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

Testing audiophile claims and myths

Discussion in 'Sound Science' started by prog rock man, May 3, 2010.
762 763 764 765 766 767 768 769 770 771
773 774 775 776 777 778 779 780 781 782
  1. KeithEmo
    Exactly... and BOTH serve a valuable purpose.

    I would assume that a recording engineer would be more interested in making an album "sound the way it should" than in "accurately reproducing the waveform of each instrument at each point in the room". But I would expect a scientist to be more interested with the second option.

  2. KeithEmo
    I think that, if we want to agree on #3, we also need to include direction/position.
    (Or we need to specify that we're talking about measuring it at the eardrum.)

    If we're measuring at the eardrum, then it is entirely true...
    Our eardrum is essentially "one dimensional" - so all it detects is pressure.

    However, if we're measuring somewhere else in the signal chain, then that is far from true.
    You can record the pressure signature of a kick drum exactly.
    However, it will sound far different if it is reproduced behind you than if it is reproduced in front of you.
    (The same "time vs amplitude" signal at the speaker is producing a different result at your eardrum.)

  3. Phronesis
    I was referring to only an electronic signal, which can be fully characterized as a function of amplitude as a function of time.

    For an acoustic signal, it seems that you would need to characterize that for every point in space, and with ears, for every point on the surface of each eardrum, so we're really talking about an infinite number of signals, which we can neither fully record nor portray. Once we get out of circuits and wires into the air, I think there's way more complexity, uncertainty, and limitations of measurements.

    And once we leave the air and enter the ears and brain, yet another order of increase in complexity, uncertainty, and limitations of measurements.
  4. KeithEmo

    I have to admit that I don't listen to headphones that often.
    I have no doubt that the Stax Lambdas are more accurate.
    I do, however, find the Koss ESP/950s to sound quite pleasant, and to be quite revealing of detail.
    (Modern Stax models are also quite pricey here in the US.)

    The Koss ESP/950's are somewhat of an odd product.
    They are relatively easy to get here in the US.
    While they list for US$1000, they often go on sale for as little as US$650 around holiday season.
    And, while they have a very cheap plastic feel to them, they are very comfortable, and seem to hold up OK....
    (They also come with a lifetime warranty.)

    Here in the US, Koss is also very friendly on the cost of replacement parts.
    For example, Stax ear pads go for $75+.
    The ear pads for the ESP/950 are very flimsy, and tear easily, but they cost $5 a pair.
    And, the last time I looked, an official Koss extension cable for the ESP/950 was US$25.

    Also, notably, the Koss ESP/950 comes with an actual amplifier to drive them (included in the $$$).
    Again, it feels like cheap plastic, but it actually works pretty well....

  5. KeithEmo
    In that case I absolutely agree.
    With an electronic audio signal you have a voltage that varies over time.
    There is nothing else.
    All you can really vary is the degree of accuracy with which you measure and record it.
    (And far too many people seem to have this idea that there is the possibility of some "magical something else" involved - which obviously isn't the case.)

    Phronesis likes this.
  6. bigshot
    Since there's nothing for me to reply to, I have a question about HDMI... A friend of mine got an upscaler for his old video game systems and it outputs to HDMI. For some reason, it only outputs video OR audio, not both at once. It had an analogue headphone output, so he plugged a bluetooth transmitter into that, but it's clunky. Is it common for HDMI to be used for audio only? Is there a reason why it wouldn't include both audio and video? Is there some sort of crazy "signal purity" theory involved here?
    Last edited: Dec 14, 2018
  7. bfreedma

    It’s not common for HDMI to be restricted to audio only, though I have seen it. Can you post the model of the upscaler and the game system - might be a configuration issue or a limitation with the specific converter. What are you/he doing that switches between audio and video? Is it a change in settings on the game system or the converter?

