1. #1
    Join Date Nov 2006
    Posts 6,795
    Rep Power 2500000000

    Default audio measurement. the answers to life, aww screw it....42.

    Let's talk about the specifics of audio measurement, how specs can be relevant, how they can be MADE relevant, and most importantly:

    "How can we completely describe the fidelity of a system".
    "...but ma, audio engineering IS gainful employment!..."

    "...If I wuz at that club where Miles played one note I would have bounced ONE BOTTLE off his shiny fucking coconut. What? He's Phil Glass now?..." -Slipperman

    "...never attribute to magic, that which can be explained by conspiracy."
  2. #2
    Join Date Nov 2006
    Posts 1,887
    Rep Power 536871071

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Excelent idea, and I'll make this a sticky for a while.


    malice
    "If the party's good enough, you can actually suck to a remarkable degree."

    B.Ohlson, oct 2012
  3. #3
    Court of The Crimson King of The Road Bought Bre-X high... still holding hope
    Join Date Nov 2006
    Location on a big ball flying through space
    Posts 4,563
    Rep Power 536871161

    Default Re: audio measurement. the answers to life, aww screw it....42.

    mmmmmm...... fresh worms!


    JLE
    "A machine is only as effective as its user"-Wilbur Wright

    "I've suffered for my music, and now it's your turn"-Neil Innes

    "I am amazed every single day how so many think that you have to constantly DO things to engineer properly. In fact, you do as little as possible, and then only the things required. The trick is in knowing what is actually required."-Terry Manning

    "WE'RE ALL VILLAGE IDIOTS AT ONE POINT OR ANOTHER. "-der Schlippermensch

    Facebook me!
    blah, blah, blah
  4. #4
    Join Date Nov 2006
    Posts 6,795
    Rep Power 2500000000

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Ok, so it is the bright pink mastodon-resurrected-from-frozen-DNA in the room, that audio component measurement, and the SELF-PUBLISHING of those measurements, is a black art, if not a black box.

    We saw, from a simple anecdotal post in the "pathetic" thread, by our own Manroom Studios, that two different pieces of gear (preamps, to be exact) had rather remarkable, and rather divergent, published specs. One of the items was a highly regarded top-shelf bit of kit, a Neve preamp, and the other was a decidedly bottom-feeder prosumer bit of nonsense....and they both spec'ed out rather remarkably well. By reading those specs, you'd be enticed to select the bottom-budget gear over the "audio royalty" piece! Yet in side-by-side listening, it is the rare punter that would actually make that same choice, particularly if the punter had no access to price or spec information!

    what gives here?
    "...but ma, audio engineering IS gainful employment!..."

    "...If I wuz at that club where Miles played one note I would have bounced ONE BOTTLE off his shiny fucking coconut. What? He's Phil Glass now?..." -Slipperman

    "...never attribute to magic, that which can be explained by conspiracy."
  5. #5
    Join Date Mar 2010
    Location Sunny Seattle
    Posts 836
    Rep Power 536870995

    Beer Thirty Re: audio measurement. the answers to life, aww screw it....42.

    Let's talk about the specifics of audio measurement, how specs can be relevant, how they can be MADE relevant, and most importantly:

    "How can we completely describe the fidelity of a system".
    Well, if I may...

    A few things to consider. The ear is (literally) a mechanical frequency analyzer made out of some gooey wetware. Everything that goes into the cochlea (inner ear) from the outer and middle ears (different issues, maybe later) goes along the "basilar membrane".

    Because of the design of the basilar membrane, high frequencies go through it (sorta, kinda like a travelling wave filter if you're that old, but only kinda and sorta, it's got active components too) at the very entrance. Low frequencies propagate all the way to the other end.

    For any give frequency below 16 or 17 kHz (for which the peak is pretty much right at the entrance) as frequency goes down, the peak displacement of the membrane goes up at a point farther into the ear. Somewhere in the 100Hz or so range, the peak reaches the other end, and again stays there.

    My point?

    Any kind of measurement that relates to what you actually hear has to be frequency sensitive, and provide at very least a spectrum of both the original signal and the added components (be they noise or whatever).

    So the first thing that I think is a given is that "one number don't mean jack".

    Now, at any point point on the basilar membrane, you can calculate an equivalent rectangular bandwidth (filter bandwidth calculated in a mildly unusual but very useful way). This is called an "ERB" (An older measurement was called a "critical band" the same comments apply with some caveats to critical bands, i.e. the "bark" scale). Signals within an ERB can mask each other. This is why a small disturbance in a signal near it in frequency can not be heard. At best, inside the same ERB, you can detect something about 30dB down from the total energy in that ERB.

