1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

Jitter Correlation to Audibility

Discussion in 'Sound Science' started by robertsong, Jun 19, 2013.
13 14 15 16 17 18 19 20 21 22
24 25
  1. blades
    We were never able to hear an audible difference due to jitter in any hi fi hear we tested.  Our conclusion is that the jitter inherent in modern audio products is simply not audible.
  2. SharpEars
  3. bigshot
    There are lots of audio equipment manufacturers who claim jitter is audible (but only in their competitors' products). The people who say it is inaudible are the ones doing carefully controlled tests and aren't trying to sell you something.
  4. Roly1650
    But if you read the entire article, he chickens out anyway, 20 ps suddenly becomes "a few hundred pico seconds", which agrees with pretty much everyone else. Nothing new here.
  5. bigshot
    When I first dove down the jitter rabbit hole, I tried to wrap my head around the slivers of time being talked about. I wanted to figure out what a picosecond was. I discovered it was so infinitesimally small, it was like trying to count all the stars in the sky at night. Every other threshold I found for timing audibility was an order of magnitude or more above the figures I was seeing cited for jitter.
    For instance, in acoustics the threshold of audibility for reverberation is defined as the length of time that the reflected sound takes to dwindle down to -60dB. Take a wild guess what amount of time is generally accepted as the threshold of human perception for reverberation decay.
  6. stv014
    More accurately, it references a paper that claims that 20 ps jitter is theoretically audible (by comparison against the absolute threshold of hearing) if all of the following conditions are met:
    - the source signal is an extremely loud pure ultrasonic sine wave
    - the audio band is perfectly silent (including no ambient noise), other than the side-band resulting from the jitter
    - sinusoid jitter is applied at a frequency that is carefully cherry picked such that a side-band is created exactly at the frequency (~3-3.5 kHz) where human hearing is the most sensitive, e.g. a 22 kHz tone is modulated by 18.5-19 kHz jitter
    This is quite different both from the spectrum of real music, and from typical jitter in most devices, and also ignores the possibility of masking.
  7. bfreedma
    Quoting articles from a company that would like to sell you a very expensive solution to a problem their own article runs away from toward the end may not be the best way to go here.
    Not exactly an independent source of information.....
  8. ab initio
    This one is pretty simple. Listen to all music at 44.1kHz. Done.
  9. blades
  10. ab initio
    7000 picoseconds sure sounds like a lot of picoseconds....
    ....until you realize it's only 7 nanoseconds...
    ...which is less than 100 times smaller than the period of a 1 MHz sinusoid....
    ... a frequency which is 50 times higher that the typical highest sound a human can hear.
    It would take such a remarkable set of specific circumstances such that a timing error that small could could add up in just the right way so that it can just be discerned by the most sensitive human ear in that I don't see a point in fussing over it.

  11. cjl
    Even claims that a few hundred picoseconds is audible should be looked at pretty skeptically. That's still a tiny amount of jitter.
  12. nick_charles Contributor
    As long ago as 1998 the Dolby labs researchers were unable to find any commercial products with jitter significant enough to be audible. By a wonderful irony the 2 or 3 commercial products that may just conceivably have jitter bad enough to be audible are all vastly expensive such as this...http://www.stereophile.com/content/mcintosh-ms750-music-server-measurements
    but even this pos has distortion sidebands almost 90db down 
  13. cjl

    And that's with 14 nanoseconds (14,000 picoseconds) of jitter, which is absolutely appalling in any piece of modern electronic equipment.
  14. esldude
    Reminds me of the strangest jitter I have seen on a device.  I have the Pioneer Universal player also in that article.  It has horrendous jitter, worse than the McIntosh at the beginning of a track, but then slowly the jitter goes down until by 12 seconds it is very low (below what my test setup would show).  I don't think JA knew this.  He uses a 33 second file for the J-test.  So the number he reported is pretty bad, but an average of 12 seconds of terrible and the rest pretty good.  This goofy player does this each time you swap tracks.  Go from track two to track three and jitter is horrendous, but slowly decreases over the first 12 seconds of a track.  Made me wonder if perhaps the McIntosh were doing something similar.  If you don't look at the jitter over time you wouldn't know this.
    I don't know how common that is in a disc player, but I haven't seen it in the other few I have tested. 
  15. SharpEars
    This pos has sidebands 88 dB down, because the signal was at -70 dB. If the signal was at 0 dB, would the most noticeable sidebands be at -18 dB and the rest of them at -30 dB? If so, these would be clearly audible!
    Update: Please disregard. I went and read the original article and the signal was actually at -6 dBFS, that graph is very misleading since it starts at -70 dBFS (i.e., lying with statistics).
13 14 15 16 17 18 19 20 21 22
24 25

Share This Page