    One thing that causes audio to not be output is a headphone jack that is “stuck” in the connected state. First thing I’d try is using a can of compressed air on the headphone jack to see if anything is making contact. Happens more often than you would think, particularly on gear that hasn’t been used in a while. Might not be the case here since you mention that audio is possible, but no harm in trying it.
  8. bigshot
    I’ll get more info from him and let you know.
  9. castleofargh Contributor
    same answers as Gregorio.
    and about this section and how to discuss, well it's your forum along with everybody else in it. I only shared my opinion and if you guys want to discuss a certain way, just do so. if enough people align on that, it will become what this section is about ^_^.
    Phronesis likes this.
  10. gregorio
    We keep seeing this sort of thing here, it's BOTH a misrepresentation of the actual science itself AND the inappropriate application of it to the listening of music. They are ABSOLUTELY NOT "saying" that "if you delay the signal the signal reaching one ear by as little as 10 microseconds we can notice a perceptible shift in the apparent location of the source"!! They are saying that it's possible to manufacture an extremely specific SET of conditions, under which it can be possible for humans to detect different arrival times as little as 10 microsecs. In this case, the set of required conditions are: Only two pure 1kHz sine waves, of identical duration and amplitude, of a specifically short duration (around half a sec), with a specific time interval between the 1kHz pulses and in isolation, with no other sounds present (such as other tones or reflections/acoustics). Change just ONE of those conditions and the time difference detectable increases dramatically, even "exponentially"! For example, the conclusion of the papers you quoted was that above 1.5kHz, ITD in human hearing doesn't operate at all, which ironically contradicts your previous assertions about the importance of ultrasonic freqs. So, the science itself has been misrepresented and then the application of it to the listening of music is also completely fallacious because NOT EVEN A SINGLE ONE of those required conditions exist with music!

    So often, evidence such as this is hauled out as supporting evidence of what audiophiles think they can hear or believe they should/could be able to hear but that presentation of the science is FALSE, either deliberately (to fulfil some agenda), out of ignorance or both. It's FALSE because it's only half the story (and sometimes less than half). Taking the example of the scientific papers relating to high/ultrasonic freq hearing response: They not only effectively present evidence of what human hearing is capable of but of what it isn't. Yes, some young adults can hear individual ultrasonic freqs but ONLY if they exceed high SPLs. The level of ultrasonic content in music does NOT exceed those required SPLs and is actually several/many times below the required threshold. The science is therefore actually providing evidence that supports the argument AGAINST ultrasonic content in music being audible, which is the EXACT OPPOSITE of how it's often presented by audiophiles!! In this particular case, the papers cited are providing evidence that we CANNOT hear an inter-aural delay of as little of 10uS with music.

    And, for the umpteenth time, it is trivially easy to test this, you can download raw un-mixed recordings, play around with them in a free editor/DAW, discover for yourself the required conditions and limits of ITD and actually fulfil this thread's title instead of perverting it!!

    Which is completely the opposite of what I believe because that's pretty much the very last thing anyone, including consumers, would want and it's clearly impossible anyway. Who would want to play back a rock/pop song (for example), which was a couple of days of unaccompanied, unedited, unmixed drum takes, followed by a couple of days of unaccompanied, unedited, unmixed lead vocal takes, then several more days of guitar takes and contain none of the samples or synths employed because they only existed as data or analogue signals and never existed as "sound waves, as measured in the air, at the listener's position". AND, even if someone did want their rock/pop song to be disassembled back into an "exact duplicate [of] the sound waves, as measured in the air, at the listener's position", technology is nowhere near capable of that feat. What consumers actually want is a manufactured, produced/manipulated montage (mix) that lasts just a few minutes, not the actual week or so of the sound waves that actually existed "at the listener's position". As far as I'm aware, no one tries to design a consumer reproduction system so it delivers an exact duplicate of the sound waves that would exist at the listener's position. Assuming high-fidelity is the goal, consumer reproduction systems are designed to reproduce (and transduce) the input signal as accurately as possible but DEFINITELY NOT, exactly duplicate the sound waves at the listener's position. Please provide some evidence/examples of systems which are designed according to your claim!
    As far the the actual input signals that a reproduction system should be trying to reproduce are concerned, they are almost exclusively designed to "a)" or more accurately: To create the conditions for a listener to experience the desired perceptual experience.
    1. How much money do you think you would you have to budget to get around the laws of physics?
    Scientific measurement mics are NOT "far better than anything likely to be used in a studio", they're far WORSE! I worked in a world class studio numerous times whose mic collection was valued at about $10m, the most expensive scientific measurement mics are about $10k -$20k, peanuts compared to the amount of money the top studios have invested in mics. So why don't they ever use scientific measurement mics? A scientific measurement mic is designed for a specific purpose, to measure a specific aspect/component of sound waves, such as frequency content, very high SPLs or very low SPLs but they can ONLY do so at the expense of sensitivity in the areas they are NOT designed to measure. So for example a specific measurement mic might measure very high SPLs much better than any studio mic but at the expense of much worse frequency response and/or self-noise and output sensitivity. Overall therefore it would be far WORSE than a studio mic because the signals we are recording need low self-noise, reasonably high output sensitivity and reasonably good freq response otherwise it simply will not be able to record all the components of these signals!

