General question regarding THD measurements

Discussion in 'Headphone Measurements' started by pila405, Aug 22, 2016.

  1. pila405

    pila405 New

    Joined:
    Aug 15, 2016
    Likes Received:
    1
    Trophy Points:
    3
    Hi,

    While some sources like Tyell\Innerfidelity and most of the users on these forums seem to get results between 0.1% and 1%, some sites\measuring systems get results which are 10 times smaller. For example; Sennheiser stating the THD on the HD800 is lower than 0.02% in 1KHz 1Vrms (102db on the HD800), same as measured by Jude on HF (on the comparison with the Focal Elear\Utopia), while on this site and Innerfidelity, it seems to be more like 0.2%. Same with other headphones (Focal Elear\Utopia included)

    How come?
     
  2. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    The phones are much better than 0.1% distortion, it's just that it's not easy to measure. Trying to measure the HD800 midrange distortion you easily get into the limits of cheap measurement rigs. For example with my UMIK I measured less than 0.05% D2 at 95db at 500Hz and even then I'm sure it was limited by the measurement rig.
    It also comes down to how the distortion measurements are created. With longer FFTs you can get rid of some of the noise to some degree. This is why the tone FFTs always show a lower noise floor than the sine-swept generated distortion measurements.
    Generally the lowest distortion measurements are most indicative of performance, as long as the levels are the same. Maybe different types of coupling can throw off distortion measurements in some cases and some transducers, but as long as it is comparable, lower should always be more accurate.

    Regarding Tyll's results. I believe they are simply limited by the processing and the mic capsule. Jude's measurements are weird because his plots go down to levels that aren't even possible to measure and you can't easily see where they are limited by noise. It is vital to understand the relationship between FR and SNR at a certain frequency.

    I should also add that it's super important to understand what orders of harmonic distortion you're dealing with. Very low levels of higher order harmonic distortion may sound worse than higher levels of 2nd or 3rd order. In amp builders there is also some discussion of which order is preferred, 2nd or 3rd order. I seem to prefer low levels of D3 in the bass over the same low levels of D2, from experience with different transducers. High levels of D3 sound really shitty, likely even worse than D2.
     
  3. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    Tyll and some of us do not measure using 1 Vrms @ 1 kHz. Instead we use 90 dBSPL @ 1 kHz, which may push the headphone harder depending on it's sensitivity. The harder you push the headphone the more it will distort.
     
  4. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Also: Regarding some measurements on this website: It is likely that we will see changes, improving the performance of the measurements soon. If you know where to look you might know already.
     
  5. 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
    It's also unknown what SPL (or V) Jude's distortion measurements are at. Distortion levels and curve characteristics will be different at different SPLs. It is absolutely crucial to know the output for distortion measurements, otherwise they are useless.

    As for Sennheiser provided specs, they are marketing. I would propose that you take a look their provided X-axis lengthened, 100Hz cut off, 1/2 or 1/3 octave averaged block FFTs frequency response graphs which seem to show amazing neutrality and bandwidth specs which state 14-44,100 (-3dB). Neither of these pieces of information are consistent with any other non Sennheiser provided measurement of the HD800. Moral of the story: 1) manufacturer's specs tend to be fudged. 2) Jude is a measurement noob, despite having very expensive measurement gear.
     
    Last edited: Aug 22, 2016

Share This Page