Merv's Politically Incorrect Audio Blog

Discussion in 'SBAF Blogs' started by purr1n, Dec 26, 2018.

  1. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    56,481
    Dislikes Received:
    97
    Trophy Points:
    113
    Location:
    Antarctica
    It depends upon what you mean. I like objective numbers when they mean something. I disregard objective numbers when they become meaningless. I do not think that THD+N better than -80-85db matters or has any correlation to hearing. Certainly not near -90-100db and lower.

    What I mean by "measurement extremist" is those who actually care about performance between -95db, -115db, -128db. No one can hear that shit. Heck, even instrumentation a decade ago couldn't detect this stuff. For a designer, this may be important. For end-users no.

    People should feel to subscribe to the "measurement extremist" approach. However, such an approach is only of academic interest to SBAF. There are forums out there better suited to this approach. Any discussion taking such an approach at SBAF will be shut down. No, it's not a democracy here.

    Yup same thing. It's black and white to the extremists. No middle ground. There are people who measure, there are people who listen, and then there are people who measure and listen.
     
    Last edited: Jul 7, 2019
  2. OJneg

    OJneg The Most Insufferable

    Friend BWC Executed
    Joined:
    Sep 25, 2015
    Likes Received:
    3,750
    Dislikes Received:
    36
    Trophy Points:
    113
    Location:
    Grand Rapids
    If all audio measurements (and their advocates) disappeared tomorrow, I would be no worse off.

    If even the small core of audio gear reviewers and forum-goers that I consider trustworthy disappeared tomorrow, I would be totally in the dark.

    This is why I say measurements are ghey. I proudly don the title "Subjectivist Extremist", given the totality of the circumstances.
     
    mscott58, gandhisfist and bixby like this.
  3. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    56,481
    Dislikes Received:
    97
    Trophy Points:
    113
    Location:
    Antarctica
    Measurements are good for making shortlists of products to consider in the Internet age where there are few brick-and-mortar stores. I'd say they are more reliable than random Internet impressions and should mostly be used as an additional data point to subjective impressions.

    Measurements are also a good tool to separate the wheat from the chaff. If we had independent measurements of the A-GD Singularity 19 DAC, we would not have purchased it. I've declined to accept IEMs and headphones because I knew I wouldn't like how they would sound based on measurements from known sources.

    I use measurements to select drivers for speaker projects. This approach hasn't failed me yet - with the understanding that Seas tends to be the most honest, JBL next, and Fostex smooths the crap out of their graphs.
     
    Last edited: Jul 7, 2019
  4. atomicbob

    atomicbob dScope Yoda

    Friend BWC MZR
    Joined:
    Sep 27, 2015
    Likes Received:
    12,250
    Dislikes Received:
    4
    Trophy Points:
    113
    Location:
    On planet
    wise person use all data input:
    1. valid, reliable measurements
    2. valid, reliable impressions from trusted ears
    3. own ears whenever possible
    4. comparison of own listening experiences with other data inputs for calibration, further understanding of personal preferences and how it relates to other listeners' personal preferences.

    Best when one knows the difference between relative, absolute and preference.
     
  5. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Friend
    Joined:
    Sep 28, 2015
    Likes Received:
    2,874
    Dislikes Received:
    4
    Trophy Points:
    113
    They achieved more than .8 correlation between listener preference and on/off-axis measurements for speakers. I vaguely remember the standard suite of electronics measurements not delivering anything close to that in terms of correlation.

    The problem with the Olive/Welti curve is that it's tied to a specific measurement rig. My beef with Olive is that he does nothing to call people out for using his target curve improperly. There have been hundreds of people asking me about the SW target curve and all I can say is that it's well... flat. I can tell about the process how we came to find it, but without knowing our measurement rig, ANY curve would be useless and most likely - misleading.
     
  6. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    56,481
    Dislikes Received:
    97
    Trophy Points:
    113
    Location:
    Antarctica
    It would seem to appear that others, rtings, oratory, etc. are indeed using a measurement rig similar to Olive's no? Or maybe slightly different model numbers on the coupler, but close enough? I am still trying to ascertain this. I would have a hard time believing that these guys wouldn't at least make sure that raw curves measured the same as Olive's.
     
  7. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    56,481
    Dislikes Received:
    97
    Trophy Points:
    113
    Location:
    Antarctica
    You are asking people to be to too sensible. :)
     
    mscott58 and pavi like this.
  8. insidious meme

    insidious meme Ambivalent Kumquat

    Friend
    Joined:
    Sep 26, 2015
    Likes Received:
    3,129
    Dislikes Received:
    3
    Trophy Points:
    113
    Location:
    Sector 8023 of the Third Quadrant
    It beats an indefinite time in mid fi purgatory.
     
    pavi likes this.
  9. Psalmanazar

    Psalmanazar was: dingleberry; now: chill + interesting

    Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    3,689
    Dislikes Received:
    120
    Trophy Points:
    113
    Remember that Olive also used the god fucking awful sounding anorexic tin can HD 518 and the anti-music LCD2 to get that bassy target response based on the Harman's JBL boom boom curve. Of course people could easily prefer a bassy eq from that crap to make it less tin can or dead. He also didn't play the records he used on flat speakers for them first so I assume most of his listeners had no idea what each recording was supposed to sound like with regards to the production. There was no designation on what "trained" means. Do those trained listeners know what the instruments on each recording are supposed to sound like in actuality and how they sound on the record? Did they have any idea what the rough instrumental levels of each particular recordings were in each genre? Do they want to hear them or do they want some phat bass car stereo? Maybe with some tame pop vocal stuff, a v-shaped hifi loudness button curve actually does sound more involving. There are so many variables to even consider and problems with that ridiculous curve. Get a flattish pair of speakers, mess with the bass like that and tons of records will sound like boomy, muddy poop.
     
  10. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    56,481
    Dislikes Received:
    97
    Trophy Points:
    113
    Location:
    Antarctica
    Hahaha. Check out Oratory's suggested LCD2C curve. He's suggesting +5db EQ shelf in the bass by 60-70Hz starting from 200Hz and below to meet the Harmon target. I believe Oratory's measurement rig uses the same parts as Olive's rig. Yes, that's moar bass to the LCD2C. That much more bass on an LCD2C would make me want to commit suicide.
    https://www.dropbox.com/s/4341dbn3nrcou4k/Audeze LCD-2 Classic.pdf?dl=0

    As an aside, anyone who would suggest +5db bass on the big Audeze headphones seriously needs to get his ears and head checked.

    There were datasets posted on IF. Didn't look liked trained listeners to me based on the wide variances. There were dudes who preferred -2db on the bass and others who wanted +12db. (I might be exaggerating, but I'm pretty sure it was wide). Maybe the volunteers took a two hour Harmon golden ears computer based training course, possibly with shitty headphones, so who knows. The bottom line is that the Olive curve is a consumer curve averaged from the preferences of two-thirds random fuckers and one-third people who might have known what they were doing. I'd rather Olive grabbed the sound engineers from the more reputable film and TV post houses here in El Lay. I only mention film and TV because their mix stages tend to be better calibrated (Dolby Atmos, DTS, etc.)
     
    Last edited: Jul 7, 2019
  11. tomchr

    tomchr MOT - Neurochrome

    Joined:
    Oct 28, 2016
    Likes Received:
    505
    Dislikes Received:
    19
    Trophy Points:
    93
    Location:
    Calgary, Canada
    Home Page:
    A correlation of 0.8 is very strong, so clearly measurements are reflective of the perceived sound quality - at least in the case of on/off axis measurements of speakers. Thus, I find the notion of "measurements are useless" to be a bit misguided.

    Generally r < 0.40 is considered weak; 0.40 < r < 0.59 is considered medium; r > 0.60 is considered a strong correlation.

