NwAvGuy's Dream Cums Tru: Schiit Magni 3 Heresy

Discussion in 'Headphone Amplifier Measurements' started by purr1n, Nov 22, 2019.

  1. 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
    [looks like you got this] what about one Topping AIO which had horribad output impedance, which Amir conveniently did not test and did not mention. Had this been a PS Audio or Schiit product, he would have had a field day.

    What about Amir excusing AC mains noise on preferred vendors, while calling them out on vendors he did not like.

    @julian67: Yes, I take this personally because people like you, who don't have the time or inclination to study complex measurements from TOTL insanely good instrumentation, end up bedazzled and befuddled. What ASR doing isn't science.
     
  2. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,653
    Trophy Points:
    113
    Location:
    On planet
    Typical US suburban residential spaces measure down to 30+ dBA when the area is quiet.
    10 dBA is exceeding hard to achieve. Bell Labs had such an anechoic chamber many years back. Microsoft and a few other now have anechoic chambers that achieve this level and possibly below.

    Here are some real acoustic measurements from my acoustic sound controlled space using an ACO 7046 mic capsule on a Josephson C617 body and system calibrated using Bruel & Kjaer 4231 mic calibrator. Scroll down to the last three screenshots:
    found here
    The overlay are Noise Criteria curves.
     
  3. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    @Lyander and @purr1n I was talking about a best-case scenario, like cherry-picking a minute at night where no cars or trains pass. I still kinda doubt it's that much louder in most suburban areas, although I admit that I live in a relatively secluded area. On the one side there's a big forest, so there's no house within 10km or so and on the other side there aren't too many, either. Obviously living in a city it's worse and I just assumed that many people on here live in the suburbs rather than in the city directly. Even then I guess 30dBA for a best-case scenario should be doable in most cities at night. Although 30dBA could be 50, maybe 60dB SPL considering rumble from diesel cars, trams, buses, maybe a river, etc, so maybe we're saying the same thing here.

    @atomicbob I don't really hear anything when none of my electronics are on, at least at night*. Do you hear stuff in your lab? Again, I'm not talking about the occasional passing car, train or plane, but continuous noise. What did your mic show when put in a box shielded from environmental noise? Are you sure you weren't hitting the noise limits? ACO specs it as 12dBA self noise.
    I was only in anechoic chambers twice, but in neither case I felt it was that much quieter than a quiet room at night. But I admit I was never alone on those occasions.

    Do you have a source for that? I kind of find it hard to believe and I'd find it shocking if the hearing deteriorates that much with age. Hearing thresholds in the 500Hz-4kHz region between -5db and -10dB are what I'd think of as "normal". In my own tests that seems more likely, at least. I doubt I'm that much of an outlier in that regard, I think most people without hearing damage should have pretty much the same hearing threshold.

    It would be interesting to measure the environmental noise in a home environment. Maybe a dynamic mic with a low noise pre?

    *At least when the heating is off. When it's on 30dbA seems realistic to me.
     
    Last edited: Nov 27, 2019
  4. MisterRogers

    MisterRogers Ethernet Nervosa

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    2,063
    Trophy Points:
    93
    Location:
    Aurora, CO
    Home Page:
    Oh hell yea it does - on the bottom and top end. Also strangely, I seem to have developed hearing loss at the exact frequencies my wife speaks ;)
     
  5. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,653
    Trophy Points:
    113
    Location:
    On planet
    My day job is Acoustic Research for a major medical device company. I work in sound controlled spaces, make NIST traceable calibrated acoustic measurements and have been in anechoic labs operated by colleagues around the area, several at Microsoft, several at Boeing Aerospace, NWAA labs, etc. It is my job to know the sound levels of environmental spaces from residential to urban, private, commercial, industrial, aircraft, etc.

    A dynamic microphone isn't going to cut it. You need an actual measurement microphone with exceptionally low self noise. I listed such a mic system in my post above. The 7046 mic capsule has one of the lowest self noise available for a measurement microphone with 900+ year stability and so far is proving after many trips to Scantek (calibration service.)
     
