Ultrasound audibility

Discussion in 'Audio Science' started by Serious, Jul 7, 2019.

?

Is Serious losing his mind?

  1. It sure seems like it

    10 vote(s)
    23.3%
  2. No, more research needs to be done on the limits of human perception

    27 vote(s)
    62.8%
  3. Maybe

    12 vote(s)
    27.9%
Multiple votes are allowed.
  1. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    TL;DR: I can consistently detect the difference between a tone in the 30-60kHz range being played and digital silence in a blind test.

    As mentioned in my profile post here I did a blind test between a 30kHz sinewave and digital silence, could hear a difference and got a 10/10 in the blind test. Prior to that I didn't think that I should be able to hear a difference, but I did.
    I just repeated the experiment and got perfect 10/10 scores all the way up to 60kHz. I also measured the output of both my DAC and the speaker drivers to see if there were any subharmonics (or other types of distortion) that I could measure that I could've heard, but I couldn't find any. Again I tried to match the level to around 80db at 20kHz. According to my own measurements I should have decent output to about 60kHz.
    In general I mainly felt a bit of a pressure on my head with the tones, a bit like a headache and a bit like when the acoustic reflex sets in (which it certainly does at those levels for sounds in the audible range). I wouldn't say I heard a high pitched tone, but I felt the noise floor increasing and maybe felt a bit uneasy in general. Detecting the tones got more and more difficult past 40kHz and 60kHz wasn't easy. Not as difficult as the 24 vs 20 bit files, more on the level of 512kbit/s .ogg vorbis and lossless redbook. The 70kHz tone was odd in that I'm sure I heard a very quiet difference tone, but to the best of my ability I couldn't measure one. It sounded like a tone in the 2kHz range. I could not hear a difference between silence and the 80, 85 and 90kHz tones.

    I know it's no proof, but I took a screenshot of a 10/10 score I got in one of the blind tests at 50kHz. I'm not making this up and hearing the difference here was a lot easier than the difference between 20bit and 24bit files. I later added 70, 80 and 85kHz files.
    [​IMG]
     
  2. bixby

    bixby Friend

    Pyrate
    Joined:
    Sep 26, 2015
    Likes Received:
    3,987
    Trophy Points:
    113
    Location:
    Northern Colorado
    I cannot hear much above 13.5 khz reliably these days, but have never tried anything above 20khz.

    I knew a now old departed friend that ran an audio shop in AZ when I lived there. He was gaga over Tannoy's new super tweeter at the time. Said it made a big difference in overall perception of the speaker system to which it was added. I never heard it in a system, but this guy was pretty straight shooter. Perhaps some of those above 20khz can be perceived, maybe not heard, but felt.
     
  3. Lyander

    Lyander Official SBAF Equitable Empathizer

    Pyrate Contributor
    Joined:
    Feb 25, 2017
    Likes Received:
    10,961
    Trophy Points:
    113
    Location:
    Philippines, The
    First choice vote just by way of poking fun, but kidding aside there are enough ways to worry about gear even without people worrying about ultrasound.

    Waves are waves (this is ridiculous oversimplification but I'm not quite able to put the thought into words), I'm inclined to think that they should certainly be felt given ideal conditions— that said I doubt this will make any significant difference for most music unless it's ridiculously sparse.

    ... I'm thinking of Vrioon; Garns recommended the album some time ago in a random thread and I gave it a listen on a lark, the intro track had me yelp on first listen. For most other music I'm thinking it might not be particularly easy to discern changes in feeling just by virtue of so much else going on lower down the spectrum.

    Other applications, though, might be very interesting. I've been tempted to do a study on how environmental infrasound (or more broadly noise pollution in general, but infrasound looks cooler) might be in part responsible for the prevalence of certain mental conditions in urban settings, but the sheer PITA of designing the research and the ungodly amount of stuff on the list for gathering and cleansing data (not to mention it seems to be edging well into pseudoscience) put me off. I know that the guy behind Akira specifically included ultrasonic content in BluRay release as a manner of influencing audience feeling, never bothered checking how much BS that was since I was skeptical from the get-go.

    That'd be fun, have a bunch of people watch an old anime film in a controlled setting and invest a load of time determining whether or not stuff in the 40kHz region made scenes feel more impactful or what.
     
  4. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,653
    Trophy Points:
    113
    Location:
    On planet
    What was your listening chain?
    What sample rate used for the playback chain?
    When you say you measured the output for subharmonics what measurement system was used? Very few microphones can actually measure reliably above 30 KHz. Have a look at Earthworks microphones or some specialty ACO Pacific or Bruel & Kjaer.
    My bet is that you are actually detecting an inter-modulation distortion rather than the 50KHz tone directly.

    I am a proponent of 60 KHz BW to accurately reproduce the leading edges of percussive sounds. In the process I've run into many system / measurement limitations. This is hard to do accurately.
     
