Thoughts about distortion behavior

Discussion in 'Audio Science' started by purr1n, Jul 1, 2019.

  1. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    49,524
    Dislikes Received:
    81
    Trophy Points:
    113
    Location:
    Antarctica
    From here: https://www.superbestaudiofriends.o...easurements-and-review.915/page-4#post-262702

    All distortion is bad.

    However, I'd gather that a distortion pattern where successive harmonics are lower than the one just prior is typical or ideal. SET amps have this distortion pattern. People always like to talk about tube distortion. But the BIG one is that dynamic driver transducers (including compression drivers) also have this distortion pattern. And transducers' distortion a magnitude or two (or three) higher than that of amps an DACs.

    Hence I would posit that a 3rd harmonic higher or even than the 2nd harmonic is not typical or normal. This is why so many IEM nerfs such as @shotgunshane, @Kunlun, or myself long for the next Sony EX1000 which uses a DD, or why @Hands find the well-regarded CFA Andromeda a mess, or why @OJneg prefers the Jupiter and remarks on its glorious mids (which for a BA driver, actually has even 2nd and 3rd order distortion in that region)


    Even GRAS doesn't get it right in the highs as it produces coupler induced peaks which vary in amplitude and frequency based on insertion depth. This is why I urged @crinacle to stay with this cheapo DIY rig for the high measurements on the IEMs.
     
    Kunlun and Lyander like this.
  2. Lyander

    Lyander Too sensitive for SBAF

    Friend
    Joined:
    Feb 25, 2017
    Likes Received:
    4,357
    Dislikes Received:
    23
    Trophy Points:
    113
    Location:
    Pasig, Philippines
    Ahh, gotcha. So the EDC3 is good for the ~$100 category but the mids-on-down 3rd order>2nd order performance is actually far from ideal? I'm aware that there are some who opine judicious amounts of distortion might actually make something more euphonic, glad that my head was on straight for that.

    Logically speaking it makes sense that distortion should act in that way (i.e. successive orders lower in amplitude) unless something odd is going on in the enclosures, just curious what factors might influence a transducer to have greater higher-order distortion? I know amps can also sin here, but also know that they are an entirely different beast in this regard and way beyond my ken at the moment, so not even gonna touch that.

    But the Verite! (/sarcasm).

    Would be curious about finally getting a pointed answer about that really, I should have been more direct in the thread. Personally I don't give much of a bother about distortion that low in the spectrum since.... I myself might have trouble caring or even hearing mud that far down, but I understand there are definite benefits to LCD2C A-grade bass. Never took a pair home but kinda fell for an LCD2 during a meet years back, bass was the definite highlight. Can't remember whether it was fazored or not though, might have been @jacq's pair when she visited the Philippines?
     
    Jinxy245 likes this.
  3. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    49,524
    Dislikes Received:
    81
    Trophy Points:
    113
    Location:
    Antarctica
    Further thoughts on "all distortion is bad". I would like to correct this.
    1. Distortion after a certain point is bad.
    2. Certain distortion patterns are worse (or more atypical) than others.
    And perhaps controversially: no distortion is actually bad.

    The reason is that all audio recording and playback gear produce some amount of distortion. Who signs off on the final mix and master? The producers, maybe the execs, the artists? Whatever they sign off on, it's being played back in the mastering room on stuff that makes distortion. Now, look around the sound houses. What do they use? Speakers with dynamic drivers, maybe compression drivers with horns.

    I think there's something to be said regarding amps with a similar type of distortion pattern as the transducers. Pass's amps, SET, the insanely expense Dartzeel monoblocks, etc. And why I never cared some the Adcom Class ABs (high 3rd order).

    You can argue that 3rd order harmonics dominating or being even to 2nd order harmonics on BA drivers has its charms. Odd order distortion will square up a waveform. Kind of like turning up the sharpness dial. It's a novelty.

    I'll show you an LCD-3 with C-grade bass in a bit... I measured such an LCD-3 several weeks ago but have not posted.
     
    Last edited: Jul 1, 2019
    Ash1412, Kunlun and Lyander like this.
  4. Lyander

    Lyander Too sensitive for SBAF

    Friend
    Joined:
    Feb 25, 2017
    Likes Received:
    4,357
    Dislikes Received:
    23
    Trophy Points:
    113
    Location:
    Pasig, Philippines
    I sing in the shower. A lot. Without the reflections off the tiling making me sound less like dying cats, I'd be a hell of a lot quieter— I can see the argument for absolutely no distortion sounding horrible. Kidding aside:

    Hah, this is an argument that loads miss, I think. Heck I did too until relatively recently. Yeah, this is what eventually made me give up on chasing neutral sound, since for all I know the person behind the mix could be aiming for a sound I don't appreciate despite liking the music itself. Or maybe I'm trying to rationalise not minding recessed mids. Also there's the matter of distortion being inaudible past a certain point, or it not mattering unless you crank volume. Latter was already addressed elsewhere though, would be redundant here.