  6. 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'm sure you already know this, or maybe not, but 30db SPL is pretty fricking quiet (again, brain subtracts ambient noises and sounds very well - I suggest to go to a sound room with sound treatments, and then step out - for a while you will suddenly notice how much noise there is). I hope you understand that db scale is inherent logarithmic in terms of actual sound pressure, and just because it's 30db (and not close to zero) doesn't mean it's even a moderate sound level.

    Lets say your room actually is 10db. If you listen to music with 92db peaks (reasonable expectation given your OB woofer and Voxactiv distortion vs SPL), which is pretty darn loud, then we are still talking about 82db SINAD, which the vast majority of "shitty" measured gear at ASR still complies with.

    upload_2019-11-27_13-0-6.png

    Also, what point can you confidently say that you yourself can detection distortion? 0.1%, 0.01% 0.005%, 0.001%?
    FWIW, a lot of the TOTL Crown CDI amps commonly used in Atmos mix stages are more like 0.05% distortion (they don't use Amir approved gear), yet the mix engineers at ILM are somehow able to pull off some pretty awesome stuff in movies like Avengers Endgame.
     
    Last edited: Nov 27, 2019
  7. 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
    And what's the SINAD of a microphone / preamp / console (ensuring enough headroom to avoid clipping)? Pretty shitty.

    And the de-noiser plugin on Pro Tools doesn't count. It's cheating, not to mention destructive
     
  8. julian67

    julian67 Facebook Friend

    CBC
    Joined:
    Oct 5, 2015
    Likes Received:
    229
    Trophy Points:
    43
    Location:
    England (Proper England - Vilayet).
    No, I didn't do blind tests. I don't own the Pi DAC but I do own the Lepai. I just picked them quickly from the lists as stuff I remembered reading about at ASR. I guess I should/could have been more circumspect and only picked stuff I used but I was trying to make a general point. Anyway, in use the Lepai is not exactly borderline and hardly requires blind testing to distinguish it from a typical household name, boringly well designed and manufactured Japanese home Hi-Fi amp i.e. Yamaha, NAD etc. My Lepai served a purpose when I needed something tiny to use in a small room with a lot of unwanted noise (busy road junction right outside). It now resides in a box awaiting the day I either win a small lotto prize and buy a houseboat, or once again find myself living in the 24/7 rumble of the concrete jungle.

    I didn't post in order to offend anybody, only to make the point that measurements can be very useful even if we don't like the person making them. Your response? Meeow!
     
  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
    I really didn't want to start a discussion about chasing numbers, as far as I'm concerned I might be able to hear 2nd order distortion 100db down listening to a 440Hz sine with my HD800, but do I care? Nope, as long as distortion isn't 10% at reasonable listening levels like with the PAP woofers I don't care too much. I don't give a shit about the next objectivist amp where the analyzer has troubles measuring its distortion, I'm far too much of a subjectivist for that.

    I don't want to derail the thread further, but the measurement mic and noise level discussion intrigues me. I've been meaning to build a low noise mic system for a while, but then again it won't be cheap and I don't think it's that useful.
    I've been wondering about this. Obviously I'm a noob with zero experience, so just tell me straight if I'm wrong here, but the 7046 is just a 1/2" mic capsule. Wouldn't a larger mic have much lower self noise? I have 0 experience with those large diaphragm condenser mics, so I thought a dynamic mic with their large diaphragms could have similarly low self noise provided you amplify the signal a lot.
    What prevents the dynamic mic from having a low self noise? What if we take a dynamic headphone driver and use that as a microphone? They are still reasonably light and can have large diaphragms, so there's less thermal noise. If we amplify that signal (and assuming that we have gear with good enough SNR), why would it have more noise than a condenser mic? In my head I just thought the SNR of the preamps becomes the problem, not the dynamic mic itself.
    Basically my thinking goes like this: You have surface area vs thermal noise and you have moving mass vs efficiency. In order to get low self noise you want a large surface area with low moving mass.
     
  10. 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
    tl;dr Happy now?

    upload_2019-11-27_13-26-57.png
     
  11. julian67

    julian67 Facebook Friend

    CBC
    Joined:
    Oct 5, 2015
    Likes Received:
    229
    Trophy Points:
    43
    Location:
    England (Proper England - Vilayet).
  12. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    The few ASR reviews I've seen vary in quality. This is the first time I see the ones you mention. And they are poster childs of what not to do.

