What is Neutral Frequency Response

Discussion in 'Tales from the Bully Pulpit' started by purr1n, Sep 26, 2015.

  1. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    60,655
    Dislikes Received:
    105
    Trophy Points:
    113
    Location:
    Antarctica
    As an aside, I do tend to purposely use subpar recordings to subjectively test headphones. There is a good reason for this which I will explain in a bit.

    This of course comes down to the matter of what is neutral frequency response. There has been a lot of misunderstanding of neutral. One such misunderstanding has been the straw-man (falsely assumed of neutrality proponents) argument from none other than Steve Guttenberg that neutral frequency response is an attempt to get closer to the artists' vision. (When I play bass guitar or piano, I could give a rats ass about how it sounds on the HiFi - it's the performance that matters.)

    Neutral frequency response is very simple. It's simply a reference, albeit difficult to exactly define subjectively with 100% accuracy, which one tries to shoot for. Subjective observations of neutrality will vary slightly from audiophile to audiophile. Subjective observations of what is neutral vary little among recording engineers. (UE was not kidding, making stuff up, or futzing around when they said their reference monitors were tuned by sound engineers at Capitol Records.)

    Even though the "neutral" bullseye may be fuzzy subjectively, at least we know there are people who try to shoot for it. Because of personal tastes of sound engineers (or even directives from Studio Execs), we will see variances in how material is mixed and mastered in terms of tonal response. For example, the guy (don't remember his name) who does a lot of the Patricia Barber stuff prefers a darker sound (with close mic'ing) vs. the guy who did the Bowie Let's Dance remaster a few years ago. Natalie Merchant's Tigerlily is dark compared to the MFSL Pixies stuff.

    So there's always going to be variance. What I love about "neutral" is that this variance will almost always sound acceptable on a neutral system. I like to use MFSL Nirvana Lithium to test upper midrange issues. A few tracks from Radiohead The Bends I also like to use for this test. The K550, at least the "defective" pairs which I've heard, massively failed this test. You can call it shitty music or whatever; but the fact is, I feel the interpretations are spot on from the recording engineers for these two discs. Certain performances are supposed to sound aggressive. But on bad system or system with upper midrange issues, this is going to sound nasty.

    I like to use Natalie Merchant's Tigerlily to make sure systems are not too dark. Systems which are too dark tend to suck the life out of that record. I also use Talking Heads Naive Melody (remaster) to test for a variety of many things (too much bass, vocal sibilance, percussion nasties, etc.)

    What I don't use is classical music. Classical music is generally bandwidth limited.* Now comes the question: is a transducer which sounds good on classical / girl+guitar / flamenco guitar, but not with modern popular music a good transducer? IMO, the answer is an unequivocal no. I'm not into mid-fi.

    What it comes down to is that a neutral system will be the most flexible with the most recordings for most (but not all) people.


    Does this invalidate personal preferences? The answer is also an unequivocal no. Heck, I happen to think the HD800 works really swell with Tigerlily. The HD800 really helps liven up that recording. Although I'm sure others may prefer Tigerlily from the LCD2 instead.

    BTW, I do prefer a very slightly dark sound. The Paradox s/n 001 fits the bill. So did my modified Jades. The UERM is slightly too "reference monitor" for me (it's based on the Yamaha NS10 studio monitor), but I do like the UERM. As of 2015, I am using modified Sennheiser HD650s.

    Finally, as along as a transducer has smooth response, I'm generally OK with it and can adapter. So even then, I wouldn't say that neutrality is my number one priority (another misunderstanding which folks have attributed to me.) Number one priority is a smooth response (no peaks). EQ'ing a screwy frequency response is difficult. EQ'ing a smooth response is easy.
     
    Last edited: Oct 11, 2015
  2. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    60,655
    Dislikes Received:
    105
    Trophy Points:
    113
    Location:
    Antarctica
    Haha, I can't believe I wrote that. There so much good stuff scattered around here. Lots of different angles to take on neutral, and I feel the subjective angle described above is a good one.

    An objective angle is another; but it needs to be supported by the subjective observations. Amps and DACs should always measure flat more or less. If the equipment is not for pro-use (where it would be slotted in long chains of other gear), a little roll-off at the extremes would be OK and probably imperceptible. Distortion and other non-linear characteristics are going to have a larger impact on sound, even in the perceived tonal balance sense, than 1.5 db roll-off at 20kHz. This is what Hydrogen Audio guys fail to understand when they nitpick FR graphs from RightMark Audio Analyzer.

