Planar Speaker Asylum

Welcome! Need support, you got it. Or share your ideas and experiences.

Return to Planar Speaker Asylum

Message Sort: Post Order or Asylum Reverse Threaded

determining probability in a single non-blind testing

99.155.82.3
Posted on March 8, 2012 at 12:01:30
farfetched
Audiophile

Posts: 1066
Location: Cleveland!
Joined: October 13, 2010
 Yes, I posted this here on purpose, for reasons having to do with the post actually lol; I simply prefer this data set, as it were.

Listening today, I was wondering, I am "confident" that -- after really more than a decade -- I have finally locked in my system to some semblance of the "absolute sound."

I say this for this reason and this reason only:

*Before* I analyze the sound, my system consistently, on lots of different kinds of recordings I might add with an added emphasis, simply makes me almost gasp, *the same feeling I get* when I sit in Severance Hall.

Of course, not the "same" as in exact, I mean, I actually feel wow, here it is, that feeling of goosebump space, of that space, the gasp, that kind of feeling of sitting and the music is here, because it and you are here, etcetera etcetera I leave the flights to younger men...

So, call this, of course, what *you* will. I remember Hoffmann, Doug I think, ? The recording engineer dude, saying "breath of life," and I am imagining that is what I have finally seemed to have gotten....

Of course, this is a single non-blind test, and it is my own, as it were testimony.

I have seen people smitten with merely what they have often enough to understand that subjective certainty is not a necessary correlate to any intended objective scale of interest or worth. Or, to put it differently, I've met a lot of people who I am not sure some words mean what they think they mean.

But: given that certainly we are not all a bunch of blind sensoriums popping off about random non-connected events, do you think you could quantify the populational probability of any given audio report being accurate? And, what weights to attending circumstances (eg: "person is a drummer," "person is a 79 year old retired banker wondering if the gods shall let him into his Valhalla and damn he loves this new wife," etc.) can be affixed?

I see no reason why single non-blind testing cannot be quantified. Surely, by now we can determine, for example, a "Tellig Unit," simply by dint of his absolute record, as it were.


We could start by toying with the notion, what is the cap on the probability? 70% possibly accurate report?

And what *is* that which is accurate in the report? Forgoing all fine grain, what would be the simple baseline threshold of interest?

Fidelity is a sin qua non, I am imagining. I could care less if a reviewer finds he can "bop around the kitchen" or "boogie" with something. I friggin boogie to my 4 dogs dashing around my wood floored old house; but while having to do with music, that has nothing to do with the question of The System, which allegedly is the function of the reviewer. Unless one fancies oneself the audio Sun King, ala "le hifi, ce mois!" and thinks a system is somehow revealed in how a wife arches her eyebrows or waves her hands. Spare me the pasha routine, but I'll admit, that's me.




/ optimally proportioned triangles are our friends


 

Hide full thread outline!
    ...
RE: determining probability in a single non-blind testing, posted on March 8, 2012 at 15:59:49
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 You may be interested in reading Toole's book. It contains a summary of the Harman research on subjective audio quality. There is actually a lot of consistency to these supposedly "subjective" impressions. Whether expert listeners or people who have no interest in audio, rankings are similar; the main difference is that those with experience are more consistent in their rankings (they know what to listen for). The only people who can't consistently rank loudspeakers, it seems, are people with > 10 dB (IIRC) hearing loss in the midrange; their rankings are idiosyncratic and inconsistent. That's a surprisingly large percentage of the population.

I could point to limitations in the studies, but I think the general point holds: subjective impressions aren't random, and they can be quantified, as you surmised.

 

RE: determining probability in a single non-blind testing, posted on March 8, 2012 at 16:24:35
farfetched
Audiophile

Posts: 1066
Location: Cleveland!
Joined: October 13, 2010
 excellent! I know Kevin Voecks is convinced of this, and my first "serious" speakers were the Snell D's he designed.

One of the things that irks me about the Absolute Sound magazine, is its continuous (!) pursuit of new terms and ideas about what is heard. As we are claiming, yes, that is indeed possible, but it would be nice to establish some sine qua non's, no?

And these parameters can probably be intelligently enumerated.

I am fairly certain, for me, the "ingredient" had to do with finally nailing the lowest octaves. It brought with it an apparent train of finally obtained qualities, such as a fine dynamic "ease" and spring-iness, if that makes sense. I swear, sometimes I can tell a pianist, for example, is leaning into his keyboard, not ebcause of creaks, but because, well, the same way you sense body movement in general, if you close your eyes and "listen" to someone moving. Hard to explain
/ optimally proportioned triangles are our friends


 

I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 9, 2012 at 06:50:29
Mando
Audiophile

Posts: 248
Joined: September 18, 2007
 I agree with this last statement. A lot!
There is a certain sense of place and presence that seems to be provided by info below 30 hz.
I added two Rel subs (stereo) to my modded 3.6R's and all of a sudden there was all sorts of immediacy and depth and "feeling the hall".
Presence of musicians and instrument in my room.
I have the subs crossed over at 25hz (my 3.6's will measure down into the 25-30 hz range)and it is amazing what that has added to the presentation.
And yes, it is hard to quantify what it means to be enthralled.
But I know it when I hear it, and so do my audio friends!
Best regards,
Mando

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 9, 2012 at 15:26:11
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 A lot of people report that.