    BUT the filters are, politely put, steep. An ERB is more or less (and there are some mistaken "ERB" scales for low frequencies, say below 200Hz, so watch out) about 1/4 octave or so, or about 60-70Hz at low frequencies. (the spaciing is more or less uniform at low frequencies, and starts to grow when a 1/4 octave becomes larger), and 1 ERB above a signal, the filter leakage is about 25dB down. 1ERB below, even more, about 45dB. Two erb's below, more like 90. Above, the 25 db drops by another 15 db/ERB per ERB, at normal levels. (At too-loud levels, it may not drop at all!!!!)

    The point?

    Well, if we take a 1 ERB wide noise signal, and add another (different) noise signal in the same ERB to it, we can't hear that with only 5dB of SNR (think of SNR as enhanced THD if you don't know the term, it can be described later). HOWEVER, if we take a signal at 15kHz that is loud, and another "probe" tone at 2kHz, you may hear that tone when it's in excess of 60 or 70dB down from the 15 kHz signal, and this is limited mostly by equipment and room noise. It could probably get to 90dB under perfect circumstances, but nobody's done that yet.

    The point? Between 5 dB and 90dB, SNR means jack (and ThD less than jack built) without knowing the signal spectrum.

    Hence: My proposal for the first thing to be presented:
    Signal Spectrum
    and
    Noise Spectrum

    I'd use a broadband multitone with anharmonic tones so that the difference tones are easy to see, myself.

    Hmm, my fingers are tired.
    Grumpily,
    jj,
    Stalwart defender of Fourier, Fletcher, and Shannon, and who speaks, at most, for himself.
  6. #6
    Join Date Mar 2010
    Location Sunny Seattle
    Posts 836
    Rep Power 536870995

    Default Re: audio measurement. the answers to life, aww screw it....42.

    By reading those specs, you'd be enticed to select the bottom-budget gear over the "audio royalty" piece!
    Well, nobody can stop people from making useless specs. See my comment in the "pathetic" thread for the way various manufacturers actually measure their specs.
    Grumpily,
    jj,
    Stalwart defender of Fourier, Fletcher, and Shannon, and who speaks, at most, for himself.
  7. #7
    Owns and wears an Elvis jumpsuit Owns a kick-start dildo
    Join Date Jun 2008
    Location In a little African country do
    Posts 2,360
    Rep Power 536871068

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Nice intro about the ear... saves me a lot of typing.

    Next item: Electronics.

    Lets for a second forget about the ear, and think about what we can measure, and what mathematically defines quality in a signal product.

    Noise and distortion are the first to crop up. Then frequency response and then dynamic range.

    This is known, and measured to death in electronic systems including our precious audio equipment. What seems to be missing is an index of what it really means, and what combinations of it does to the signal.

    I think the problem is that some systems performs great in a certain test if you test for it - only. Take frequency response for instance. If you give it a sweep, or even a chirp, and analyse it at the other end, you don't see the effect that normal audio energy in the spectrum has on the whole. It is like testing a piece of metal for environmental endurance by getting a bucket of seawater, and keeping it in it for a month, while it only lasts 1 week bolted onto a pier.

    That may be why a well built system with modest spec on paper actually outperforms another with much higher spec sonically.
  8. #8
    Join Date Nov 2006
    Location 10.1.1.3
    Posts 2,175
    Rep Power 494928028

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Serious question, who when buying gear actually reads/considers the specs?

    Is it necessary?
    "Art is the expression of imagination, not the reproduction of reality." - Henry Moore

    >> York St Recording Studios <<
  9. #9
    And Then there Were Three Dog Night Group sex therapist
    Join Date Nov 2006
    Posts 4,878
    Rep Power 536871171

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Serious question, who when buying gear actually reads/considers the specs?

    Is it necessary?
    Problem is that as it was shown in the other thread, specs don't tell much. If there was a standard derived for audio measurement, or at least the specs would be measured by a third-party the situation might be better.
    When in doubt, mumble!

    EVERYTHING SOUNDS LIKE SHIT IF YA LISTEN LONG AND HARD ENOUGH.
  10. #10
    Owns and wears an Elvis jumpsuit Owns a kick-start dildo
    Join Date Jun 2008
    Location In a little African country do
    Posts 2,360
    Rep Power 536871068

    Default Re: audio measurement. the answers to life, aww screw it....42.

