Can't make my measurement rig work properly

Discussion in 'Measurement Setups, Systems, and Standards' started by przem, Sep 9, 2016.

  1. przem

    przem MOT: Shipibo.Audio

    Joined:
    Feb 11, 2016
    Likes Received:
    70
    Trophy Points:
    13
    Location:
    Poland
    Hello everyone,

    Couple of weeks ago I decided it's time to make a measurement rig, which I did. But I can't make it work properly for some reason - I've tried different things, read those forums and it still doesn't work the way it should, details below.

    In advance I'd like to apologize for the the mess and chaos in this post, I didn't know how to tackle it, and I'm a bit tired and angry that it doesn't work... :)

    First, about the rig:
    Microphones are WM61As from ebay.
    Mics are connected to a circuit board with capacitors, resistors, etc. (pic below)
    Casing is made from plumbing pipe and plywood. On the inside of the plumbing pipe I've put a bitumen mat for damping (but that proved to be useless)
    As for the microphone housing, I've tried few different ones, but none of them seemed to work properly.
    Rig is connected to the Scarlett 2i2 through XLR connectors.

    Ok, so what's the problem?
    I compared my FR graphs to those from Innerfidelity and other people, and they don't look similar, there's either too much bass, or a massive drop in the mids, or there are too much highs, depending on mic housing I use.
    For the needs of this post I used B&O Beoplay H6 headphones and compared them to Innerfidelity graphs:
    http://www.innerfidelity.com/images/BangOlufsenH6.pdf

    Let's start:
    First, I put the microphones in the wooden tubes I made and attached them to CDs padded with PVC foam
    [​IMG]

    Excuse the "ear", didn't work out the way I planned, but it seems to have very little influence on the measurements anyway.

    [​IMG]

    Inside of the rig.

    With this setup, FR graph looks most similar to Tyll's, but there's way too much bass:

    [​IMG]

    So I figured I'll add more damping to tame the bass and I put this synthetic wool.

    [​IMG]

    Which had little influence on the graph.

    [​IMG]

    There's still too much bass. So I went on to add soft memory foam inside for some more damping.

    [​IMG]

    But that only bumped the mids a bit and had no effect on the mighty bass.
    [​IMG]

    So I figured that the problem must lay in the wooden housing, and binned it. Instead I went for a soft memory foam circle with pvc dense foam glued to it on both sides to stiffen it up.
    [​IMG]

    And that did make a significant change to the FR graph.
    [​IMG]

    As you can see, bass is now tamed and more in level with higher frequencies, but I get a massive drop around 200Hz, which definitely shouldn't be there (or at least shouldn't be that visible).

    So I continued to tinker with that and put the synthetic wool back in, which did help a bit to get rid of that drop, but boosted bass again
    [​IMG]

    Lastly, I figured I'll try just memory foam, without any stiffening, but that didn't help me much either:
    [​IMG]

    Graph for this:
    [​IMG]

    It tamed the bass as I wanted, but the drop is still there + highs are messed up now as well.

    I tried few different sources and headphones, recalibrated everything numerous times in ARTA and results are always skewed in a similar way, so it's either ARTAs fault (which I find doubtful), or my faulty rig design, or the materials I used.
    I ran out of ideas and I hope for your insight super best audio friends. As you probably figured out by now, I'm totally new to all this, so please don't be too hard on me. :p

    Thank you everyone for reading and your (future) input.
     
  2. philipmorgan

    philipmorgan Member of the month

    Pyrate BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    3,790
    Trophy Points:
    93
    Location:
    In the wind, so to speak
    Home Page:
    Do you have any circumnaural headphones other than the B&O to measure with? Might help troubleshoot what's going on if you were measuring a more commonly-measured headphone.
     
  3. przem

    przem MOT: Shipibo.Audio

    Joined:
    Feb 11, 2016
    Likes Received:
    70
    Trophy Points:
    13
    Location:
    Poland
    I could get a hold of Fostex TH900 and some Beyerdynamics - most likely DT 770 pro, would that be of help?
    Or maybe you could tell me what headphones you prefer and I'd see if I can get a hold of them.

    Thank you very much!
     
