2016 July/August NY area measurements by Serious

Discussion in 'Audio Science' started by Serious, Jul 30, 2016.

  1. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    I dunno about the Plankton-O-Matic Meter, but this FIR filtering stuff I was mocking around with brought more Rezolooshun to my HD600s somehow. It does not cure cancer though.

    Wifey liked it too.
     
  2. 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 all close enough. As long we we don't have Sean Olive's proposed mega-bassazilla target, I'm fine. Why do I sense that Harmon lined up JBL K2 speakers in small room with the microphone three feet from them?
     
  3. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    LOL! Yeah, I dunno about those other mega-bass targets. B&K works for me. Great call.
     
  4. OJneg

    OJneg The Most Insufferable

    Pyrate BWC
    Joined:
    Sep 25, 2015
    Likes Received:
    3,923
    Trophy Points:
    113
    Location:
    Grand Rapids, MI
    Fixing FR does bring out more Rezolooshun, but it doesn't turn turds into diamonds. Nor silver into gold.
     
  5. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    True. One needs to start with a decent set of cans and even then manage one's expectations. IMO it applies to all. There is only so much one's uber-DAC and uber-Amp can do with a set of crappy cans or such.

    Some things scale well. Some others not so much.
     
  6. 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's just one example and the HD600 usually ends up looking fairly similar with many measurement approaches. I also think that @Hands changed his methods since the HD600 measurements. His HD650 measurements do look more like my in-ear-mic measurements than your coupler measurements. But his 2014 HD600 measurements do look weird. Well, his HD650 measurements from back then also look normal, so maybe his HD600 was different?
    But yes, in-ear-measurements will always look slightly different from one person to another which is why Hands mainly uses them to back up his subjective impressions.
    On the other hand I felt his Ether and Dharma plots more accurately captured the midrange and treble balance to me and I don't think this is simply a matter of using the right target curve for the coupler measurements.

    Your HD600 measurements do look very similar to my in-ear-mic results. Less bass extension, more 2kHz and generally more energy from 4kHz on with mine. But my own V2 measurements for the HD800 (both stock and unmodded) don't look like my IEM results at all.
     
  7. 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:
    Yeah, I went through a couple different compensation methods early on. I think the HD600 was right in between that and my final compensation. I think I posted a mixture of things for that headphone. Then I got a new WM61a mic in and a new tri-flange tip to put it in that better fit my ears. More flush with canal opening, more secure overall. That changed some results too. I've been pretty consistent and had the same setup since the beginning of 2015.
     
  8. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Quick update: I tried to compensate the FR for the effects of the in-ear measurements a little, to where I felt it was more accurate. I also improved the microphone compensation a little, at least that's what I feel. The compensation curve consists of a 3db notch at 5kHz and a 16db (!) shelf after 12kHz. The notch may not be needed, but I don't think I hear a peak here with my HD800 and I also generally felt it was more accurate with the other measurements, so I went with it. I think that's (part of) the concha gain and a slight bump here is preferable.

    Here's what my HD600 looks like now:
    HD600.png
    I think the channel imbalance is real. I might try to improve this when I get a chance. It sounds somewhere in between subjectively. I added a bit of felt under the pads to bring them closer to a like new condition, which made it sound brighter with less 3-4kHz emphasis, but also more of a 5kHz bump. I definitely hear a dip at 6kHz with the HD600 (when listening to sweeps).
    I think the rise after 12kHz or so is realistic because of the 24kHz resonance. Still, I would say the HD800 sounds the airiest.

    Here's what the Utopia looks like now:
    Utopia.png
    For the Utopia graph I also tried to compensate a little for the 6-9kHz hole that I originally had in my measurement. For this I increased the region between 6 and 9.5kHz by about 6.5db. I know it seems weird to change the FR in a way that seems closer to right subjectively, but I really didn't like how the measurement came out and I'm pretty sure that if I repeated the measurement I would get at least 5db more in this region. The sharp nature of the dip in the original measurement does seem to confirm that it was an artifact.
    I think this measurement looks a little bassy, but it could be accurate. Maybe 1db less would be more accurate.

    Finally, here's what my HD800 looks like now: (The mods are still the same.)
    HD800X.png
    50Hz spike is the terrible 50Hz power hum that I get with my in-ear measurements. About a 1db step after 2kHz, which I preferred to flat. It's hard to get consistent measurements, especially above 6kHz with the in-ear measurement method. I hear a bit of a 7.2kHz spike, which seems to be related to an interaction between headphones and my ears because I also hear it with the HD600s. The rest of the treble seems to be very flat.
    Overall it's hard to compare the tonal balance to the HD600, because not all things are equal. I would say my HD800 is maybe the most mid-centric of the three headphones, but not in a bad way. I'm fairly certain that the stock HD800 cable (which I'm still using) sounds bright and artificial, while the HD600 cable actually probably sounds slightly mellow or even mid-centric. From memory the HD650 cable sounded even warmer. The HD600 also sounds more damped with less attack on transients. Less snappy. Both the HD600 and modded HD800 pass the 'Keith Don't Go' (or RAM - Fragments of Time or Porcupine Tree - Trains or Nirvana, etc.) test for me. The HD600 actually sounds slightly more problematic on some material, while the HD800 sounds brighter on other stuff. I actually find the HD600 to be rougher up top, but that might just be me.
    It might be worth a shot to buy the HD600 cable and terminate it with HD800 plugs and a 4 pin XLR plug on the other end. I much prefer the XLR to 1/8 inch plug with 1/4 inch adapter that I made when reterminating the HD600 cable to the XLR to 1/4 inch plug that I made with the stock HD800 cable. In other words, the stock HD800 cable sounds like ass. I think SPC sounds the worst.