    ...or at least the specs would be measured by a third-party the situation might be better.
    ...or the specs based on a final production unit, with all the corners cut as per the paper pushers, and not the immaculately assembled prototype.
  11. #11
    SoundBlaster™ Expert Regular person but for the extra nipple
    Join Date Nov 2006
    Location On the bright side of life!
    Posts 4,159
    Rep Power 536871147

    Default Re: audio measurement. the answers to life, aww screw it....42.

    that two different pieces of gear (preamps, to be exact) had rather remarkable, and rather divergent, published specs. One of the items was a highly regarded top-shelf bit of kit, a Neve preamp, and the other was a decidedly bottom-feeder prosumer bit of nonsense....and they both spec'ed out rather remarkably well. By reading those specs, you'd be enticed to select the bottom-budget gear over the "audio royalty" piece! Yet in side-by-side listening, it is the rare punter that would actually make that same choice, particularly if the punter had no access to price or spec information!

    what gives here?
    My experience is that, no matter how good or how bad the specs are, or how good or how bad a device actually is, switching to a better power supply improves its abilities, often without changing the published specs.

    Of course if there's already a top-notch power supply included there's not much room for improvement anymore.

    Mac
    .
    Mac

    Don't be worried about your anosognosia.

    Being a professional is not about how good you are at your best,
    but how good you are at your worst.

    Creating complex solutions is easy.

  12. #12
    Owns and wears an Elvis jumpsuit Owns a kick-start dildo
    Join Date Jun 2008
    Location In a little African country do
    Posts 2,360
    Rep Power 536871068

    Default Re: audio measurement. the answers to life, aww screw it....42.

    My experience is that, no matter how good or how bad the specs are, or how good or how bad a device actually is, switching to a better power supply improves its abilities, often without changing the published specs.
    GIGO principle. Garbage in, garbage out, and that applies to everything you feed into the amplifier...
  13. #13
    wardrobe malfunction investigator "Mock Chicken" joke?...could be American.
    Join Date Apr 2007
    Location Ireland
    Posts 43
    Rep Power 106665536

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Problem is that as it was shown in the other thread, specs don't tell much. If there was a standard derived for audio measurement, or at least the specs would be measured by a third-party the situation might be better.
    The great thing about standards is that there are some many of them. As we know, a manufacturer will pick the standard that gives the most impressive figures. Test gear is expensive, so we rely on our ears most of the time. Null tests and blind A/B tests can give us a lot of useful information, but it's easy to mess up the procedure. For these reasons, defining sound quality on paper is challenging. But not impossible.

    I do use specs as part of the process of elimination, but it is at best a blunt instrument.

    I always thought that full bandwidth and isolated phase change was not audible. Clearly I need to look up "all pass filter".

    j_j, could you recommend a book or a link on designing listening tests please?
  14. #14
    Roadside Outcast Dept. Always picked last for sports
    Join Date Nov 2006
    Location On location...
    Posts 4,906
    Rep Power 536871172

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Serious question, who when buying gear actually reads/considers the specs?

    Is it necessary?
    Only as far as determining candidates for a listening test.

    All purchasing decisions are then made on results of those tests.

    Cheers,
    Tim
    Don't forget, we are all engaged in a battle to the death against mediocrity.

    The best radio mic system that money can buy is ALMOST as good as a $20 cable.

    One of the most important things to remember about sound is:
    'Sucks' is always conducted better than 'Rules'. - Pimp-X wisdom


    Never underestimate the power of stupid - Blackie C (RIP)

    Ego and talent seldom go hand in hand... Talent and humble on the other hand... - Zoesch

    Weedy ignores this simple bit of glaring obviousness because he is an "ELECTRIC BASS GUITARIST"(coughcough)
    and views the kick drum as a "bass riff rhythmic pattern suggestion generator" - Slipperman
  15. #15
    Join Date Nov 2006
    Posts 6,795
    Rep Power 2500000000

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Tim, that's very interesting, but it glosses over the fact that there's this big, complex, supply chain out there that is driven by customer knowledge and expectations (or lack thereof). The published spec drives the supply chain demand. Stores will not stock stuff for you to listen to if the manufacturer's marketing hasn't created a demand for it.

    This is of course all smoke and mirrors, but it's a reality. And of course, there's the online market where people buy cold.
    "...but ma, audio engineering IS gainful employment!..."

    "...If I wuz at that club where Miles played one note I would have bounced ONE BOTTLE off his shiny fucking coconut. What? He's Phil Glass now?..." -Slipperman

    "...never attribute to magic, that which can be explained by conspiracy."
  16. #16
    Roadside Outcast Dept. Always picked last for sports
    Join Date Nov 2006
    Location On location...
    Posts 4,906
    Rep Power 536871172

    Default Re: audio measurement. the answers to life, aww screw it....42.

    dwoz,

    I hear you about the consumer-driven demand, but I don't go to the stores...although I get that many people do.