    2. And I would also reiterate that, from a scientific point of view, some technology does NOT change very rapidly or hardly at all in some cases. So your assertion is irrelevant, misleading and effectively a fallacy unless you specify what technology you're talking about and whether it applies to the specific equipment or point you are making assertions about. Microphones are one area of technology that has NOT changed very rapidly, in fact the fundament technology has hardly changed at all over many decades. This is in stark contrast to digital technology which has changed/advanced massively over the same period. I'm extremely surprised you are so ignorant of this, even to the point of "reiterating" that same fallacy?

    Last edited: Dec 15, 2018
  11. Phronesis
    Someone shared this short article on auditory memory with me, which I think is quite good:


    There are apparently various models of auditory memory, and this model has three types of memory:

    - Perceptual auditory storage: Very short-term sensory memory for a fraction of a second, which is quite detailed, but it's raw unanalyzed detail

    - Synthesized auditory memory: Memory which goes back several seconds, has substantially less sensory detail than perceptual auditory storage, and has some general features of the sound stored based on analysis of perceptual auditory storage

    - Generated abstract memory: Long-term memory which can be stored indefinitely, and lacks sensory detail, instead capturing the gist of what was heard as an abstract representation

    These different types of memory have important implications for listening impressions and tests.

    When we do typical controlled tests, we're usually comparing A and B based on synthesized auditory memory from music excerpts which are several seconds long, so there's a loss of sensory detail involved. In addition, the memory reflects the level of analysis specifically associated with synthesized auditory memory, so some features of the sound will be abstracted out and other features won't. It's NOT like simply replaying detailed recordings of A and B and comparing them.

    When we describe the sound of a headphone that we haven't heard recently, we're working with generated abstract memory, which is quite schematic and lacking sensory detail. So when we say something like "the bass is strong but a bit bloated, the highs are somewhat rolled off and smooth, and mids are clear and realistic," it's those attributes that we've remembered and are recalling, not a detailed sensory image of those attributes.

    It's interesting how all of this stuff with memory happens seamlessly at a subconscious level without our being aware of it. But we can readily notice all of this if do some real experiments and thought experiments and pay attention to how our memory is working.
    Last edited: Dec 15, 2018
  12. KeithEmo
    I agree with you 100%.

    They are saying that a difference that small would only be likely to be audible under very specific circumstances. However, you cannot say for a fact that those circumstances will never occur in an audio recording. I'm pretty sure that most synthesizers these days can make a pure monaural sine wave at 1 kHz... and they could certainly play a sample containing one... and I've definitely heard electronic music where the artist played a pure tone and "made it fly back and forth" by adjusting the relative phase between the channels. And, unless you can say that we will NEVER encounter that situation in music, then you must concede that "it may be audible on some recordings". Bear in mind that we're talking about 1 kHz here... and not anything ultrasonic.

    As you say, it is trivially easy to prove, and the scientists who conducted that test have already done so.
    I'm pretty sure that, if you or I were to repeat that same test, we would arrive at the same results.
    (I've read summaries of several tests that replicated those results... and that minimum value.)

    Please note that I said it was POSSIBLE.
    You responded that it was not possible.
    I responded by presenting published test results showing that it is in fact possible.
    Now you respond that "nobody would want to do it anyway".
    (I would agree - at least that it's quite unlikely to matter in any but a few really obscure situations.)

    You seem to persist in confusing what equipment currently in use can do with what's POSSIBLE.

    In SCIENTIFIC terms, a modern recording studio is NOT "the definitive level of current technology"; and, of course, even current technology is by no means a definition of what's possible.

    If I wanted to record cymbals, from six inches away, with a response up to 50 kHz, first I would check the specs for currently available microphones. If none of them was able to do it well enough, I would request a few quotes from the folks who make high speed measurement equipment. Then, if they didn't sell something off-the-shelf that could do what I wanted, I'd consider designing one. (A microphone is simply a device for measuring air pressure over time. First I'd think about using multiple currently available microphones, each dedicated to a particular frequency range, and combining their outputs digitally. Or, I might consider making a microphone, or having one custom manufactured, specifically designed to do what I want. I suspect that using a thin platinum diaphragm, and tracking its position with a LASER interferometer, might be practical and economical. Or, perhaps, a very thin Kapton diaphragm, sputtered in platinum, would be more practical. Another possible solution might be to fire an electron beam through a small chamber, and read the air pressure from the density of the air itself, based on how much the beam is attenuated. I'm also not convinced taht I'd rule out piezoelectric crystal and polymer colutions. Those are just some simple possibilities - off the top of my head.) And, yes, developing such a device might cost more than the cost to build a typical high end sound studio. (Luckily, in science, we don't have to worry about "consumer practicalities". The latest gravity wave detector, which finally worked after several previous failures, cost just over $600 million.)