    Let me explain.

    LP-2020 ASR Review

    The only thing useful in that entire review are the THD+N vs Pout sweeps for 8 and 4 ohms, and this is what one may conclude from them:

    1. Channel 2 (right channel) is defective in the device he measured.
    2. Considering the only non-defective Channel (left channel), it is evident the LP-2020 out performs the SMSL SA100 in its linear operating region.
    3. The linear operating region seems indeed 8W for 8 ohms and 10W for 4 ohms.
    4. THD+N is about 65 dB for 4 and 8 ohms, or 0.05%.
    LP-2020 manual and specs:
    https://www.parts-express.com/pedoc...-mini-amplififer-with-power-supply-manual.pdf

    The manual specs the device at 20W which is too optimistic, but < 0.1% distortion is not.

    Given what I see I WOULD recommend this product considering price and with moderate power delivery in mind.

    In the mean time, you and most every other average reader will walk out of that review feeling that this is a POS. Conclusion? The review is fairly misleading.

    EDIT: Amir should learn how to make THD+N vs Freq sweeps. The THD+N vs Pout for different frequencies was specially painful and meaningless to me.

    PiFi DAC+ Pi Sound Card Review

    It comes down to a comparison against the HiFiBerry DAC+ Pro. The Pro has 6 dB better THD+N performance.

    The PiFi has less 2nd harmonic distortion, but the Pro has much less 3rd harmonic distortion. In fact, odd harmonics are higher on the PiFi all around. Maybe something is clipping. Who knows. It would have been great if he had tried testing both devices at -3 or -6 dBFS.

    Conclusion about the review? Is a bit incomplete and it gives the impression that the PiFi is shit. It is a competitive market, but 0.005% THD+N at 0 dBFS is not shit IMO. And we leave the review with an uneasy feeling since the number might have been too pesimistic if 0 dBFS was used for THD+N.

    EDIT: Note that even REW recommends -3 dBFS for THD+N measurements (with DACs in mind). It's written all over the generator tool.
     
    Last edited: Nov 29, 2019
  13. julian67

    julian67 Facebook Friend

    CBC
    Joined:
    Oct 5, 2015
    Likes Received:
    229
    Trophy Points:
    43
    Location:
    England (Proper England - Vilayet).
    Nope. I owned and used one for years before that review and it's pretty fair.
    As well as the measurements the review offers subjective impression and opinion i.e.

    "Conclusions
    There is nothing horrifically broken in Lepy LP-2020A. It is hard to imagine doing much better given the price that it aimed for. I understand there is another version with the proper TI Tripath amplifier. If so, that may be a better buy as I have heard it produces more power which was the main issue here."

    Power being the main issue is on the nail. The big value proposition is accurately and fairly described "...hard to imagine doing much better given the price...". My original point was that there was a lot of hype and essentially uncritical praise for this product but the measurements offer some insight into the reality. It's an OK piece of kit within very tight constraints of budget and environment and other hardware, but otherwise look elsewhere.

    Anyway, I'll just apologise now for the outrageous heresy of not hating the bad person who does only bad things . I don't want any more catty claw marks on my beautiful fizog.
     
  14. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,653
    Trophy Points:
    113
    Location:
    On planet
    @Serious : dynamic microphones can have low self-noise but the ones close to useful require 60 to70 dB of amplification. Further, their frequency response flatness do not approach that of condenser measurement microphones. Too variable for measurement work. If it was that easy you would find dynamic measurement microphones on the market.

    Yes, the 7046 has 12 dBA of self noise. Here is a partial list of the usual suspects:

    ACO Pacific
    Bruel & Kjaer
    DPA
    Earthworks
    GRAS
    Josephson
    Microtech Gefell
    Neumann

    Your assignment is to go find the residual noise levels of the various offerings and learn why 1/2 inch is the most commonly used measurement microphone size. Hint: all electroacoustic transducers become directional at high frequencies and all have omni directional patterns at low frequencies. Where the transition occurs is tied to physics, the size of the sensor.