  5. Syzygy

    Syzygy Friend

    Pyrate
    Joined:
    Jun 13, 2018
    Likes Received:
    2,144
    Trophy Points:
    93
    Location:
    DFW, Texas
    Yeah, I know the feeling you're describing, as I used to be able to hear low ultrasonics in my youth as well, but I'd never tested with such high frequencies. In fact, I never tested at all, since it was mostly pain-inducing, but dependent upon the amplitude. But I could definitely hear them…active ultrasonic alarms, some dog whistles. Higher frequencies were more "sensed" than heard.

    As I stated in your profile post, unless you've got a specialised chain designed to reliably reproduce those frequencies, I'd bet you're hearing a subharmonic, which itself could still be "normally" ultrasonic to humans, say near 20kHz. It could also be the amplified vibration of a filter component.

    When I heard the ultras, I could tell frequency differences; that is, compared to similar signals. Do you sense differences at the different selected frequencies, or are they essentially the same sensations at 30k and 50k? If they're the same, I'd wager some component is vibrating resonant with those high frequencies and it's getting amplified.

    Another test to do would be to see if you can borrow matching components (amp-for-amp, etc.) and see if it goes away with the borrowed one. That'd isolate it to your particular components.
     
  6. 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:
    My money is on IMD in the audible spectrum.
     
  7. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    What microphone with the speakers? Post RTA/FFT with the 30kHz tone playing using 48kHz bandwidth.
     
  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
    Gungnir MB into Ragnarok into my DIY OBs (meaning Voxativ drivers). 192kHz sample rate, both through Wasapi event and Directsound.
    I used my UMIK-1 microphone set up about 15cm from the driver while I was sitting about 5m from the speakers to give me some headroom for the SNR. Getting significantly closer than about 15cm requires me to use a different mic stand or I'd have to hold the mic myself. For the FFT I used REW set to 1M length. But either way both just showed the mic noise floor which seems to be higher than 30dbA, while the distance only gives me about a 15db headroom, so I cannot say that I didn't hear IMD.
    I don't have a tweeter that goes much higher than 30kHz so the Voxativs are actually the drivers with the best HF extension that I have. According to my own measurements using a WM61A there should be output up to 60kHz before it starts rolling off at a much steeper slope. I can't say how linear it is up to 60kHz though since I don't have a proper mic cal.

    I also fed the DAC output straight into my ADC running at 192kHz (Terratec DMX 6 Fire USB).

    I can try again with some pink or white noise overlaid on the tones at say - 60db to try to mask IMD. My room was pretty quiet with the loudest sound being my LCDs power supply (BenQ XL2420T) and my neighbors pump for the garden pond - at least until the rain started.

    @Syzygy I think they sounded different to me, but they also got quieter, so I can't say for sure. 70kHz was definitely the odd one out and I'm sure there was some form of IMD. The output sounded like a tone around 2kHz that was amplitude modulated by a very low frequency tone, around 0.5Hz or so. Getting closer to the driver made it louder and it went away when I got far off-axis. Even then I couldn't see it with my UMIK-1.

    I'll post the FFTs when I get back home this evening. I might redo them though since I just got noise.
     
  9. james444

    james444 Mad IEM modding wizard level 99

    Pyrate Flathead IEMW
    Joined:
    Jun 3, 2016
    Likes Received:
    2,100
    Trophy Points:
    93
    Location:
    Vienna, Austria
    I'm definitely in the camp of "more research needs to be done".

    There are indeed a few rare studies that point to possible ultrasonic perception in humans:
    From what I've been gathering, these are quite controversial and research is far from conclusive though.

    However, it may not be long until clarification. A new imaging tool based on optical coherence tomography (OCT) enables us to "watch the ear hear":

    https://www.sciencedaily.com/releases/2018/10/181017111011.htm

     
  10. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    I did not know the 8" Voxative AC-1.6 could go well into 30 kHz, let alone as high as 60 kHz. That is impressive.

    I do not even think many dedicated 1" tweeters can do well in those ranges. My understanding is that metal domes break up horribly around 20 kHz and even lower. Silks usually just roll off. Usually the more expensive beryllium tweeters can cover highs and ultrasonics more smoothly, but even then 60 kHz is quite high.

    Again, did not know the 8" German (paper?) Voxative could be used at such ultrasonic ranges. Amazing.
     
    Last edited: Jul 7, 2019
  11. Armaegis

    Armaegis Friend

    Pyrate BWC
    Joined:
    Sep 27, 2015
    Likes Received:
    7,466
    Trophy Points:
    113
    Location:
    Winnipeg
    None of your mics go that high either. Calibration won't matter if they can't detect anything to begin with.

    Most amps also deliberately roll off past 20kHz (something something stability oscillations voodoo). I imagine dac output stages do as well. I can't speak for @schiit here but I'd venture a half box of doughnuts that they do.
     