    And, yes, I agree that building such a device might be extremely impractical... but it would NOT be IMPOSSIBLE.
    (And, unlike recording gravity waves, there might simply not be enough interest to fund developing it.)

  13. bigshot
    The upscaler is called OSSC https://www.videogameperfection.com/products/open-source-converter/

    He has a whole bunch of vintage systems that patch into a converter/switcher that switches between them and converts to HDMI. When he plugs that in direct, he gets both sound and picture.

    All of the audiophool pseudo-scientific justifications are getting trotted out now! We've gone from inaudible frequencies to inaudible phase shifts. Next we can talk about inaudible jitter, inaudible levels of distortion and inaudible stuff that we can hear but we're too much of a cave man to know how to measure yet! The guy who has never set foot in a recording studio is telling the guy who runs one how primitive recording studios are. Welcome to Bizzaroland!
    Last edited: Dec 15, 2018
  14. bfreedma
    That’s an impressive upscaler, but it’s also a lot more complex than the typical plug and play versions. It looks like there are specific settings for each game system and display combination. There’s an active support forum - I’d suggest posting there if my shot in the dark below doesn’t get it to output both audio and video. It also looks like it’s a bit of a work in progress, so make sure the firmware is up to date.


    Sorry I can’t offer anything more specific - would have to be in front of his setup to really diagnose the issue.
  15. gregorio
    1. There's what, roughly 1 million commercial tracks/songs released per year and probably a similar number of Films, TV programmes and commercials? Please provide a single example, in the whole of recording history where those circumstances have occurred.

    2. Sure they can but what's that got to do with it? You need 2 identical 1kHz tones, NOT just a (one) pure monaural sine wave at 1kHz, they need to be panned hard right and hard left and then one delayed by 10uS. That's certainly doable BUT
    2a. Firstly, that's extremely rarely ever done because unless the listener is sitting precisely between the speakers then the amount of "back and forth" will be different for each listener, which is why we use amplitude panning and almost never psycho-acoustic panning (phase based panning) and Secondly, on those very rare occasions when it is done, the relative phase is adjusted in the several millisecond range, NOT the micro-second range AND ADDITIONALLY, you clearly haven't thought about the laws of physics, AGAIN!! Sound travels though air at the speed of 0.343 mm (less than 1/64") per micro-second. Therefore, if your head is just 0.343 milli-metres away from dead centre between the drivers (speakers/HP drivers) then our 10uS difference in arrival time between the two 1kHz sine waves could fall below the demonstrated threshold. To put it another way, if your head is 3.43mm away from dead centre between the drivers, the actual arrival time of the 10uS difference would be either 20uS or 0uS! Thirdly, when played by speakers, the 1kHz tones at the listening position will also have room reflections, which completely breaks the required conditions!
    3. I CAN say that we will NEVER encounter that situation in music and therefore I do NOT concede that "it maybe audible on some recordings"! Creating that set of conditions would not be artistically desirable AND even if it were, in practice how many consumers ensure their head is within 0.343 mm of dead centre between the drivers and will never listen to the piece/song with speakers? If it's none, then no consumer would be able to hear that desired effect anyway, so what would be the point of going to the effort of creating it? Clearly it's a ridiculous assertion but I'm perfectly willing to be proven wrong, just provide a single example.
    4. I stated it was not possible to hear an arrival time difference of 10uS in music. I did note that you said it was POSSIBLE, which is I'm responding because your assertion is incorrect!
    4a. Which is correct!
    4b. No you DID NOT!! You presented published tests showing that it is ONLY possible to hear a difference in arrival time of 10uS under conditions which do NOT exist in music recordings (or listening to music recording). You're just going around in circles, stating something is "possible" that in fact does not exist, which of course is a nonsense fallacy. If gravity didn't exist, then pigs could fly. Science (and everyone else) has it all wrong, flying pigs ARE "possible"!!

    Can you please explain how it's "POSSIBLE" to record a rock band (or in fact anything) with a microphone that hasn't been invented yet?? :beerchug: ... :popcorn:

    Last edited: Dec 16, 2018
762 763 764 765 766 767 768 769 770 771
773 774 775 776 777 778 779 780 781 782

Share This Page