  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
    Oh, the pain of getting a realistic acoustic seal for closed headphones while trying to get "anechoic" results in the mids and treble. Depending on how long and thick your WM61A cable is you might try sticking it in your ear canal to see how the bass looks like on your head. Dummy heads probably also don't get the most realistic seal. Making a good coupler that works for all headphones isn't easy. I'm not convinced that Marv's V2 coupler is realistic, either.
     
  5. przem

    przem MOT: Shipibo.Audio

    Joined:
    Feb 11, 2016
    Likes Received:
    70
    Trophy Points:
    13
    Location:
    Poland
    @Serious I tried, but can't point the microphone in the right direction inside my ear it twists and turns, any suggestions? I might ask someone to tape it to my ear, lol...
     
  6. philipmorgan

    philipmorgan Member of the month

    Pyrate BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    3,790
    Trophy Points:
    93
    Location:
    In the wind, so to speak
    Home Page:
    In my limited experience, Sennheiser HD6xx headphones are less sensitive to coupler placement, seal, etc. Those would be good ones to try to see if you're issues are related to coupler seal or mic compensation or something else.
     
  7. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    You can still try to measure. It's only really to check for the seal of the pads. The WM61A isn't directional until the very end of the audible spectrum (in a free space). If you want to do real in-ear mic measurements I would use an eartip and a thin cable and seal off the end facing the ear canal.
     
  8. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    It's not possible to get the exact same results of Tyll because Tyll runs a dummy head with an artificial ear. On top of that, Tyll runs compensation to translate the measurement into somewhere between a diffuse and free field curve measurement - or something like that.

    Finally make sure you have a good seal between the area in front of the microphone and the area behind it; or better yet, eliminate the cavity / tube behind the plate by using a stack of foam without a tube.
     
  9. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Another tip. I also use strategically placed strips of foam to simulate a imperfect seal, from hair and from the nape of the neck. Seal will effect bass measurements headphones with enclosed internal areas to the ear (sealed traditional dynamics or orthos). The effect on the bass will also depend on driver characteristics.
     
  10. przem

    przem MOT: Shipibo.Audio

    Joined:
    Feb 11, 2016
    Likes Received:
    70
    Trophy Points:
    13
    Location:
    Poland
    Thank you guys so much for all your replies, you're gold and your help is invaluable.
    Great, I have an access to HD650, just need to find a cable for it. Will try to measure them this week.

    @Marvey I do realize it's too much to ask for a plumbing pipe and ebay mic to make up for a professional, expensive equipment, I used Tyll's gaphs as a benchmark, and just hoped to get slightly more similar results, but I'm pretty new to this and didn't realize a lot of this stuff you guys are writing about.

    I followed your advice and stuffed the inside of the rig with stacked foam discs, and used scraps of foam to create an imperfect seal and the results are much more accurate than before in my opinion, graphs below.

    [​IMG]

    The one above is for the stacked foam before I simulated an imperfect seal, bass is more tame than in previous measurements, drop around 200Hz also isn't that visible, and highs look fairly similar to those from Tyll's measurements.

    [​IMG]

    Here's the same setup, but with a piece of memory foam between the disc and and the earpad. Low frequencies are now in line with 1kHz as they should be, highs got messed up slightly though, but it's still much better compared to my previous tries.

    Thank you again, Marv, that was super helpful.

    @Serious I did put it in my ear too as you said and the initial measurements look promising too, I'll tinker with this more tomorrow, but I like the results. What I don't like though, is that I still hear the pink noise long after I stopped measuring. :D
    Graph for this:
    [​IMG]
     
  11. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    This is my real measurement rig: (there is a CD under the foam disc). Yup, all my secrets here have now been revealed.
    DSC_2939.JPG

    This is my fake measurement rig. I posted this because I got sick and tired of certain people e-peening how expensive and IEC / ANSI / ISO /AES approved their measurement systems were when it was obvious that they were idiots. It's the man, not the machine. Especially in this case where transducer errors are so high, on the order of several magnitudes, compared to electronics.
    DSC_01255309.jpg
     
  12. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    @przem Someone has to make an universal headphone measuring jig and sell it for 100$ (without mic) so we can finally have standardized measurements. That still doesn't mean that I will stop using my special piece of foam (that I got for free) :p.
    (I've thought of such a thing but it's really difficult to design a good coupler, so difficult that I've stopped using a coupler and started using my own ears.)

