Well to me, stuff with Harman-like upper mids starts sounding shouty at higher volumes and becomes fatiguing very quickly, and stuff with strong lower treble presence gets overly bright/sharp and fatiguing.
     
    • Like Like x 6
    • Epic Epic x 2
    • List
    Last edited: Aug 1, 2021
  11. MaxBuck

    MaxBuck New

    Joined:
    Sep 20, 2021
    Likes Received:
    15
    Trophy Points:
    3
    Location:
    La Quinta, CA
    I'm not in any way married to the "Harman curve" or indeed to limiting my consideration to speakers that "measure well," but I've found that the speakers I like best (KEF, Revel, lesser extent Elac) do actually tend to "measure well." Not sure what that really means, but IMO people should buy what they like, regardless. Also not sure what the purpose is of arguing over this preference stuff.
     
  12. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Take your speakers that you like that measure well. Then add some upper midrange emphasis and a ton of bass. Voila, Harmon target.

    Harmon target doesn't measure well. It's a consumer preference target that given who and where you ask, is either 70% preferred or 70% disliked.

    f**k Sean Olive.

    FYI, Oratory at Reddit actually has a good target that "measures well" - perceptive neutral - the Optimum Hi-Fi target.
     
    • Like Like x 5
    • Respectfully Disagree Respectfully Disagree x 1
    • List
    Last edited: Sep 21, 2021
  13. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The purpose is to take stand. Not everything is good, even if it's from a hack like Dr. Olive who obviously needs to publish AES papers once or twice a decade to keep his cushy job at Harmon.

    We totally get the different preferences here on SBAF, even going to far to espouse system synergy with proper selection of components and having personal goals when it comes to sound. However, sometimes we need to draw a line in the sand. Not be spineless. Evil (bad sound) must be stopped in its tracks.

    You mentioned some good speaker companies who actually have an idea of what good sound is like. How would you feel if one day you woke up and all these speaker companies started making speakers voiced to the Harmon target? I don't know if you are aware, but the Harmon target has had a disproportionate influence in the personal audio sphere to the point where people thought it was a some kind of variant of neutral. All the while Dr. Olive just stood by never correcting folks that this target was actually a consumer preference of randomly selected people. He could have done the right thing, clarified his research; but instead he just basked in his own ego.

    This poll was a response to sources claiming that over 70% of people preferred the Harmon curve. Obviously, in this community, it's the other way around where 70% dislike the Harmon curve.
     
    Last edited: Sep 21, 2021
  14. Tekker

    Tekker Facebook Friend

    Banned
    Joined:
    Aug 1, 2020
    Likes Received:
    133
    Trophy Points:
    43
    Location:
    Netherlands
    It’s ironic how everyone at ASR sees the Harmon Target as neutral, just because Amir said it’s done by scientific research.
     
  15. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The research is valid. The questions to ask are how the research was conducted (who were the subjects) and what was the intent. Also, AES papers are hardly peer reviewed. No one ask him any hard questions. Anyway, if I were going to make the next Dre Beats, I definitely give the Harmon target a huge consideration.

    I don't think Amir realizes that the Harmon target sounds very different from the ideal target of his speaker spin-o-rama thing.
     
    Last edited: Sep 24, 2021
  16. roshambo123

    roshambo123 Friend

    Pyrate Contributor
    Joined:
    May 26, 2018
    Likes Received:
    2,762
    Trophy Points:
    93
    Location:
    Los Angeles
    The Harmon curve is market research. The cause of the effect is the key thing, which is naïve consumers with less experience are likely prefer the 'Vivid Profile' just as I recall they express preference for sound systems that are demonstrated at louder volumes. Americans also like over-salted food, but you're not going to fool someone who has sampled many well prepared dishes. So, if Olive found anything, it is simply that crude tricks can garner a response from those who don't know any better.
     
  17. wormcycle

    wormcycle Friend

    Pyrate
    Joined:
    Aug 13, 2016
    Likes Received:
    1,506
    Trophy Points:
    113
    Location:
    Toronto, ON, Canada
    The thread is a bit stale but I recently started experimenting with convolution filters on Jriver and Roon, mostly downloaded from AutoEq. My main beef with Harman curve is not so much the target, although the lower bass looks silly high, as how people are trying to EQ headphones to meet this target. And that includes smart people like Tyll.

    I tried three headphones: LCDX2021, HD880 600Ohm and HD650 with and without Custom Cans copper mode.
    I am planning to post some impressions in appropriate threads but in general each and every EQ included 27-30Hz, +6-7dB, q=0.71 added blindly just to follow the target. And with no exception it sounds bad, it creates audible distortion. But when I reduce this band by -3dB, I like the sound with both Oratory or Inner Fidelity convolution filters. HD650 sounds clearer without teh Senn veil.
    I also remember that the original Oratory EQ for HD800(S) included +12 dB boost at 30Hz, just to meet the Harman target and it sounded exactly the way you would expect, like shit.
    The Oratory LCDX2021 convolution filter, with -3dB in low bass, sounds actually slightly cleaner to my ears than the Audeze preset.
     
  18. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,771
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Trying to boost low bass with headphones which have severe low-end rolloff is counter productive. All we would be doing is increasing excursion without much increase in output, and any increase in output would accompanied by massive distortion not only in the bass, but also with the rest of the audio band - this because magnetic gaps are not linear.
     
    • Like Like x 6
    • Agreed, ditto, +1 Agreed, ditto, +1 x 1
    • List
  19. Zardoz

    Zardoz New

    Joined:
    Aug 26, 2023
    Likes Received:
    15
    Trophy Points:
    3
    Location:
    California
  20. AxeFigo

    AxeFigo Acquaintance

    Joined:
    Sep 7, 2023
    Likes Received:
    30
    Trophy Points:
    18
    Location:
    China
    I'm always kind of sad about these capitalists use statistics as hypes with so many hiding plots, and people could fall into the traps of so-called authority easily without knowing the truth. The Harmon curve is more like an average statistics, and I know you get it, which of course would be someone's preference sounding, but for the most of the time, it can't be the one, cuz that's what statistics means. The major of statistics is to analyze stats and give a rough direction, which is sad to me that some companies make it an authority and people naturally buy it, but we have no ways to do anything, which is the sadness of post-modernity.
     
    Last edited: Sep 15, 2023

Share This Page