    That said, once the various measurements indicate that the equipment does not contribute anything audible to the sound, the arms race in measurements becomes a numbers game. The big question is where that audible limit lies for the various measurements. There's considerable individual variability there, so there probably isn't a "one size fits all" solution. So some of us choose to strive for perfection. At least then we can confidently say that the amplifier/DAC/whatever does not contribute any colour to the sound, so you'll hear the colouring of the transducer and that's it.

    Some of the numbers arms race also comes down to marketing. I think it's hard to argue these days that a piece of equipment with 94 dB SINAD is "state of the art", even if 94 dB SINAD is amazing and would have been state of the art years ago.

    Tom
     
    Priidik likes this.
  12. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    56,481
    Dislikes Received:
    97
    Trophy Points:
    113
    Location:
    Antarctica
    ^ This is what I am against. Where measurements become a numbers game. Promoted by measurement extremists.

    Except it doesn't. I can point out various amps and DACs that measure "state of the art" in terms of SINAD and are "recommended" by the top dog measurement extremist, yet I can tell the difference in level-matched blind tests.

    I don't expect the measurement extremists to be able to hear differences. Mainly because I see little effort or documentation on how they carry out their so-called blind tests. Also, if one already has sight of the measurement results, they are going to be convinced that there is no difference because of strong expectation bias. Studies have shown that bling factor does influence people, likewise, the measurement factor probably does the same thing. It cuts both ways.

    Going back to @OJneg's extreme subjectivist stance, I would have to correct myself somewhat. I'm more apt to go subjectivist for DACs because they all measure so well with the exception of a handful. About the only thing we can really gather from the distortion FFTs is whether a DAC employs a resistor ladder or is sign-magnitude. But really, I've never seen any measurement from any half-way decent measuring DAC that tells us how it going to sound. With speakers and transducers, it's different though because the errors are so gross.
     
    Last edited: Jul 8, 2019
    mscott58, Josh83, numbercube and 3 others like this.
  13. Senorx12562

    Senorx12562 Case of the mondays

    Friend
    Joined:
    Oct 26, 2015
    Likes Received:
    2,300
    Dislikes Received:
    53
    Trophy Points:
    113
    Location:
    Limbo
    As long as things that measure worse sound better and vice versa, I can't be sold on either extreme side of this "debate" (conundrum? dilemma?). Which of course means I have to go with what sounds best to my very imperfect ears and my own package of psychoacoustic biases, I guess. Very unsatisfying, but there it is.
     
    Last edited: Jul 8, 2019
  14. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    56,481
    Dislikes Received:
    97
    Trophy Points:
    113
    Location:
    Antarctica
    None of this stuff is state of the art, and using that definition, we can say that a well laid-out CMOY using a sub $2 opamp is a state of the art headphone amplifier. The last major leap in technology was the transistor. Heck, not even the crazy pipelines and prediction branches of modern CPUs is state of the art.

    I'm not even aware of any recent developments in amplifier circuit design other than Jason's take on John Broskie's constant transconductance output stage. Everything else is just the same blocks or variants of those glued together.
     
  15. Riotvan

    Riotvan Got lost for three weeks at Delft City Hall

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    2,338
    Dislikes Received:
    6
    Trophy Points:
    113
    Location:
    The Netherlands
    I usually go by subjective first and if something sounds off i measure to see what the fuck is going on. Music is my priority and there it's all about the subjective qualities. Keep it simple, there is nervosa on both sides.
     
    JayNYC, LetMeBeFrank, allegro and 2 others like this.
  16. OJneg

    OJneg The Most Insufferable

    Friend BWC Executed
    Joined:
    Sep 25, 2015
    Likes Received:
    3,750
    Dislikes Received:
    36
    Trophy Points:
    113
    Location:
    Grand Rapids
    You can either maintain that (1) any sonic improvement in audio equipment is due measurements of X, Y, and Z being better, or (2) that all audiophile subjective listening is moot because hoomans can't tell the difference due to, sighted bias, audible memory, etc.