    For transducers, depending upon the measurement method used, there should always be a target response to achieve neutrality. A flat line isn't necessary that target. At the listening position with speakers, the target response I like is the old B&K target (http://www.bksv.com/doc/17-197.pdf) see page 4. [I've seen HF'ers and a few people here incorrectly refer to this as the Olive-Welti curve. The O-W curve is the target / compensation for their headphone measurement system. Sean Olive did not invent this target curve. The B&K target was formally presented to AES in 1974, and sound engineers were already aware and using similar type curve since the 1950s or earlier.] A flat response will sound like shit at the listening position. This B&K target also somewhat works with the V2 measurements for headphones on this site.

    Now we hear people say speakers should measure flat. Yes, speakers should measure flat if the measurement is taken at 1 meter away with the microphone leveled with the tweeter. That's how Stereophile does it. That's the standard that many speaker builders use. This method is useful because Stereophile or the speaker-builder does not know how your room is going to influence the sound. Hence the "standard". When I build speakers, I always take measurements at the listening position. That is because I can. Studios, when they precision EQ their speakers for neutral response, also take measurements at the listening position (using B&K target or similar curve.)

    Different methods, different techniques, different targets. All attempt to achieve neutral frequency response.
     
    Last edited: Sep 26, 2015
    trl, drbobbybones, Changeling and 4 others like this.
  3. Koth Ganesh

    Koth Ganesh Friend

    Friend
    Joined:
    Sep 26, 2015
    Likes Received:
    1,961
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    Chennai, India
    Marv, that was very useful for me...now let me try to boil it down to an example I understand for now..... Out of all the HPs I own over the various types - planars, dynamics, electros - I find the Paradox Slants approach what you talk about in the first post with a bias towards slight darkness which to me is very pleasing. Slightly behind it is the HD 600 (without the darkness). Are these observations on the right track?
     
  4. Riotvan

    Riotvan Got lost for three weeks at Delft City Hall

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    2,452
    Dislikes Received:
    6
    Trophy Points:
    113
    Location:
    The Netherlands
    My experience with something that sounds flat and neutral to my ears is that it might not wow you at first but it will most likely not piss you off either. In fact i find it really allows me to focus on the music and just forget about the gear. Every single time there is some fr dip, peak or what have you i end up focussing more on that and in extension my gear. Which can set off all kinds of audio nervosa and takes away from the enjoyment of music.
     
  5. Thad E Ginathom

    Thad E Ginathom Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    7,622
    Dislikes Received:
    13
    Trophy Points:
    113
    Location:
    India
    * ??? :confused:

    The asterisk implies that you meant to enlarge on this. Please do!
     
  6. NoStream

    NoStream Acquaintance

    Joined:
    Sep 27, 2015
    Likes Received:
    61
    Dislikes Received:
    0
    Trophy Points:
    18
    Location:
    Oakland, CA
    I've been working a bunch with EQ and headphone correction software lately and thinking about this sort of stuff. If you look at Sonarworks' software, their reference curve is based on flat-measuring speakers in a well-treated room... i.e. what studio engineers are working with. And the curve they come up with is very similar to the HD600s and 650s rather than the 800s. (They actually EQ up to -5 dB at 6k on the 800s, vs. my old curve with -3 dB.) I'm beginning to think that idea is correct. Headphones with response like speakers at 1m tend to sound really bright, since speakers in a room see more absorption than headphones on a head.

    The Olive curve is also interesting but is just based on listener preference, not flat response. So that seems less useful for "audiophiles."
     
    ultrabike and purr1n like this.
  7. Psalmanazar

    Psalmanazar Most improved member; A+

    Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    4,003
    Dislikes Received:
    121
    Trophy Points:
    113
    Olive's headphone curve for Harman is inaccurate. It's simply the equalization that their testers preferred with the HD 518 and LCD 2 listening to Steely Dan, rap, and vocal pop. The bass is not "speakers in a room warmth" as Tyll put it in that video presentation on headphone flatness; it's elevated. What Harman did find though was that both trained and untrained listeners regardless of age and location preferred a more neutral response.

    Put on a pair of headphones whose bass frequency response is voiced close to the Harman curve such as the Focal Spirit, NAD HP 50, Philips X2, or several closed cans that come close to it. Play some decently mixed, not brickwall mastered harder guitar rock or metal with audible bass guitar and kick drums with actual bass in them. I'm recommending guitar riff based music due to the way it's mixed. What you'll hear is that the instrumental levels are screwed up. The bass drums and sometimes the bass guitar will usually be slightly boomy, often louder than the guitars which will come out recessed. This music was of course not mixed to sound that way. What you hear through those headphones actually goes against the development over the last fifty years of more powerful, higher gain amps to make guitars heard over loud drum kits. The Harman curve slightly shelves the midrange frequencies where the human voice and thus most melodic human instruments reside. It's the opposite of honky; it suppresses the music.