    Okay I've honestly not heard enough gear in the home setting to be able to imagine what that sounds like, but the analogy works. Samsung cellphone camera postprocessing, basically? Ew.

    I perceive timbre as a function of FR and decay mostly, but distortion would alter it in a big way too wouldn't it? I'm willfully exposing ignorance here but given how distortion is effectively the activation of signals higher up the spectrum given fundamentals at X, if I've not misconceived it, how might higher D2/D3/D4 affect pitch perception, and might this have anything to do with how sonic images become "fuzzier" in a sense at higher levels? IIRC it's stuff in the upper frequencies that make things sound "tighter", "more present", or cleaner-cut, so would this be the reason why D3 sharpens things?
     
    Grattle likes this.
  5. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    49,524
    Dislikes Received:
    81
    Trophy Points:
    113
    Location:
    Antarctica
    Chasing neutral sound is still good. Any self-respecting sound house will have neutral sounding speakers. Some will even calibrate. I'll ask the head engineer where I work what he does with the big primo mix stages. I see huge JBLs and Crown amps.

    I think there is a certain sense to trying to mimic the distortion characteristics of the speakers in the sound rooms. This is not to say that transducers with different characteristics cannot be enjoyable.

    I think I mentioned this elsewhere. If in an alternate universe, Quad (electrostatic) took off and owned the professional market instead of JBL, then all the music would be mastered differently.
     
  6. Psalmanazar

    Psalmanazar Non Voting Friend

    Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    2,991
    Dislikes Received:
    102
    Trophy Points:
    113
    There would be no movie explosions, no shouts, no screams, no bombastic orchestral scores and dynamics would be all over the place.

    For music mastering if they replaced the B&Ws, ATCs, Egglestonworks, and Keith Klawitter KRKs? The loudness war would be e even worse or compression wouldn’t be used at all and nothing would gel together in most recordings.
     
    Lyander, Ash1412, OJneg and 1 other person like this.
  7. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    2,638
    Dislikes Received:
    4
    Trophy Points:
    93
    Piles of burnt panels outside studios.
     
  8. OJneg

    OJneg The Most Insufferable

    Friend BWC Executed
    Joined:
    Sep 25, 2015
    Likes Received:
    3,744
    Dislikes Received:
    34
    Trophy Points:
    113
    Location:
    Grand Rapids
    I was willing to chalk up my preference of the Jupiter (compared to Andro) as purely a matter of tonal (i.e. FR) signature, but the THD characteristics of each could at least be a contributing factor now that you have shed some light on it.

    A point that I would add to the general discussion might be that the way THD plots show us information on a FFT is not quite how we process information as time-domain creatures. In signal processing it's understood that you can pass a tone though a DUT and that the nonlinearity of the DUT "creates" a spurious tone related to the input tone at some given harmonic. When we listen though, we're not hearing that additional tone separate from the fundamental, but with the fundamental. Just as we don't hear a guitar pluck with its separate harmonics; we just hear the pluck and integrate all the accompanying harmonics into what we call the sound of the guitar.
     
    Grattle and Lyander like this.
  9. OJneg

    OJneg The Most Insufferable

    Friend BWC Executed
    Joined:
    Sep 25, 2015
    Likes Received:
    3,744
    Dislikes Received:
    34
    Trophy Points:
    113
    Location:
    Grand Rapids
    I am glad that we're in the habit of showing THD plots with each different harmonic component vs frequency. Manufacturer specs that claim < X.X % are inadequate to show us the true character of the devices nonlinearity. I would suspect that a component that added slightly more distortion with an even curve would sound better than a device that had spikes of distortion (especially when those spikes usually correspond to peaks/nulls in FR)

    Circling back, I would posit that the THD characteristics can be more or less relevant than people might initially think when looking at them. While it's easy to dismiss a dominant third harmonic as being 60dB below the music, it might be indicative of a more sterile or dissonant quality of the DUT. In other words, THD can be an indicator of certain sonic qualities rather than the reason behind those qualities. The true reason might be the nature of the design which that indicator is simply pointing to, but we might not be able to quantify.
     
    Last edited: Jul 2, 2019
    Psalmanazar likes this.
  10. Lyander

    Lyander Too sensitive for SBAF

    Friend
    Joined:
    Feb 25, 2017
    Likes Received:
    4,357
    Dislikes Received:
    23
    Trophy Points:
    113
    Location:
    Pasig, Philippines
    I looked it up, so for similarly uncertain people DUT = "device under test". Knew I'd heard it before but had to confirm context.

    Yup, sorry if I was unclear but I agree that because the way we perceive is mediated by temporal elements, i.e. we process things as distinct saccades (or whatever the appropriate analogue would be here— slices?) within measures of time, there will be those close enough to simultaneous that we don't necessarily perceive them as separate, so for things like distortion and the effect this might have on timbre (by way of how it influences the behaviour of harmonics), they do alter how we perceive an individual signal might sound. Or at least I think so, based on what I know of measurements. Also not sure I didn't butcher that sentence.

    I'd be curious how distortion (THD I guess, because it's be a remarkable pain in the pooter to isolate the effects of each) might correlate with prolonged ringing in CSD measurements. Could be a matter of design too I guess, since I know there are headphones that pull off tasteful decay quite well, ergo distortion not being universally bad in the context of personal listening.
     
    Grattle and crenca like this.
  11. Ash1412

    Ash1412 Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    485
    Dislikes Received:
    2
    Trophy Points:
    63
    What I would say from trying out planars, dynamics and BA and comparing them against each other is that subjectively, lower THD matters for lower frequencies, while as the frequency rises, the behavior/complexity of the distortion starts to take over in importance. Although most people would consider HD650 bass as its worst quality but still passable since most of the distortion is 2nd order, I'd even go so far as to say its bloom is mildly headache-inducing for me (nothing to the level of closed-backs though), while planar bass never comes close to sounding fatiguing, at least when perforated/open/non-vacumn pads are used.

    However, I couldn't differentiate, beyond the HD650 softness, any more distortion between it and planars when judging upper midrange on up. And while Andromeda's distortion in terms of THD is pretty good, the BA signature is immediately present the moment you put them on. A worse case is the MD Plus which I had for loaner here for a week. While I liked the Massdrop Plus's resolve, any remotely loud song pushed its upper mids into sounding almost radio-like for me.

    I'm not completely agreeing with @purr1n about it sounding sharper since my subjective opinion of BAs is more like Hands description where it sounds like crumpled paper, as far as midrange goes, but BA distortion does start going from crumpled paper to sounding like a bit of sharpening when treble and transients are concerned, somewhere between the palatable bloom of 2nd order distortion and the transient metallic timbre of Focals. So maybe it's really a problem of what range people are focusing on when listening and the resulting impression combined with differing language use e.g. @shotgunshane described the MD+ as being "clear" or the Focals are described as "clear" while other reviews describe planars as "clear", which if one considers the above could mean wildly different things. But tbh, we're still lightyears away from crappy reviews that describe sound as "meaty" and "crispy".
     
    Last edited: Jul 2, 2019
  12. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    49,524
    Dislikes Received:
    81
    Trophy Points:
    113
    Location:
    Antarctica
    Or fleshy.
     
    data.noodle likes this.
  13. Armaegis

    Armaegis Friend

    Friend BWC
    Joined:
    Sep 27, 2015
    Likes Received:
    3,663
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    Winnipeg
    We've got all sorts of DSPs and doodads that muck around with FR to emulate something else or tune to a certain target curve. Is there anything similar in the realm of THD? Assuming your host transducer has lower(ish) whatevers than the target that you're trying to emulate, it could be an interesting experiment.

    I know there's all sorts of VST plugins that can create harmonics and distortions and tone emulators and whatnot, so I feel like it's doable.
     
  14. ChaChaRealSmooth

    ChaChaRealSmooth SBAF Gearmaster

    Staff Member Friend
    Joined:
    Dec 6, 2018
    Likes Received:
    2,389
    Dislikes Received:
    10
    Trophy Points:
    93
    Location:
    La Mirada, CA
    I'm kind of wondering how this would come to be. Aren't harmonics and timbres of different instruments by nature determined by some level of "distortion?" The reason why a full grand piano sounds different than an upright, etc.

    Not sure if this also deals with distortion, but reverberations aren't necessarily bad IMHO. Concert halls and recital halls certainly have a sound to them that is much different than my not-acoustically-treated house. Kind of similarly, I'm learning that I really enjoy the slight reverb of the Verite.