    Anyway, do you like the results of the in-ear measurements with the compensation?
     
  9. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Wanted to share a quick experiment I did with you guys:

    I wanted to explain why I think the speaker curves need the B&K target (or similar) and the in-ear measurements don't. Maybe @Hands will find this interesting.
    What I did was to measure the Riva S from across my room. In a nearfield listening situation, or without enough backwall reinforcement the Riva S will sound (and measure) slightly bright. Brighter than my HD800. But that wasn't the case here - it was relatively close to the back wall and at about 6m distance. (The great thing about the Riva S is how well thought out its off-axis response is. The off-axis response is probably more impressive than on-axis, but that's not what this is about.)
    BTW: This is not about the Riva S measurements at all. I could get them to look very different if I wanted to.

    I did two measurements each. Two measurements with the microphone in my ears with my head facing the speaker directly and two measurements with the microphone at the same position (or as close as I got), but without it being in my ear. I calibrated the in-ear results with the new calibration file. I could've taken more than 2 averages, but they were already pretty close. I cut off the measurements at about 55Hz because they were useless below that. The Riva S just doesn't go any lower.

    The measurements do change slightly as we change the angle to the speaker closer to something we'd find in a stereo setup. I personally think the speakers should be voiced to counter these changes - that would be a slight dip between 1-3kHz, but nothing major. Maybe 1-2db. Of course that's a very different topic.
    (For those of you looking at this for the Riva S measurement, let me tell you the treble extension would be much better if I tried to properly measure the FR of the speaker.)
    Idk 6 1dboct100Hz.png
    The main thing I wanted to show you was just how closely the 1db per octave roll off from 100Hz target (similar to B&K, but a tiny bit bassier) with an omni mic and the in-ear measurements match. I used the 1db/octave target because the B&K target didn't match it quite as well. Essentially the whole region up to 2kHz (where the outer ear gain starts to matter) is an almost perfect match. I also attached the difference between the two curves. After 5kHz the in-ear curve actually goes below the omni mic curve but this will depend on a number of factors. I wouldn't put too much trust into it here. That the in-ear curve looks so weird after 10kHz is because of the steep compensation that I use to make sense of the headphone in-ear measurements. That doesn't seem to work as well for the speaker measurements.

    So what does this mean for me?
    When the in-ear measurements for a headphone look flat up to 2kHz, it will most likely sound neutral up to 2kHz. After that it starts to get more complicated. I feel headphones should not measure with the same ear-gain bump from 2-5kHz as the speakers, at least not at the ear canal opening. (And also not at the ear drum, like I explained before somewhere here.)
    It also says that speakers that measure flat with no compensation will sound too bright. The exact target that you should be using will also depend on a number of factors. Sometimes I prefer the B&K target, sometimes I prefer the 1db/octave/100Hz target. Simply EQing with an omni mic at the listening position will not yield the same perceived frequency response for every room/speaker. Even studios seem to forget this. I think Darin Fong has a few studio presets in his software, all of which sound completely different.
     

    Attached Files:

  10. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC
    Hi I know this is old thread but I was just comparing the HD800 measurements you have here as I have been catching up on measurements,
    And like to know why you have the amplitude at only about 74db, when in the measurement rig thread most others use higher levels around 90db.
    I have not noticed much difference on open headphones,
    But closed headphones do seem to look better at lower volumes..

    Anyways I have measured my HD800 and can verify less ringing in the CSD than others here so I have to attribute this to the extra mods I used, notably the dynamat extreme addition.
     
  11. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    I don't think the level was calibrated, but I definitely didn't run them at 90db either. As long as the SNR is high enough the FR shouldn't change. With speakers, especially small bookshelf ones you might even see compression effects at 90db. When it's late at night I'll even measure my speakers at 50db. Apart from distortion in what way did the closed headphones look better at lower levels?

    Personally I don't think the dynamat mods are usually visible in the CSDs for headphones or at least I haven't seen a difference yet. Smoother FR = better CSDs.
    What might be different is the vibrational behavior, but you need an accelerometer for that. Another interesting thing might be making CSDs from the excess phase plots, although there should really be nothing with headphones.
     
  12. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC
    Thanks for info, I have noted on some closed cup headphones,
    the cup cavity has less effect on lower volume.

    I was testing a portable and by mistake the volume was very low and saw a different FR.
    I decided to test all my headphones to look and grow accustomed to the data as it applies to the sound..

    Some members are testing around 90db, but I not feel that is very realistic,
    so I been testing with 1khz around 85.

    On updating the REW program,
    I seem to prefer it now over ARTA,
    As it now seems easier,
    Plus I like the CSD better now that it is colored.
    I not home but I would like to post the CSD for your analysis as it would be very interesting.
     
  13. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Yeah, I really like the CSDs now that they have rainbow colors. That's something I've been waiting for for years.

    Do you still have the measurements for the closed phones? That would be very very weird...
     
  14. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC
    I deleted it as an error but it was the bass area and it was the HiFiman Edition S that I modded with wood cups and increased the bass by slight removal of the paper backing at a corner, look:
    [​IMG]
    They are nice but the headband not ideal,
    [​IMG]
    plus I have settled on the Denon mm400 as my best portable.
    I need to rid of 2-3 other Frankenstein units lol.
     
  15. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    If it's the bass it's most likely a SNR thing. Lower SNR will at some point just make things look more linear until it gets buried below the noise floor. Usually there shouldn't be any difference in FR between low volumes and say 90db. You might get compression at 130db or higher depending on the headphone, but generally the level won't change the FR for headphones.

    BTW: I'm planning to get more serious about IEM measurements and also refine my in-ear method for headphones. That'll take a few months though and I'll need to calibrate my microphones first, too. The new MiniDSP dummy head thing also looks very interesting at least as a general reference.
     
  16. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC
    I remember the reasons of why not to use dummy ears as it is different with every ear so I hope they chose a good shaped ears lol.
    I really like the idea for uniformity, but The flat coupler Plat has advantages of being tuned and wider acceptance.
    I do hope those ear couplers are good though, but they using silicone so time will tell what types of resonances they will introduce.
    I hope they at least measure out like Tyll setup.

    Anyways, here is my HD800 CSD within a 35db range :
    [​IMG]
    You can easier see that the overall signature is more bass and treble by the colors, so it has the HD800 "air" since there is no SBAF rug liner that you gave me (thanks) as I gave that to member thegunner.
    Yet there is no 5-6k resonance just like yours:
    HD800X CSD.png
    Nice and clean in the aspect of all upper spectrum before 1ms.
    :)
     

    Attached Files:

    Last edited: Nov 4, 2017
  17. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC
    I Forget to add that I do not use the SD resonator mod either,
    as my unit had a different material to be like the stock 800S sound in the lower spectrum.
    HD800 S MOD FR.jpg
    As you can see my coupler is like member Hands which shows a bit more bass than member Marvey.
    I am not sure if I will change my coupler response yet.

    Forgot to mention that I sold this unit as well as the HEKV2,
    And am currently with the Dennon 7200 I modded.
     
    Last edited: Nov 4, 2017
  18. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    My coupler and CSD settings are very different, so hard to compare. With your measurements there's still a lot of that 6kHz peak. That it doesn't reach 1ms also depends on the settings. What does stock look like?

    Hands measurements will look more like mine. For my HD800 something like this:
    I'm fairly certain that the difference between two ears is way less than the difference between flat plate couplers and the ear. For general FR the ears are probably better. For CSDs in the treble region they're almost useless, but they will show us how much of the 6kHz ringing there still is. The peak will be much worse with the ears.
     
    Last edited: Nov 5, 2017
  19. Maxx134

    Maxx134 Dunning–Kruger effect poster boy

    Pyrate
    Joined:
    Oct 27, 2015
    Likes Received:
    818
    Trophy Points:
    93
    Location:
    NYC

    You are using both the SD mod And the SBAF rug liner for your reduction in resonance or ringing. That is a double wammy of damping.
    This to me sligtly mutes or dampens liveliness, making it sound a bit relaxed/tame.

    I am not talking about peaks, but with the trebles elevation..

    You say mine has a peak without mention the whole trebles range elevated after, which does not qualify as mine having a peak.
    There is no peak but an elevation without dips of stock.

    For me to have a higher 6k elevation without ringing, like yours, is truly a feat, as mentioned I not using the rug liner which kill the Soundstage air after 6k, or the SD mod.

    It is preferable by many to have more bass than flat, which why many consider the 800S an improvement.

    We know it(800S) does not address the ringing as good as the SD mod,
    But none of the past mod choices keep the trebles air or presence/brilliance region beyond 6khz.

    This is a major strength of the original 800, to have this region elevated.
    So I am of the opinion, that keeping the upper range and lower range boosted slightly,
    does give more lively sound as those other methods..

    I tried the SBAF mods which are on your 800 and they sounded excellent,
    But also very tame in comparison.

    You are of the opinion that a flat FR is best.
    I can agree with this only for a measument tool.
    No headphones I tested are like this.
    Both the Susvara and the Abyss Phi have elevated bass.

    Also, your dips in the trebles are to me, more of a result of the coupler whether it is your ear or a flat coupler.
    I don't believe it is the driver causing those dips, but the interactions inside.

    Anyways I do value all your opinion and my only real point is that I eliminated the 6khz ringing without loss or dampening of the upper spectrum.
     
    Last edited: Apr 28, 2018

Share This Page