    And finally, here is a useful chart for everyone as a typical rule-of-thumb for environmental noise:

    environmental noise - typical.png
     
  15. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    This is classic Amir as well. Many of his reviews are ambivalent.

    Amir does not always do bad things. But he does many things wrong. Those two ambivalent reviews you mentioned are examples. Specially the LP-2020A review.

    How to make it better?
    1. Do proper THD+N vs Freq sweeps.
    2. Lower the power to a more representative value of the linear operating region of the device when doing THD+N tests in general for a give load.
    3. Recognize when the device is likely faulty before publishing a set of measurements, and act accordingly before reaching final conclusions.
     
  16. 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 know, I meant specifically making a mic for low noise recordings, not a measurement mic. Bandwidth doesn't really matter in my use case and neither does the frequency response flatness. 50Hz-5kHz -3db bandwidth with a peak at 2kHz? Whatever...
    Maybe some time I will build such a rig to make recordings with a better-than-human-hearing self noise. I guess I could try a dynamic mic with a lot of gain like you mentioned. Could come close to those kilobuck measurement systems in self-noise.
    For the most part I really just want to see how quiet my room really is at night. So far I've always hit the noise floor of my mics (at 25-35dbA). Maybe low level driver distortion results could be telling aswell, but I think other people have tried that in the past and found nothing useful.

    BTW: Quiet rural nighttime means outdoors, right? Animals, transportation in the distance, etc could really mess up the noise. Indoors that should be quieter than a recording studio IMO. Plus electronics (in the recording studio) usually make some noise when turned on.
     
  17. Thad E Ginathom

    Thad E Ginathom Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    14,134
    Trophy Points:
    113
    Location:
    India
    Before you maybe go for refinements with equipment, just start an app on your phone.

    Here, in sub-urban Chennai, Southern India, 4am and we we have just had a wet monsoon night. It's cool enough to have the door open and I'm seeing 42db. Mostly frogs! Normal day time, with ceiling fan and or AC, we'd be over 50.

    Worryingly, when I type, it see the meter go over 70! Well, my wife knows how loud I type, and tells me, but I don't. That's the hearing damage.

    Ha! 61 dB of rain just hit. I should go to bed.
     
  18. Thad E Ginathom

    Thad E Ginathom Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    14,134
    Trophy Points:
    113
    Location:
    India
    PS... Honestly, for non-legal, personal use I think phone apps are fine for environmental noise. The frequency range is going to be limited by the mic. I even bought a plug-in mic once, and have an app that can use its supplied calibration measurements (no, it wasn't expensive) but I never really used it. Mostly we are interested in nothing more accurate than the rough table posted above, or in comparative measurements.
     
  19. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    At 42db you might be hitting the self noise of the mics in your phone. Hard to say if it's not calibrated. The iPad goes to maybe 35db as do most cheap SPL meters. Whatever app and mic combination I try that I have here, all of them are limited by noise. My Shure SM58 into a midrange all-in-one-external soundcard was the best I had, but the soundcard broke. Make a recording and see if you can hear anything other than noise. Human hearing has far lower self noise than most microphones.
     
  20. JustAnotherRando

    JustAnotherRando My other bike is a Ferrari

    Contributor
    Joined:
    Aug 15, 2017
    Likes Received:
    1,373
    Trophy Points:
    93
    Location:
    Hong Kong
    By total coincidence and just for curiosity, I fired up a sound meter app on two phones two nights ago and lay them next to a cheapo sound meter (I happened to be in the process of switching from an old to a new phone and was reinstalling apps).

    The sound meter was reading ~40dB with living room ambient. The old phone was 50dB and the new phone ~30dB (no way was the room in the 30s!) I was wondering if this difference would remain linear, so I started randomly generating noise (nothing properly repeatable, just rubbing my hands together then clapping) until the meter reached the 70s. The old phone consistently read 10-15dB over the meter, the new phone went into the same ballpark as the meter. My conclusion was that phones are really random.

    Oh, and as the conversation veered into urban sound levels- in my living room at night, I get ~34dBA. This is a high floor apartment in an unusually quiet part of Hong Kong (zero street noise), anything nearby with fans in it turned off.

    [Edited to make the presentation of numbers a bit clearer.]
     
    Last edited: Nov 27, 2019

Share This Page