Introducing the Randy Headphone Measurement System

Discussion in 'Measurement Techniques Discussion' started by Hands, Oct 20, 2018.

  1. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    After many, many delays and much procrastination on my end, I am finally ready to introduce details and preliminary measurements for my Randy headphone measurement system!

    I am still trying to decide if "Randy" is going to be just "Randy," the name of the rig, or if it should be an acronym for something.

    $5 Paypal gift to whoever comes up with the best acronym for Randy, and an acronym I actually want to use.

    Behold the splendor of Randy!

    [​IMG]


    Background on the hunt for the Randy Rig:

    First, some background. My current method of measuring headphones uses a small mic, glued into a tri-flange IEM tip, which I insert into my ear canal. I try to keep the mic as flush with the opening of my ear canal as possible. This plugs into a little battery box and then into the mic-in of my Creative SB1240.

    One big downside to this is that it isn't comfortable. In fact, it hurts pretty bad, especially since I made the mistake of re-gluing the mic and tri-flange tip with super glue. Super glue isn't nearly as pliable at hot glue!

    That, and the mic starts to slip out of place once the ear wax starts attacking it. Or something like that. It's annoying.

    Another big downside is I can't fit this mic in my right ear properly, so that's why you almost never see right channel results. If you do, it's because the headphone was designed in such a way that I could just wear the headphones backwards, or similar, and measure the right channel on my left ear.

    On the upside, this method used a real head and ear. Sure, my ears are my ears, not an average ear from the results of 300+ humans and certified by old wizards that sit on an audio council. And I actually quite like how the results look with the mic at my ear canal, especially when using measurements to help describe what I hear. I only ever compensated for some bass droop from the battery box.

    The MiniDSP EARs rig looked promising, and it still has a lot going for it in terms of price and physical consistency for each user (mic consistency TBD). But I found it just a hair too small for my taste, width and height, and the ears appear to be a simplified representation of human ears. I wanted something more true to a real human ear, and even the to the contours of a real head rather than a flat coupler if possible. Also, I didn't like the way they used an ear canal. The way I see it, if you aren't going to accurately simulate the ear canal on a measurement rig, don't bother. But I'm really just a monkey pushing buttons compared to the real experts, so take my opinion for what little it's worth.

    This led to the following requirements:

    1. Measurement results in the same vein as my current rig, if nothing else for ease of transition. I'll never be able to get a perfect 1:1 match, but I wanted this to keep on a similar course.

    2. Ears meant to more closely mimic real human ears, averaged or not.

    3. A head more closely proportioned to that of my own. Preferably, something that actually resembles a head and not a flat coupler.

    4. If ear canals are not accurate to real canals (basically impossible on a budget), the ability to use mics flush with entrance of ear canal.

    5. Ability to measure both channels without sticking anything in my own ears ever again.


    Enter the Randy Rig

    After some searching, I came across the Binaural Enthusiast B1-E Dummy Head with mics and battery box. While meant to be used for binaural recordings, I figured it would suit my purposes.

    This seemed to tick most of what I wanted for a few hundred bucks, and with how the battery box works, it was almost a direct drop in replacement for how I hooked up the in-ear mic rig.

    The head has a width more similar to mine (still a hair less wide, it seems) but did require me to add some cushion at the top to add more height for appropriate fit. I'm also lazy and don't like to have to adjust the headband when I transfer headphones from my head to the dummy head.

    What I really like is that the ears on this are modeled after real ears and are somewhat similar to my own. Of course, this means I still won't have results for that of the average human ear, nor do these ears have the pliability of the latest, greatest ears that GRAS uses. This also didn't cost me thousands of dollars.

    I had to remove the mount attachments, which was no problem. I also had to move the mics so they sat flush at the entrance of the ear canals. Since I'm not a robot, they're only as closely matched to the ear canal entrance as I could humanly manage. But I imagine that will be a minor source of variability between the two ears.

    I believe this utilizes the Primo EM172 mics, also seen in at least some of the 3Dio binaural measurement rigs.

    Once I had everything stood up from a hardware perspective, I got to work on a compensation curve.


    Preliminary Measurement Results (Custom Compensation Curve, Left Channel Only)

    Since I was pretty happy with how my results were turning out with the in-ear mic as of late, I wanted to try to base it off that. I'll admit that I also applied some influence from the target curve @Serious made for Tyll's measurement results a while back. (Thanks, Serious!)

    The process behind this mostly involved me measuring a few headphones, a mix of open and closed back over-ears, and dumping the results with 1:1 and 1:3 smoothing using the current, in-ear method. I then gathered raw, uncompensated results of the same kind from the Randy rig. I took a difference of the two and made some custom tweaks of my own, to put it simply. The 1:1 results are used for everything below 1KHz, with the 1:3 results used above 1KHz and in tandem with said tweaks of my own.

    So far, I've only been able to work out the compensation curve for the left channel. The mics aren't perfectly matched, which isn't too surprising, so I still have yet to work that out. However, I wanted to get feedback on what I have so far with the left channel before moving on.

    Here are a few headphones with the applied custom compensation curve compared to the in-ear rig:

    HD6XX Left Randy vs In Ear.png
    HD58X Left Randy vs In Ear.png
    EMU Purpleheart Left Randy vs In Ear.png
    MDR-MA900 Left Randy vs In Ear.png
    Mobius Ballistics EQ Left Randy vs In Ear.png


    I'm happy with how the HD6XX and HD58X results turned out. Those headphones are really easy to measure, in that they are very consistent and are less likely to show massive variations between measurement rigs. You can almost always identify similar traits for these headphones regardless of what they were measured on. Almost no need to worry about fit and seal, so long as you have adequate clamping pressure.

    I expected, and received, greater variation between the two rigs with the MA900 and E-MU Purpleheart in particular. The MA900 can sound noticeably different depending on how you fit it on your head. With the in-ear method, I have physical feedback ensuring I have it seated like I would normally. I can't do that on a dummy head.

    In fact, I did get at least one measurement of the MA900 with less of a treble spike using the Randy rig. But I thought it might be more accurate to get an average of a few varying positions. I didn't do that so much on my own head, because I know how I like it to sit. (Should I go back to including a graph of all the measurements takes I averaged together in this case?)

    The E-MU Purpleheart is barely over-ear, but I tried to fit them similarly. Nonetheless, because of it's on-ear/over-ear in-between nature, I expected some variability. This one is a bit tough to get just right, but it's a good use case.

    I know the Mobius is somewhat of an unknown right now, but I wanted another sample to use for getting some average deltas between the two measurement rigs. It works well as a closed, over-ear headphone with minimal fit and seal fuss. So, while not well known, you can see the results nonetheless (with ballistics EQ setting).

    Besides, with a different head and different ears/pinna, you're never going to get the exact same results. I personally think these are perfectly serviceable results, given unknowns and variability will always play a role in headphone measurements. But what do you think?


    Steps Moving Forward

    1. Based on feedback about left channel results, will finalize that compensation curve.

    2. Move onto right channel results from there. Target being that the left channel should measure the same on both ears.

    3. Distortion results still TBD. This mic does some things better than the in-ear mics (D3, D4 mostly) and some things worse (can't handle as much SPL, slightly higher D2 "floor" overall). I may need to incorporate the MiniDSP EARs for distortion results to get an accurate picture.

    4. Still reviewing CSD results for Randy rig. So far, so good, though. Will share later.

    5. Hoping to have moved on to using Randy rig exclusively by 2019 and giving my poor ears a rest.
     
    • Like Like x 31
    • Epic Epic x 1
    • List
  2. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,589
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Holy hell, this looks like what I was trying to achieve, except way easier. Looks much better than the EARS and I actually want one now!

    I was actually planning on trying to DIY a dummy head - eventually. But this looks awesome and actually worth the money. Just a bit unfortunate if it really uses the Primo mics. I tried those and to be blunt, I much preferred the way the results looked like with my WM61A. There was literally no redeeming quality to the EM172 vs the WM61A, except its ease of use.
     
    Last edited: Oct 20, 2018
  3. Biodegraded

    Biodegraded Friend

    Pyrate Contributor
    Joined:
    May 28, 2017
    Likes Received:
    7,952
    Trophy Points:
    113
    Location:
    Vancouver BC
    Roughly Accurate Noggin Duplicate, Yeah.
     
  4. direstraitsfan98

    direstraitsfan98 D2Girls v2.0

    BWC
    Joined:
    May 11, 2018
    Likes Received:
    296
    Trophy Points:
    63
    Location:
    Canada
    Is it supposed to look like a gimp?
     
  5. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    I think you can order it without the mics and battery box. It's not too hard to get your hand up in the head to move or replace mics.

    I've considered putting the EARs mics in there for ease of use through USB and better distortion results. But I'm already this far into it and am not sure I want to deal with the hassle. I'd probably have to find a way to add some width to the mic body too so it'll fit, like maybe wrapping some adhesive foam around the edge of the mic's body.
     
    Last edited: Oct 20, 2018
  6. Taverius

    Taverius Smells like sausages

    Pyrate
    Joined:
    Dec 27, 2017
    Likes Received:
    3,026
    Trophy Points:
    113
    Location:
    Rapallo, Italy
    What a handsome lad he is!

    Have you thought about replacing those big fat screws with countersunks?

    I can't quite get a handle on the scale but they seem like they might make it harder to get a good, consistent seal on stiffer pads.
     
  7. LetMeBeFrank

    LetMeBeFrank Won't tell anyone my name is actually Francis

    Pyrate
    Joined:
    Aug 4, 2017
    Likes Received:
    3,758
    Trophy Points:
    93
    Location:
    Jackson, Mi
    Really Awesome Noise Dispensing Instrument

    R.A.N.D.I.
     
  8. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    I have considered it but am worried about screwing things up.

    I have also contemplated that it may be good to have something on the coupler that can disrupt a 100% perfect seal. I might even try to add a small amount of fake hair around the ear eventually.
     
  9. 9suns

    9suns [insert unearned title here]

    Pyrate
    Joined:
    Jan 19, 2017
    Likes Received:
    1,754
    Trophy Points:
    93
    Location:
    EU
    Random Asshole Negates Direct Involvement = RANDI :confused:
     
  10. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    Some updates on developing the Randy system.

    I was testing out measurement consistency on the right channel, and things were looking noticeably off...to the point I really didn't want to try to make a crazy different comp curve for it.

    Based on what @Serious said, I assumed maybe the mics were very poorly matched. I double checked the mics in my EARs, and those seemed pretty consistent.

    So, I swapped the EARs mic into my Randy rig. Upside is Randy is now USB-compatible. No more 9V batteries needed!

    Downside is that left-to-right consistent is still not where I want it.

    I think what's happening is that these ears are modeled after real left and right ears. I know my ears have slight differences. One takes a larger IEM tip than the other, and measurements turn out differently depending on the ear I use.

    I guess it makes sense that Randy follows a similar rule. That's one upside to the EARs, its left-to-right consistency from a physical perspective. I just wish they were modeled more closely after real ears, had more width to the rig, and had a curved coupler more similar to that of a real head. You win some and lose some.

    On top of that, the mics from the EARs measure differently than the stock Primo mics, so I have to re-do the comp curve all together. And despite my best efforts, I'm going to end up having a separate curve yet for the right ear on the Randy rig.

    Such is life, I suppose. Still, getting 90% of the way there with dual channel consistency is better than me not being able to take right channel measurements at all.

    I'm going to email the guy that made this head and see if he can't get me an exact replica of the left ear and mirror it for the right side.
     
  11. geniekid

    geniekid Facebook Friend

    Contributor
    Joined:
    Mar 13, 2018
    Likes Received:
    137
    Trophy Points:
    33
    Location:
    Chicago, IL
    Rigorous Audio Noggin DummY
     
  12. gbeast

    gbeast Mighty Moral Power Ranger

    Pyrate
    Joined:
    Nov 8, 2015
    Likes Received:
    1,723
    Trophy Points:
    93
    Location:
    Phoenix
    Home Page:
    Sorry to be random but I actually have that head and it is in my closest never to be used again. If someone wants to buy it lemme know. It is not in pristine condition like hands.