    It's also very possible that I'm actually deaf.
     
    Jinxy245 likes this.
  15. Lyander

    Lyander Too sensitive for SBAF

    Friend
    Joined:
    Feb 25, 2017
    Likes Received:
    4,357
    Dislikes Received:
    23
    Trophy Points:
    113
    Location:
    Pasig, Philippines
    Arguable. The way I understand it is that distortion as operationally defined here is the addition of elements not inherent to that which is being recorded. Vocal harmonics, instrumental overtones are not a product of reverb or gear distortion etc but are instead of the instruments themselves, so while concert halls, bathroom walls etc might accentuate these overtones, they exist in the original signal to begin with and are thus do not qualify as distortion. The way I understand purr1n's post above is that because the final mixes that engineers sign off on are evaluated on systems where a measure of distortion is present, playing back the same on a system with as-close-to-zero-as-physically-possible distortion will generate an unnatural sound, or at least one very much unlike the sound aimed for.

    Or that's my understanding of it anyway, could be wrong or misreading again. And yes some reverb can be awesome, I rather like a bit of it myself I think.
     
    Last edited: Jul 3, 2019
  16. ChaChaRealSmooth

    ChaChaRealSmooth SBAF Gearmaster

    Staff Member Friend
    Joined:
    Dec 6, 2018
    Likes Received:
    2,389
    Dislikes Received:
    10
    Trophy Points:
    93
    Location:
    La Mirada, CA
    That's the word that I was forgetting. I think I need to drink less when I post. And perhaps at a time of day when I'm more awake.

    Your point about the overtones being a product of the instruments themselves, and then the recording/playback gear having distortion makes a lot of sense to me.

    That being said, I'm not sure if I fully understand @purr1n here. Might require some more days of thinking.

    Oh well, I'm enjoying my music at this very moment so....
     
    Lyander likes this.
  17. Ash1412

    Ash1412 Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    485
    Dislikes Received:
    2
    Trophy Points:
    63
    A lot of it comes back to whats talked about in this video by an Audio Precision guy, so the literal measurement company (I think it's been posted by someone here already):

    Two points the video makes that struck me were:

    1. "Modern" measurements are conducted under very static, standardized conditions meant to facilitate calibration, replication and is primarily concerned with complete overload/meltdown/distortion above 1% situations.

    2. Actual music's energy spectrum is way closer to noise than the signals used in measurements.

    So actual music probably has a lot more complexity going on in it than we assume matters, a large portion of that being timbre, distortion, noise, etc... from everything that makes up the sounds and the recording chain. If anyone knows about analog film, they'd know the phrase "detail in the grain", the film equivalent of "signal beneath the noise" for vinyl folks and what happens when compression algorithms can't see through that grain (the better ones just blur it away, the worse ones cause artifact in the grain) or denoise filters are applied in bad digital masters. Perhaps what @purr1n means is that gear that try to cancel out their own distortion also tend to filter out "distortion" in the recording as well, instead of trying to recreate it. e.g. Mike Moffat's stance on delta-sigma or a lot of members here's dislike of high feedback amps and planars, even stats, so all things that measure insanely well below common standards for distortion.
     
    Last edited: Jul 3, 2019
    Psalmanazar, Jinxy245 and Lyander like this.
  18. OJneg

    OJneg The Most Insufferable

    Friend BWC Executed
    Joined:
    Sep 25, 2015
    Likes Received:
    3,744
    Dislikes Received:
    34
    Trophy Points:
    113
    Location:
    Grand Rapids
    A lot of DAWs can utilize "tube emulators". I think I recall some smartphone DSP settings that did the same thing to "enhance" the listening experience. :confused:
     
  19. Armaegis

    Armaegis Friend

    Friend BWC
    Joined:
    Sep 27, 2015
    Likes Received:
    3,663
    Dislikes Received:
    2
    Trophy Points:
    113
    Location:
    Winnipeg
    But now I'd like to see something where you can input a target THD measurement from another headphone, and try to pull (convolve?) your current headphone towards it... At the same time as the FR chicanery.
     
    Serious likes this.
  20. Serious

    Serious Inquisitive Frequency Response Plot

    Friend BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    1,671
    Dislikes Received:
    0
    Trophy Points:
    113
    Location:
    Munich
    Or use a bunch of measurements to create a variable digital filter of some kind that creates opposite phase harmonics at just the right levels to cancel out the distortion. But I'm sure that's also been done before.
     

Share This Page