    That actually looks better than the fake one. So is it like the other one but with creatology foam? Solid plate behind the creatology?
     
  13. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Pinna sized piece of felt, on top of Creatology high-density foam, on top of a CD, on top of a stack of medium density foam. All secured to the edge of a wire frame rack.

    The top piece of felt was added to bring extraneous ringing and other artifacts down to where it was most similar to pork belly. CSD results without this piece of felt, directly on the Creatology foam, were too ringy. A side benefit is that I could measure supraaurals more accurately than before as the felt mimics a weak seal on pinna.

    I never really used the C.U.N.T. It was always semi-B.S. used to mock people who couldn't stop mentioning G.R.A.S. or Audio Precision every other sentence, as if expensive tools suddenly imparts one with the aptitude, patience, and passion to do this stuff.
     
  14. Madaboutaudio

    Madaboutaudio Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Singapore
  15. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,777
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I know. I love that. Jude's so obsessed with SuperBAF that he so easily takes the bait.

    And oh!, that cute baby blanket. I hope we didn't get rid of it.
     
  16. Madaboutaudio

    Madaboutaudio Friend

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    545
    Trophy Points:
    93
    Location:
    Singapore
    The C.U.N.T is so special that Dictator kim forgets to cens.... nevermind.. maybe it was before everything went nuclear.
     
  17. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    That actually looks more similar to my V2 style coupler, mainly because the felt is ear shaped. I preferred a slightly thicker felt here and even that wasn't realistic to my ears. But it did measure similarly to the chicken that I measured, but not similar to when I measured my cheeks or the top of my head. Maybe I have to shave my head for science?

    Here are my two couplers without awesome DOF:
    tmp_11116-20160912_003130462018634.jpg
    The V1 type coupler on the left absolutely needs air behind it. I still have a slightly thicker piece of the same material that might be a little better. This coupler didn't work with the TH-X00.
    The V2 style coupler is different in that I use a cut piece of a latex or similar glove over the CD. Right now there are two pieces of closed cell foam on there. It's been long since I've used it but I thought two pieces were "better". The blue one is slightly thinner. The blue in the back and black in the front seemed to give the best measurements.
    With my weird modded (semi open) Superlux can, the V1 coupler got more realistic results as the V2 got too strong of a seal (even with a thick piece of felt) elevating the whole region below 2kHz by 5db or so, especially the bass. Both couplers don't work for every headphone.

    The ear canal opening measurements have the benefit of the most realistic seal and accurate response up to about 4kHz, from where on ear resonances make it harder to interpret the measurements.

    In the end none of these measurement rigs will measure the Gaurav-factor and putting headphones on a coupler will never be as fun as putting vinyl on the turntable.
     
    Last edited: Sep 11, 2016
  18. przem

    przem MOT: Shipibo.Audio

    Joined:
    Feb 11, 2016
    Likes Received:
    70
    Trophy Points:
    13
    Location:
    Poland
    100$ would be way too much! Parts for mine cost pennies (minus the Neutrik XLR plugs and Scarlett 2i2, but those could be easily replaced). I'd still probably pay that 100 for a reliable design, though. I hope "someone" does it one day.

    I agree with that in general, it's just that I'm making leather accessories - so far I was making headbands mostly, which don't affect the sound in the slightest - but lately I started learning to sew earpads, and those can change the sound of the headphone a lot, and I wanted some objective, comparable data instead of my subjective feelings.
    With the advice of you guys I should be able to build a setup sufficient for my needs.

    @Marvey I love the inscriptions on this, especially "Fast Uncompensated Joint and Unbiased Device Evaluation Technology" bit, it all makes sense to me now. Impressive. :) Also, looks like I've been using wrong types of foam for all this (too soft), shame they don't sell that creatology foam in Poland, gotta find substitutes.

    Thanks again for all your input and help guys, you saved me a lot of pain!
     

Share This Page