    Good to see my dream come to fruition through someone else. Don't have the time to be into this stuff as much anymore.
     
  13. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,623
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Is there a ear canal or are mics mounted flush where ear canal would be?

    You got PM.
     
  14. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,623
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    J.U.D.E.

    Justifiably Undoing Dishonest Ears
     
  15. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    There are ear canals as seen in diagram here:

    [​IMG]



    However, by default, the stock mics (if purchased with mics) are mounted at the entrance of the ear canal, just as I configure it myself:

    [​IMG]



    I think the main reason for this is the mics come installed in a relatively large plastic housing, so they occupy most of the ear canal space:

    [​IMG]



    If you unscrew the plastic housing, you'll find the mic is a small piece itself. Actually, it's almost the exact same size as the mic in the EARs if you include the rubber piece that goes around the EARs mic. Tiny capsule.


    LOL! Defeats the purpose of me wanting to name it Randy, but I like it nonetheless. Will keep the rename on the table as an option.
     
  16. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,623
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    I don't think (really need more data) I am a big fan of ear canals and would much prefer flush mounting. G.R.A.S. might have gotten the closest to modeling a real ear canal (no high Q resonances), but I am not sure of the importance of this for accurate measurements after appropriate compensation.
     
  17. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    Right, I'm sticking with the flush canal mount myself for ease of getting decent results. Results seem good enough with right compensation curve. And this head allows you to do just that, no problem. (Though, reaching up inside the head is a bit of a pain with how tightly packed it is with foam. And watch out for the screws!)

    It's the slight physical variations between each ear that make dialing in a comp curve easy. I emailed the guy that made this about getting a flipped left ear to have total consistency. I'm more keen on the results the left ear produces. In hindsight, the included mics + battery box weren't too bad.
     
  18. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,623
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    The miniDSP still took quite a bit of work to get proper L and R curves. The difficulty was was made worse by deltas introduced by headphone positioning. I did get good at consistent positioning with practice. I actually averaged the L + R deltas of several headphones, assigning different weights. Some headphones like the HD800 seem super sensitive to positioning in the last half octave up top. Others are less sensitive. Just an royal PIA because different headphones and very slightly different positioning yielded different L + R deltas.
     
  19. Hands

    Hands Overzealous Auto Flusher - Measurbator

    Staff Member Pyrate MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,285
    Trophy Points:
    113
    Location:
    Colorado
    Home Page:
    The EARs at least has the advantage of symmetrical ears, as simplistic as they are. The BE head seems modeled after a real pair of left and right ears, including all subtle differences between them from the actual subject. Kinda cool, I guess, but not exactly what we'd want here. I think our brains are pretty good at calibrating out those differences, so a rig shouldn't hinder channel consistency results.

    So, taking headphone positioning and mic differences out of the equation, the left ear might measure a peak around 10KHz, and the right ear will have a HUGE drop between 7-10KHz and shift the spike to around 12-13KHz. You can swap mics between channels, try different mics entirely...always the same thing for me.

    And this is with headphones that measure consistently, like the HD650.

    I'll have more to share on this later, since I'm re-doing the curve since I swapped in the EARs mics. Hoping I can just get a symmetrical ear from the dude that makes these and call it good. If not, the curve for my right ear is going to be funky.

    Totally agree on the pain of measuring placement, fit, and seal sensitive headphones. The MA900 is a pain for me. I usually just get like 15 different measurements in varying positions and have a spreadsheet ready to go where I drop in the results and get an average result. I mean, I take some care making sure the results are valid, but I can churn stuff out without too much work or time these days.
     
  20. spwath

    spwath Hijinks master cum laudle

    Pyrate BWC
    Joined:
    Dec 13, 2015
    Likes Received:
    7,894
    Trophy Points:
    113
    Location:
    Madison, WI

Share This Page