    I do audio-post for film and television for a living, so the gear that I'm looking at isn't usually seen in the prosumer retail outlets anyway.

    I deal directly with the manufacturers or their agents.

    They are more than happy to provide information/specs without the marketing hype, and supply units for evaluation.

    The specs are useful for determining which unit I should be evaluating...given that I live and work about as far from anywhere as it is possible to get.

    Cheers,
    Tim
    Don't forget, we are all engaged in a battle to the death against mediocrity.

    The best radio mic system that money can buy is ALMOST as good as a $20 cable.

    One of the most important things to remember about sound is:
    'Sucks' is always conducted better than 'Rules'. - Pimp-X wisdom


    Never underestimate the power of stupid - Blackie C (RIP)

    Ego and talent seldom go hand in hand... Talent and humble on the other hand... - Zoesch

    Weedy ignores this simple bit of glaring obviousness because he is an "ELECTRIC BASS GUITARIST"(coughcough)
    and views the kick drum as a "bass riff rhythmic pattern suggestion generator" - Slipperman
  17. #17
    Join Date Mar 2010
    Location Sunny Seattle
    Posts 836
    Rep Power 536870995

    Default Re: audio measurement. the answers to life, aww screw it....42.

    The specs are useful for determining which unit I should be evaluating...given that I live and work about as far from anywhere as it is possible to get.
    Coober Pedy? The intersection of 87 and Rt. 4? Veever's Crater?

    More seriously, there is nothing we can do about manufacturers who spec their one test unit, hand-tweaked in a faraday cage with a special power supply and RF filtering into and out of the cage.

    But at least wouldn't it help to have some specs that might actually carry some information, about equipment working in-situ, with real signals, and some meaningful sense of what it might sound like, at least on the gross impression level. (I submit we don't get that from the "Spec's" in common use, although I will say that if there were meaningful ways to express the infamous "4", it would be a start. I'd prefer a different set, myself, though. Might only be 4, but not 4 single numbers, thank you.)
    Grumpily,
    jj,
    Stalwart defender of Fourier, Fletcher, and Shannon, and who speaks, at most, for himself.
  18. #18
    Join Date Nov 2006
    Location 10.1.1.3
    Posts 2,175
    Rep Power 494928028

    Default Re: audio measurement. the answers to life, aww screw it....42.

    I'm probably being naive, but i thought what created demand for
    a product were people saying "i've used this and it sounds
    great"... or at worse "this product is GOING TO sound awesome".. ?

    If specs CAN be made to be informative of how a unit will
    sound to someone, how would it appeal to someone like me?
    I'm not an EE by any stretch so i can't see how specs are ever
    going to be something i look at. I don't even know what "ideal"
    specs should be and frankly don't care. Let me try it.
    From tracking to mix. That's how i'll use it and end of the
    day that's what i care about.
    "Art is the expression of imagination, not the reproduction of reality." - Henry Moore

    >> York St Recording Studios <<
  19. #19
    And Then there Were Three Dog Night Group sex therapist
    Join Date Nov 2006
    Posts 4,878
    Rep Power 536871171

    Default Re: audio measurement. the answers to life, aww screw it....42.

    Might only be 4, but not 4 single numbers, thank you.)
    Yeah, that where I had problem with Ethan's statement.

    You can probably categorize system properties within 4 "domains" and measure it accordingly but that not equates to having only 4 numbers (or 1 graph and 3 numbers) and pretending to know all about the system.
    When in doubt, mumble!

    EVERYTHING SOUNDS LIKE SHIT IF YA LISTEN LONG AND HARD ENOUGH.
  20. #20
    Join Date Mar 2010
    Location Sunny Seattle
    Posts 836
    Rep Power 536870995

    Default Re: audio measurement. the answers to life, aww screw it....42.

    If specs CAN be made to be informative of how a unit will
    sound to someone, how would it appeal to someone like me?
    The question is exactly that, how can what we CAN measure (which is a lot) be reduced or abstracted so that you don't have to be an expert to understand the measurements.

    One thing you can count on, it won't be one number, or one plot.


    I'm not an EE by any stretch so i can't see how specs are ever
    going to be something i look at. I don't even know what "ideal"
    specs should be and frankly don't care. Let me try it.
    From tracking to mix. That's how i'll use it and end of the
    day that's what i care about.

    It might be possible in the long term to create a suite of things (we're talking digital here, for the calculations) that you could use to measure what you like, learn its characteristics, and then use THOSE to match to real specs. Not simple, but perhaps not impossible.
    Grumpily,
    jj,
    Stalwart defender of Fourier, Fletcher, and Shannon, and who speaks, at most, for himself.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts