DAC Amp Measurement Methodlogy and Equipment Discussion

Discussion in 'Headphone Amplifier Measurements' started by schiit, Nov 5, 2015.

  1. schiit

    schiit SchiitHead

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    6,229
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Valencia, CA
    Home Page:
    Measurements and certainty? LOL. Those who worship at the altar of the dScope need to know a few things:

    1. It is stunningly easy to get bad measurements.
    2. Measurements are not always reproducible.
    3. Measurements from unit to unit will differ.

    I can't tell you how many times I've gotten some real head-scratchers from:

    1. Bad cables.
    2. Interference (power cable over signal cable.)
    3. Someone welding in the next unit in the building, causing huge noise on our AC lines.
    4. Ground loops.
    5. Bad settings on the analyzer.
    6. Analyzer software bugs/crashes.

    This is after literally thousands of hours on a Stanford SR-1.

    So does this mean measurements are meaningless? Not at all. Just that you have to be careful about getting good measurements--especially when measuring products of an unknown performance level. For example, it's easy for me to go, "Huh, that's a weird result for a topology that should be measuring (like this.)" For Marv, it's a lot harder, because there is no general reference. I don't envy his job...and everyone should be profoundly thankful he's taking the time to do this, and do it right.

    (Aaaaand...as far as good measurements correlating to good sonic performance, let me swap out that Magni 2 for a Vali, and you tell me which you like better...and which has 100X higher distortion--yes, 100X, not 100%.)
     
  2. firev1

    firev1 Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    37
    Dislikes Received:
    0
    Trophy Points:
    18
    Location:
    Singapore
    Running signal generation and acquisition on the same machine is never a good idea IME Especially with the kind of hobbyist level stuff we are using.
     
  3. Ormia

    Ormia Rando

    Joined:
    Nov 5, 2015
    Likes Received:
    0
    Dislikes Received:
    0
    Trophy Points:
    1
    Once the unit arrives, if needed, I can probably 'rig' up something to make some academic measurements as well. I don't have audio precision, but I can use a combination of National Instruments Hardware and MATLAB/Lab View in my laboratory.
     
    Last edited: Nov 5, 2015
  4. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,602
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    What National Instruments Hardware? Me would like to know.
     
  5. Ormia

    Ormia Rando

    Joined:
    Nov 5, 2015
    Likes Received:
    0
    Dislikes Received:
    0
    Trophy Points:
    1
    Last edited: Nov 5, 2015
  6. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,602
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    Seems to be 16-bit. You may need a 24-bit interface.

    I was thinking something like this:
    http://www.ni.com/datasheet/pdf/en/ds-337

    However, that one is a module and not necesarily a cheap one. Factor in the cost of the chassy interface, and one end up in AP price territory.
     
  7. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,602
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    Moved this discussion here since we are talking on measurement equipment in general, instead of specific performance on a set of DACs.

    I find this discussion extremely helpful on it's own right.

    Please discuss any high performance ADC & AD/DA cards that may be suitable for audio characterization. Hopefully at reasonable prices.
     
    Last edited: Nov 5, 2015
    schiit likes this.
  8. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,602
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    One card keeps pooping out for cheap. The ESI [email protected] Anyone used it?

    There is also the E22 from Lynx with better specs, but a little more expensive.
     
  9. schiit

    schiit SchiitHead

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    6,229
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Valencia, CA
    Home Page:
    Makes sense. Sorry for the derail.

    I'll see if I can post up some comparative measurements from the SR1 and a SYS-2722 (Dave has access to one). They're usually pretty in-line, which also usually means the measurements are probably good. It may take a while to post this, though...

    Also forgot: in addition to Marv, let's also tip our hats to Tyll, atomicbob, JA, and others who go down the measurement path. It can be very illuminating...and it can cause a lot of consternation as well.
     
    bonehead and ultrabike like this.
  10. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    50,977
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    • I'll grab the Zana Deux Super and measure it for fun. I am almost certain the measurements will look horrible.
    • It could be that the measurement to subjective correlation for the Grace, Schiit, LH, and iFi units could be coincidental. Or perhaps measurements, because they only look at only very specific aspects of sound quality, are best applied with gear that is more similar to each other than different.
    • Despite the odd spurs in the high frequency of the Grace, no one is going to hear it. No one is going to hear a 9kHz spurious signal 70db down and several others 90db from the 1kHz fundamental. The Grace does measure well and it does sound good. The last thing I want is for people to go full-nwavguy on measurements. That would not be good and cause me to immediately stop doing any sort of measurements for amps and DACs.
    In general, yes, I agree. In practice the situation is different. My tower PC offers the best quality USB ports compared to three other laptops I have in the house. The Dell is the worst. The Levono is next. The Sony VAIO the best among the laptops, but still not close to my PC. I get the best and cleanest results using the PC for both signal generation and acquisition. For DACs and amps, even little things like WiFi usage or how my breakout cables are laid out out or how rusty my alligator clips are or which USB bus the devices are on matter more.

    Here's an example:

    Output from Lenovo (power not plugged in - you don't even want to see how it looks like when the laptop is charging)
    grace ibm.png

    Output from my PC:
    grace pc.png
     
    Last edited: Nov 5, 2015
    philipmorgan and ultrabike like this.
  11. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    50,977
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    On Jason's take, one thing I'd like to highlight: it's the man, not the machine. Just because someone has an AP, dScope, G.R.A.S., etc. doesn't mean they will be able to obtain quality measurements. I think most folks who have done a lot of measurements know this and sort of take the nuanced knowledge they have for granted.
     
    SKiring likes this.
  12. OJneg

    OJneg The Most Insufferable

    Friend BWC Executed
    Joined:
    Sep 25, 2015
    Likes Received:
    3,745
    Dislikes Received:
    36
    Trophy Points:
    113
    Location:
    Grand Rapids
    Never go full-Nwavguy
     
  13. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    50,977
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    Following up on the m9XX, here is measurement I took from the Sennheiser HD650 into the UMM-6 microphone fed into ARTA.

    I think I can say that I feel more comfortable that the QA400/QA190 or the breakout cables / ratshack resistors I was using as the load weren't causing anything weird with the m9XX. The tips of the high frequency spuriae seen in the direct amp measurements can also be seen in the acoustic measurements from the HD650 headphone.

    Again, nothing to go full-nwavguy on. I got the same result using the Lenovo.
    headphones.png
     
  14. Ormia

    Ormia Rando

    Joined:
    Nov 5, 2015
    Likes Received:
    0
    Dislikes Received:
    0
    Trophy Points:
    1
    Oh right. Totally forgot about the bit-depth resolution on my device.
     
  15. imackler

    imackler Key Lime Pie Infected Aberdeen Wings Spy

    Friend
    Joined:
    Oct 1, 2015
    Likes Received:
    1,449
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    Anaheim, CA
    I think he's gone into hiding until he can come back and sue everyone for copryight infringement whose tweaked his design.
     
  16. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,602
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    Meh. He proly is around low key now. He probably went through mid-life crisis. It happens.
     
  17. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    50,977
    Dislikes Received:
    85
    Trophy Points:
    113
    Location:
    Antarctica
    Or the other way around. Opamp for voltage gain. Already done since the beginning of time. Opamp for buffer. Same. What a lot of people don't grasp is that he didn't do anything new, except put the pot between the stages (and that wasn't new either), which from a practical end-user engineering point of view was 110% retarded. I must have received a dozen PMs on HF from people who didn't understand that their hard-wired low gain setting was still too high for their source output levels. What he did well was marketing himself.
     
    Last edited: Nov 5, 2015
  18. OJneg

    OJneg The Most Insufferable

    Friend BWC Executed
    Joined:
    Sep 25, 2015
    Likes Received:
    3,745
    Dislikes Received:
    36
    Trophy Points:
    113
    Location:
    Grand Rapids
    Fucking this.

    The objectivist shtick sells. O2 and ODAC wouldn't be staples on Head-Fi if it didn't.

    Let's nip this thread derailment in the bud :Violin:
     
  19. Cspirou

    Cspirou They call me Sparky

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    4,956
    Dislikes Received:
    5
    Trophy Points:
    113
    Location:
    Northwest France
    Aren't good measurments a sort of hedge when buying gear? In other words, while you can find a good amp that measures badly, you will never find a bad amp that measures well.

    If not, does someone have an example of an amp that measures well but is unusable?

    As far as equipment goes I am curious if the following would work for audio:

    http://www.sigzig.com

    http://www.bitscope.com
     
  20. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Friend MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,602
    Dislikes Received:
    16
    Trophy Points:
    113
    Location:
    Irvine CA
    The sigzig seem a little too slow. The bit scope seems a little constrained in bit depth. They are pretty kool though.
     

Share This Page