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

Estimating channel imbalance of headphone cable

Discussion in 'Sound Science' started by bobbooo, Sep 9, 2019.
1 2
  1. castleofargh Contributor
    those are different contexts. with abx we have a clear reference to get back to. when we purchase a headphone with 1dB imbalance(usually not everywhere in the frequency range, but also usually reaching more than 1dB imbalance somewhere), we have no reference. we just listen to that headphone and unless the difference clearly generates the feeling that something is wrong, we're likely to just enjoy our headphone like anybody else. and it's a good thing, because otherwise we'd have to throw most headphones and IEMs ever manufactured to the garbage. ^_^
    I also feel 0.2dB clearly in direct comparison(in the midrange!), but just as well I will often forget my EQ on the wrong position or set for the wrong headphone and spend sometimes a full day before noticing. just enjoying my music as usual.
    when discussing perception here, we have to be clear about the testing conditions because the result can massively depend on them.
  2. bobbooo
    As far as I'm aware ABX testing is really the only kind of scientifically valid test that eliminates pretty much all biases. Of course it's not perfect, but it's the best we have. Unless you know of another kind of test that eliminates all bias but is more representative of real-world usage? The fact that most headphones have channel imbalances is even more reason to have a cable with minimal imbalance of its own, otherwise the imbalances from both cable and headphones could cumulatively reach audibility (potentially 2dB if each has an imbalance of 1dB).
    Last edited: Sep 15, 2019
  3. bigshot
    I didn't actually doubt that you could detect a difference. The purpose of the dB scale is to have each unit be about a JDD. My point was that when you do tests like this, you can better put them in perspective. Would a 1dB difference have made an impact on your listening enjoyment if you weren't directly comparing and you were just listening to music in your living room? Would it have made a difference if it was just a 1dB imbalance between 200Hz and 400Hz? Would 1dB be different if it was between 2kHz and 4kHz? Hopefully you got a sense of just how small 1dB is. Now you should test to find out the context.

    Noticeable differences in a direct A/B switched comparison are the absolute edge of what matters. In many cases, the JDD is complete overkill. If you are working in a studio and you need your work to be consistent from day to day, you might need those sorts of standards. If you are listening to Beethoven in your living room, it won't make your sound any better. There are things that do matter. Transducers, room acoustics, recording quality... those are the things that will make the biggest improvement in your sound system. Doubling down on smaller and smaller numbers is easy when you don't have any idea of what those numbers represent. Worrying about wires isn't very productive. Just get the right wire for the job and be done with it. You don't need to split atoms on wires.

    Hopefully, now you'll have a feel for what a dB sounds like. Now go do the same sort of tests to determine what a noise floor of -40dB sounds like, or a 5dB imbalance in a various octaves in your frequency response. Do a range of rips at various compression levels and learn what the differences are between codecs and data rates are. Then you'll know what the numbers represent, and you'll realize that for the purposes of listening to commercially recorded music in the home, most of the stuff audiophiles worry about are largely irrelevant to their listening enjoyment. Numbers aren't abstract. They have context. But 1dB in one context might be completely different than 1dB in another. You learn about context by experimenting and careful listening in situations that resemble your target conditions (i.e. sitting on the couch listening to Miles Davis).

    The AES videos linked in my sig are a good start. He has downloadable examples all set up so you can hear it for yourself on your own system. Science is about practical application too.
    Last edited: Sep 15, 2019
  4. bobbooo
    I'll take a look at those videos when I get a chance. Just to be clear though, I have read up a fair amount about audibility thresholds of various audio parameters already, and understand there are areas that have a much bigger effect than channel imbalance. My motivations for asking the original question were partly just curiosity about the physics and psychoacoustics of the audio playback chain, and a bit of perfectionism. When you've gone through your audio chain and proved to yourself through reading resources and doing tests that most of the important things (e.g. obviously headphone frequency response, recordings with the best mastering and dynamic range, bitrates, encoders etc.) are good enough and above the audibility threshold, you naturally turn to the less important things like channel imbalance to make sure they are too. It's the same with any hobby, you start off getting the important things right, then move on to the nuances. But if we have a lassaiz-faire attitude to the whole playback chain, small JND errors could cumulatively add up to noticeably audible ones, so for me, I prefer to err on the side of caution and leave a bit of headroom, which usually costs very little if any more time or money.
    Last edited: Sep 15, 2019
    castleofargh likes this.
  5. bigshot
    Hearing is better than reading! I'm satisfied with my specs, but I'm not turning to fussing with inaudible or irrelevant specs, I'm working with signal processing that can actually improve the sound of music.
    Last edited: Sep 16, 2019
1 2

Share This Page