Sound Science Questions

Discussion in 'Headphone Measurements' started by MoatsArt, Mar 1, 2016.

  1. lm4der

    lm4der A very good sport - Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    1,461
    Trophy Points:
    113
    Location:
    Seattle, WA
    Yeah... crap.

    Point goes to Bill-P. I'm reading your post and thinking to myself "Oh, so it's Tyll's fake ear that is creating that 2-5k bump!" Then I'm thinking to myself "Oh Jesus, duh. Of course. What am I, new?"

    I seem to have left my brain somewhere, so, can I be promoted to Captain Duh? Maybe the person that changes the titles can update mine? Who is that, @zerodeefex or @HitmanFluffy?

    I seem to remember a post that captures the SBAF measurement methodology, but I can't find it. Linky?
     
  2. HitmanFluffy

    HitmanFluffy Hoping to see real genitals someday!

    Anti-SBAF PSYOPS Banned
    Joined:
    Sep 26, 2015
    Likes Received:
    1,532
    Trophy Points:
    93
    Location:
    Austin, Texas
    Whoa, I'm just a pal here. I've never had any admin powers ever :)
     
  3. Bill-P

    Bill-P Level 42 Mad Wizard

    Pyrate
    Joined:
    Sep 25, 2015
    Likes Received:
    4,801
    Trophy Points:
    113
    Location:
    San Jose, CA
    I think it (the article) got lost on the old site. :(

    And yeah, I made the mistake of thinking "flat" on Tyll's graphs actually meant something too... a while ago. Nowadays, I have wised up and started looking deeper into measurements rather than just looking at the graphs.

    Like I mentioned, I found that even when two headphones measure the same, fitment, and enclosure may still change the way it sounds. This is why a headphone that measures like the Orpheus HE90, with a flat-ish response, and the same low distortion figures, may not sound the same as the Orpheus, although tonally, I think both will be pretty close to one another.

    This is a problem with the current measurement methodology, I think. When we place the mic so close to the enclosure, we are more or less eliminating side reflections (from ear pads or enclosure, HD800 is infamous for having side reflections here) which may or may not alter our perception of certain frequencies. If we move the mic further away, then we won't be able to measure bass, because bass needs a good seal and couple to be transmitted.

    So the current measurement methods have limitations beyond what we can deal with.

    Though... a headphone that measures flat, well, does necessarily measure flat anyway. At least on most rigs where ear shapes aren't taken into account.
     
  4. lm4der

    lm4der A very good sport - Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    1,461
    Trophy Points:
    113
    Location:
    Seattle, WA
    This is exactly how I understand it. But I'm having to rethink it a bit after @Bill-P's explanations, so hopefully he (or others) can comment.
     
  5. Bill-P

    Bill-P Level 42 Mad Wizard

    Pyrate
    Joined:
    Sep 25, 2015
    Likes Received:
    4,801
    Trophy Points:
    113
    Location:
    San Jose, CA
    Mm... I think you're kinda barking up the wrong tree.

    1) Frequency response, as measured on any website, in my honest opinions, is a very good approximation of how the driver itself is doing inside a given acoustic space (enclosure? or no enclosure?)

    2) However, notice how I said "driver" and not "headphone". A headphone as a whole still has an enclosure, and this enclosure affects the sound in a very drastic way. HD650 has a frequency response pretty darn close to the LCD-2. Both have equally low-ish distortion above 200Hz, so we can safely assume distortion is out of the equation. They do not sound the same. At all. Not even with regards to timbre, though tonally, both are kinda similar. That's the easiest example to follow.

    3) Having said that, the rest of the discussion is kinda moot...

    The above is also the reason why you really can't just take any random headphone, apply a super elaborate EQ or DSP, and somehow get about the same sound. It just won't happen. The headphone as a whole is still an entirely different beast.

    It's also why you can have earphones (ear buds or IEMs) with a certain response and they won't sound anything like any full-size headphone that measures with the same graph in an ideal situation.

    Frequency response and other kinds of graphs here should be used more as a comparative tool rather than an assessment tool. Meaning... you shouldn't base your absolute judgment on how something sounds like based on how it measures. You should listen and decide for yourself whether it sounds good or not.

    Does this have anything to do with the geometry of your ears and the headphone? Well... I honestly have no idea, but I doubt that's true for full-size headphones. Most people typically hear most full-size headphones (given there's no fitting issue) about the same way, even considering their differing tastes in music and sound signature. I think it's a myth that your ears matter in this case. If they really do, then there was never a point to HRTF to begin with, because each person will have his own HRTF, and there won't be such a thing as an HRTF that applies to everything.

    In short, if you believe people hear things differently because of something that's unique to them (ear shape?), then it is also necessarily true that HRTF is futile as a generalized method of compensation, and the premise of the discussion is invalid to begin with.

    Funny how that plays out, eh? :p
     
  6. Klasse

    Klasse Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    117
    Trophy Points:
    43
    One thing to consider is that most of recordings are made for speakers, not headphones.
    As the sound goes from the speakers to your ears, it interacts with your body (the shape of your head, size of shoulders, shape of your ears, etc...) and it's equalized as a result, and that's what you hear.

    Same happens on a live venue.

    When you put on headphones, you 'cancel' part of those fundamental interactions since now the drivers radiate sound directly to your ears. So part of the body-equalization won't take place and the frequency response of the headphone should be tweaked to address this problem.

    First thing that comes to mind, they (the people making the headphone) don't know you.
    You might be tall and slim with a small round head or really short with a giant head or whatever, then addressing that probem is no piece of cake.

    You should also consider that the sound from speakers (or live venues) normally comes towards your ears coming from the front, while on typical headphones the drivers radiate sound from the sides.

    So while we still have our ears in place ready to tweak the sound when using headphones, just as with speakers or real concerts, they won't tweak the sound in the exact same manner when sound comes from the sides as when it comes from the front. So the headphones frequency response should address that problem too (although the designers don't know the shape of your ears).

    There's a trend towards headphones with angled drivers or angled pads, so the sound reach your ears in a slightly more frontal way.

    More so, the size and geometry of the driver affect the shape of the sound wave, and this affects its interaction with your ear. Using a typical 40mm dynamic driver won't yield the same results as a 56mm Ring Radiator or a big planar magnetic driver.

    Last but not least, there's no natural crossfeed on headphones. When sitting in front of two speakers your left ear hears a good portion of what's coming from the right speaker and vice versa. The way your left ear receives the sound coming from the right speaker is not trivial since there are some pretty obvious but not simple geometric reasons involved that end up tweaking the sound as a whole.
    So the natural crossfeed or more precisely, the lack of natural crossfeed on headphones should be addressed somehow and while some frequency response tweaks might improve the experience with some recordings they can also cause even more problems with others since it's a pretty complex phenomena that cannot be fully addressed via frequency response tweaks.

    Last but not least (bis)
    Preferred listening levels play a role here as well since we are not particularly linear when it comes to sound perception at different levels (Equal Loudness Contours), then what could be interpreted as perfectly flat / perfect tonal balance at (say) 90dB SPL Average, won't be perfectly balanced at 75dB Average since our perception changes.
     
  7. Klasse

    Klasse Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    117
    Trophy Points:
    43
    I know it can be quite overwhelming (everything compacted in a few paragraphs), but my post is all about frequency response on headphones which translate to timbre fidelity.
     
  8. Friday

    Friday Friend

    Pyrate
    Joined:
    Sep 29, 2015
    Likes Received:
    520
    Trophy Points:
    93
    Don't forget the unsolicited stack of unnecessarily mind-numbing journal papers
     

Share This Page