    Otherwise, selling someone a 130dB SNR DAC or 0.0000001% amplifier is equivalent to selling pure cables spun by virgins or magic stabilizing rocks for your CD player. It's probably more fraudulent actually because by your own reasoning you are selling something that doesn't make any difference. I would hazard that most guys selling cables at audio shows at least believe in their product.

    Guys like bigshot were always the most consistent logically. When random objectivist would tout his awesome measuring amp or DAC, he would just reply "so what, my Oppo headphone out of my iPhone sounds just as good."
     
  17. pavi

    pavi Almost "Made"

    Contributor
    Joined:
    May 17, 2017
    Likes Received:
    216
    Dislikes Received:
    0
    Trophy Points:
    33
    Location:
    usa
    a most fascinating discussion... thank you! this is what I'm hearing:

    1. measurements matter: they provide an objective basis and filter for subsequent subjective assessment, although they may be taken further than is useful

    2. subjective assessments matter: some, though not all, listeners are able to hear differences, make distinctions, and describe them in ways that are genuinely (and dare I say objectively) helpful to others

    3. devices that measure identically may sound different to discerning ears: perhaps we aren't yet able to measure everything that matters

    4. subjective differences are further differentiated by personal preferences

    5. all this presumes ability, competence and integrity
     
    Josh83 and Lyander like this.
  18. Priidik

    Priidik Friend

    Friend BWC
    Joined:
    Sep 27, 2015
    Likes Received:
    1,498
    Dislikes Received:
    2
    Trophy Points:
    93
    Location:
    Estonia
    Good idea to make a summary.
    1. measurements are tools to assess and help navigate; just like a hammer needs training, audio measurements methods need training. A number taken by a 'monkey' with a caliper is just as useless as a random audio measurement from suspect source
    2. when one wants to be connected to the reality perception rules over logic
    3. oh, we can measure practically everything we can fantasize about (particle physics etc wizardry); it's that we haven't found out everything that matters to our perception and by how much
    4. beliefs, biases and preferences are hindering us to perceive clearly; I think it's part of the fun
    5. when first hand experience is unavailable a super best audio friend is still the best place to get good data
     
  19. tomchr

    tomchr MOT - Neurochrome

    Joined:
    Oct 28, 2016
    Likes Received:
    505
    Dislikes Received:
    19
    Trophy Points:
    93
    Location:
    Calgary, Canada
    Home Page:
    How is it ever fraudulent to provide good measurements? I don't understand your logic. If I can make an amp that provides 130 dB SNR and -130 dB THD, and I back those numbers up with measurements, why is it fraudulent of me to sell it?

    Tom
     
  20. Psalmanazar

    Psalmanazar was: dingleberry; now: chill + interesting

    Slaytanic Cliff Clavin
    Joined:
    Sep 27, 2015
    Likes Received:
    3,689
    Dislikes Received:
    120
    Trophy Points:
    113
    Oh god.

    Yeah considering the time it takes to "learn" a pair of speakers, to learn the room, and become accustomed to the conversion using well known test recordings, yeah that's total bullshit study.

    Even within the "good delta sigma" (some of these are almost 20 years old) category of converters, there are drastic differences in perceived tone and timbre across chips, brands, that the listener must become accustomed to. The chips sound different, the manufacturers have different goals and house sounds, the implementations in various products sound different and stuff with the same chips from the same company can sound different like the Steinberg UR44 and the UR824.

    The people who say this stuff all sounds the same are the same guys who say that all "clean" mic pres sound the same and that Behringer products sound competitive with everyone else's. Such BS as even the PGA2500 in everything varies from box to box. Compare the og UA Apollo to the Prism Sound USB boxes. The original Apollos used the same parts (CS4398 and NE5532) for the D->A and sounded awful in comparison.

    The numbers and age don't mean that much. Some old stuff is still state of the art. Prism Sound still has the cleanest treble I've heard and their current converter is 11 years old , uses 35 year old op amps, and a converter chip that's at least 15 years old. Perception is what matters, not the quantification of it.
     
    Last edited: Jul 8, 2019
    bixby and JayNYC like this.

Share This Page