    The Harman curve is the the 21st century version of the 1980s diffuse field curves that the Beyerdynamic DT 770/880/990, Sennheiser HD 540, and AKG 240 were voiced around. The Sennheiser HD 580/600/650 series moved past such deliberate coloration. Why go back?
     
    Last edited: Sep 30, 2015
  8. Bill-P

    Bill-P Level 42 Mad Wizard

    Friend
    Joined:
    Sep 25, 2015
    Likes Received:
    3,734
    Dislikes Received:
    15
    Trophy Points:
    113
    Location:
    San Jose, CA
    Yeah, I'd agree with you that Paradox Slants is close.

    I think the HD600 can be polarizing. At stock, it has some weird reverb/resonance problems in the lower mids that cause it to sound very thick and warm, and may even be kinda dark as you're putting it? (this is entirely dependent on ear pad age as well, by the way, but that's for a different topic)

    An HD600 with as low distortion as possible is basically the perfected neutral sound signature IMO. It doesn't seem to do anything bad, and there is just enough of everything else. Maybe I'm wishing for a bit of low bass (even the HD800 beats the HD600 at low bass IMO) sometimes, but for most music, I think it's the de facto standard for a production non-modded headphone that's neutral.

    On that note, though, as above, I don't agree that a neutral headphone can't "wow" someone. I think the factor that "wows" people is more the "dynamic" of a headphone, or how much it can throw something out to grab you. Stock HD600 doesn't do this, as do many of the other neutral headphones on the market (Oppo PM-1/2?), but I think a neutral headphone can still be super dynamic, and then it'll wow. Hopefully you'll get to hear that soon. :) (I'm not hinting at anything, btw... or am I?)
     
    Koth Ganesh likes this.
  9. Koth Ganesh

    Koth Ganesh Friend

    Friend
    Joined:
    Sep 26, 2015
    Likes Received:
    1,961
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    Chennai, India
    Billy.... pleased as punch for the corroboration...Its tough retraining a 55 year old pair of ears but I should be there when I am 60 !! On the 600, you just articulated perfectly what I've been feeling but unable to say as much... the stock version is great but there is no "grabbing me" effect. Your LCD 2.2 does this for me in spades BTW. Finally, don't act coy:D. I am mailing a certain item Saturday (down with the flu all week).

    Thanks mate.
     
  10. Claritas

    Claritas Friend

    Friend
    Joined:
    Sep 26, 2015
    Likes Received:
    719
    Dislikes Received:
    19
    Trophy Points:
    93
    Location:
    R'lyeh, Carolina
    So far as I know, no one's made a headphone that follows the Harman curve. Insofar as FSP and HP50 are close in the bass, the result sounds more realistic to me. Insofar as they're less close elsewhere, the result sounds less realistic. I take that as an indication that they're on to something useful.
     
    Last edited: Oct 1, 2015
  11. mickeyvortex

    mickeyvortex Friend

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    38
    Dislikes Received:
    0
    Trophy Points:
    18
    My 0.02 cents- as a part time geneticist, what I'd like to see is Olive et al working with folks on the genetics of "neutrality". Biologically, there is likely tons of evolutionary selection acting on the ear components considering our reliance on the ears for hearing but also via other means such as preventing ear infections. This means "neutrality" is (and should be) constantly shifting. Variation is preferred.

    P.S. I suggest reading this article from Soundonsound on the structure of the ear and how it perceives and "equalizes" sound- link: https://goo.gl/SLSKnN
    The author uses interesting analogies with an audio signal chain. The intricacies and "players" involved are mind boggling.
     
    Thad E Ginathom likes this.
  12. Luckbad

    Luckbad Traded in a unicorn for a Corolla

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    3,194
    Dislikes Received:
    6
    Trophy Points:
    113
    Location:
    Laguna Hills, CA
    Home Page:
    The beauty of neutrality to me is that it simply doesn't exist in an objective form. Everyone is different. Everyone has different frequency sensitivities and varying amounts of hearing loss, making neutrality objectively inconsistent. Beyond that, we all have preferences.

    I have several dB of hearing loss around 8kHz in addition to quite a bit of treble sensitivity (it gets painful to me over ~4kHz before most people). The result of that combo is that peaks within 3-4kHz +/- 8kHz are brutal to me.

    Someone did a write-up on Head-Fi about balancing your headphones with equalization. I went through the exercise with sine sweeps and such on several setups and headphones and it was pretty interesting. The gist of it is that you want a full spectrum sine sweep to sound completely flat to your perception, so it ends up +/- ~2dB perfectly neutral to you.

    When I went through that trouble, I discovered that the HD650 needs very little tweaking for me compared to most headphones, which could explain my preference for them.

    Even when I got several headphones sounding flat neutral to my ears with sweeps, they all still sounded a bit different based on their capabilities. I also still boosted the hell out of sub-bass no matter what.
     
    Franco72 and Thad E Ginathom like this.
  13. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    3,074
    Dislikes Received:
    4
    Trophy Points:
    113
    I'm with Marv on the flexibility notion. For me a system is not neutral if I can distinguish a coloration which persists between different tracks. Sometimes it takes longer for me to detect it, but once I get it, then the audiophilia nervosa kicks in.

    After our engineers were able to come up with a 'neutral' target curve, the biggest test was to see if it works for studio engineers. Our notion is that a neutral sounding studio rig should be able to provide the best translation to a variety of systems. You can run a grot-box for double checking, but due to peak induced near-by frequency masking and sonic information rendered inaudible by deep-enough dips, an engineer wont be able to hear the whole picture. And thus won't be able to make all of the decisions he needs make with sufficient accuracy.
     
  14. Ferrum

    Ferrum Acquaintance

    Joined:
    Oct 4, 2015
    Likes Received:
    45
    Dislikes Received:
    0
    Trophy Points:
    18
    Location:
    Sweden
  15. The Alchemist

    The Alchemist MOT: Schiit - Here to help!

    Friend
    Joined:
    Sep 24, 2015
    Likes Received:
    1,378
    Dislikes Received:
    1
    Trophy Points:
    113
    Location:
    The SchiitR
    Home Page:
    Great post marv
     
  16. Rex Aeterna

    Rex Aeterna Friend

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    257
    Dislikes Received:
    0
    Trophy Points:
    63
    Location:
    Cinnaminson, nj
    I also fine tune my system at listening position. Im not gonna be listening at 1 meter away....


    I always used music for fine tuning. I always only suggested sine wave sweeps to check for channel imbalances or room acoustic issues but even then i use music most of time tuning a room too....but, that's just me.
     
  17. Deep Funk

    Deep Funk Deep thoughts - Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    11,840
    Dislikes Received:
    13
    Trophy Points:
    113
    Location:
    South of Amsterdam, near The Hague.
    Home Page:
    This thread really makes me think.
     
  18. Cspirou

    Cspirou They call me Sparky

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    5,388
    Dislikes Received:
    5
    Trophy Points:
    113
    Location:
    Northwest France
    I remember seeing a talk with Jerry Harvey at Google. Jerry mentioned that he doesn't tune his IEMs for neutrality but for accuracy. In other words he wants a kickdrum to sound like a real kickdrum even if the FR curve looks really ugly.

    Another thing is that people say they want things to be neutral that way you don't add any coloration and you hear exactly what the artist/producer intended for you to hear. The problem is you can never truly replicate what they heard unless you have the exact same monitors that they had. Whatever they heard on the amp and speakers does not get captured on the master tape and your home amp and speaker has no way of really knowing how much distortion the engineers heard while recording.
     
    Lyander and Vanheim like this.
  19. OJneg

    OJneg The Most Insufferable

    Friend BWC Banned
    Joined:
    Sep 25, 2015
    Likes Received:
    3,751
    Dislikes Received:
    37
    Trophy Points:
    113
    Location:
    Grand Rapids
    This isn't a bad concept, but when it comes down to tuning you need to have a wider swath of references or else you lose sight of the target. It's easy to fall into the trap of making that kickdrum sound even "better" and more "realistic". And that's a job that should be left for studio engineers (if that) rather than folks making reproduction equipment.
     
  20. Kunlun

    Kunlun cat-alyzes cat-aclysmic cat-erwauling - Friend

    Friend IEMW
    Joined:
    Sep 26, 2015
    Likes Received:
    5,248
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    Meow Parlour
    I'd love to see people's take on what an in-ear canal neutral curve should be. There are a number of resonance peaks in the human ear canal (one at 2 and a half-ish khz, for example) and occluding the ear canal makes the picture even more interesting. I think a little understanding of ear canal acoustics plus subjective experience plus research to see how groups of people hear things (there was a study out of Aus. that showed people found an iem with slightly elevated sub-bass sounded more natural, for example), could lead to some interesting ideas.

    Jerry's an interesting case, he said in that talk that he never tunes flat and that he finds flat to be boring.

    First, flat and neutral are different, I would say, because the human ear is not "tuned" flat. Second, speaking medically and not as a shade on Jerry, the man has hearing damage as do many, many people who have done his work. That has to be considered, at least to some extent in how he tunes and what we hear in his products.

    On the other hand, an audio engineer with a huge amount of experience argued that people can adapt, so an audio engineer who knows what his neutral sounds like will still be able to tune for it regardless of his hearing loss (he may ask for a second set of ears to check things over, though). I'm not sure if that's true, but maybe.
     

Share This Page