  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
    @ultrabike LOL, I know this sounds crazy, but there are third party measurements with calibrated Brüel & Kjær equipment that show decent output to 40kHz (where the graphs were cut off) and mine are mostly consistent with those, except that it keeps going to around 60kHz before dropping off. Let's please not make this about the drivers. I didn't say it was linear, but it's not a brickwall filter either. As I said in my profile post I might buy a small plasma tweeter off of eBay for experiments like those. But those probably have way higher IMD.

    @Armaegis I think 1/8th inch mics should go higher than 20kHz pretty easily. At least I can detect tones higher than 20kHz with my WM61A. Schiit specs the Ragnarok as 180kHz -3db and the Gungnir MB analog output to something like -1db at 200kHz. Using RCAs back into the ADC I can see the tones, so there's definitely output up to 90kHz+.

    I'll see how low my noise floor is tonight by measuring tones in the 10-20kHz range at much lower levels than the 30-60kHz tones I played. Those were at a digital level of 0.8, so around -2dbfs.
     
  13. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    In lieu of microphones, try FFT with the ultrasonic signals and without. Could me IMD, or could be other junk. That 8" driver isn't meant to be cleanly reproducing 30kHz+, or 10kHz+ plus for that matter, but it's going to try. Laser interferometer might be nice here to see what actually happens.
     
  14. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    Yes, that's what I did, but again the noise floor was too high with the UMIK-1. I'll also try my other mics (I only have a bunch of the WM61As, a Shure SM58 and one of the cheap electret NW-800s). I don't have the screenshots of the FFTs here on my laptop, they're only at home on my desktop.
     
  15. Thad E Ginathom

    Thad E Ginathom Friend

    Pyrate
    Joined:
    Sep 27, 2015
    Likes Received:
    14,134
    Trophy Points:
    113
    Location:
    India
    "ssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss"

    Could anyone hear that?

    Maybe.

    Was it worth it?

    Probably not.
     
  16. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    4,276
    Trophy Points:
    113
    My money's on IMD as well.

    Generally ultrasound isn't that trivial to play back and capture. And when designing signal chains it often makes sense to keep your analog paths ultrasound free, as it can cause problems (unless you're specifically designing for that bandwidth).

    I have read some research on the audibility of ultrasonic content and it was always the modulation of audible range content that was audible.
     
  17. mitochondrium

    mitochondrium Friend

    Pyrate
    Joined:
    Sep 1, 2017
    Likes Received:
    1,115
    Trophy Points:
    93
    Location:
    A Cell
    @Serious

    I think the weak points are the driver you use which may reach 30 kHz but I doubt it is very efficient up there. Other weak point is your microphone.
    It is akin to someone using candle light and a normal microscope in order to look at something the size of a molecule.

    Whether you hear/perceive 30 kHz is not for me to judge.

    I know that in order for me to hear 30 kHz I needed bats in the belfry.
     
  18. atomicbob

    atomicbob dScope Yoda

    Pyrate BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    18,653
    Trophy Points:
    113
    Location:
    On planet
    Here is an example simplified chain I use for recording percussive sounds above 20 KHz:
    20190708 QTC-1 P-Solo ADI-2 Pro 40 KHz recording system.png

    This result leads me to believe leading edge high frequency content is one characteristic I feel in the studio that is missing from playback systems:
    21 20181217 Pro Mark 5B hit to bell - Bosphorus 20in QTC-1 P-Solo ADI-2 Pro 384x32x1.png
    Note significant energy present above 20 KHz on the leading edge.

    *edit*
    I don't believe most of us are able to hear tones above 20 KHz but rather can perceive transient wave fronts having rate of change consistent with high frequency response. Square wave response is helpful in assessing system performance on this particular characteristic.

    The clip shown is a Pro Mark 5B stick hitting the bell of that Bosphorus 20 inch traditional cymbal.
     
    Last edited: Jul 8, 2019
  19. Walderstorn

    Walderstorn Friend

    Pyrate
    Joined:
    Mar 20, 2016
    Likes Received:
    1,904
    Trophy Points:
    113
    Location:
    Europe
    I wish that I could put you on my dynamic mri. See what's going on in that head of yours. Play several frequencIes and see if your brain would react to the ones you claim to hear and how it would do it.

    That actually wouldn't be dumb at all too for a new masters degree, if I was in a country where audio was more of a thing so I could find guinea p...i mean collaborators.

    Edit - Thinking more about the subject, it would require MRI heaphones with very good ANC as well but still able to play all the desired frequencies.
     
    Last edited: Jul 8, 2019
  20. ultrabike

    ultrabike Measurbator - Admin

    Staff Member Pyrate MZR
    Joined:
    Sep 25, 2015
    Likes Received:
    8,960
    Trophy Points:
    113
    Location:
    Irvine CA
    LOL! If I was being punched in the face at 30 kHz I don't know if I would hear it (aside from the lower frequency distortion that the cracking of my teeth would make). But I probably would feel it, and it might have an effect in my brain waves.
     

Share This Page