 

RE: determining probability in a single non-blind testing, posted on March 9, 2012 at 15:31:30
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 There are indeed lists of subjective attributes that have been used in subjective quality research. The main problem I have with such lists is that they tend not to be comprehensive, and also tend to assign an arbitrary weighting to achieve a figure of merit, e.g., 10% to imaging, 10% to timbre, etc. What we really want to measure here is how the human brain responds, and it's algorithms are a good deal more complex than that and only partially understood (although research has provided some knowledge of what qualities most people find important). The danger here is the Consumer Reports or U.S. News college rankings trap: an "objective" result that doesn't really measure what it purports to measure.

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 9, 2012 at 21:13:08
Satie
Audiophile

Posts: 5430
Joined: July 6, 2002
 Actually, that drove my friend to dump his exquisite Sonus Faber Cremona M because he had trouble matching its precision and fast response and bass tightness with a subwoofer that works the bottom octaves. He ended up with a Focal JM Lab Nova Utopia, that gives up some detail, truncates images into a soundstage in the space between the speakers and costs more than double. But it has the minifridge sized bass cabinet with a long throw 13" driver and great rigidity. The tight powerful and impactful bass is what it definitely "nails down". Very impressive. Too bad about the soundstage compression and lessened detail, truncated sustains, but it is 91 db sensitive.

The Nova Utopia definitely improves on image height and soudstage depth relative to other box speakers, but at $45k for the original and $90k for the current successor model, that should have been the minimum expected.

There is something to that fully powered bottom octave that is like a belly full of comfort food. That is often missing with the more precise sounding and fast soundstage masters. I can't say that I don't miss it on the ESLs and most planars, but it can be filled in with good sealed subs.

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 10, 2012 at 09:20:57
patmacav
Audiophile

Posts: 43
Location: Chicago
Joined: April 29, 2008
 I have a hard time understanding the definition of fast and tight when it comes to bass. Tubas are anything but fast and tight. My acoustic stand-up double bass is not fast and tight. It's only as fast as my fingers can move from note to note. And if I don't kill the sustain on the note, I can pluck my G string, lay the instrument down, walk across a small room and back and still hear and feel the sustain. So fast and tight; what do those terms mean and how are they measured?

When I added decent subwoofers, and I'm not talking great subs here, just decent, that was a game changer for my system. The lower harmonics interact with the upper harmonics and fill in what's missing from any sans sub system. Some folks don't like bass, some do. I'm one who craves and loves the bass "comfort food" in my musical diet. And my system plate would not be balanced or full without it.


Good day

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 10, 2012 at 17:53:58
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 That's a great way of describing the effect of the bottom octave. So your friend chose a speaker with lower resolution so it would better match the woofer? I suppose that would be the best tradeoff for some people, certainly abrupt tonal transitions scream "artifact." But they do seem to bother some more than others.

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 10, 2012 at 18:16:32
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 Controversial. There are those who say tight bass is merely less bass, and they may have a point, since boomy one-note bass is a chronic problem with woofers -- either induced by the driver itself, or by room interactions. This is particularly offensive when the resonances cause pitch shift. It's like having an out-of-tune organ pipe playing along with the music.

Then there's speed. Low frequencies are by definition slow, and a woofer that's literally "slow" will have a mechanical low pass filter. On the other hand, woofers can suffer from group delay. And they can and do ring. In my experience, this is surprisingly benign. Planar woofers are typically designed to ring, but it doesn't seem to be much of a problem, partly because they're designed to have distributed resonances which prevents response anomalies, partly I think because the backwave cancellation is minimum phase and so cancels some of it out, and perhaps because the modal ringing of the room is typically so much worse. As always, amplitude response distortions seem to have more of an effect than temporal ones of the magnitude normally encountered in loudspeakers.

And of course there's harmonic distortion, which in all but some servo woofers is audible when the woofers are pushed, typically around 10%.

So pick your poison. It's a vague term, not really technically correct, and so hard to pin down. "Wooly," "clean," "boomy," and "flat" or "balanced" might be better adjectives. But while frequency response anomalies and balance issues are easily identified, I don't think there's a very good correlation between the measured quantities and the subjective impression of definition. Not just in the bass, but in every frequency range. Why do planars sound more defined and natural than dynamics? Why do stats sound more defined and natural than planars? It certainly isn't the waterfalls. Harmonic distortion at modest levels? Box resonances? The nature of the self-noise, which in the typical cone is sort of a fuzz, but in ribbons or planar magnetics is either a frequency-dependent high Q metallic ringing or a tin foil/mylar sound effect?

I'm afraid I'm answering this with more questions than came in . . .

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 10, 2012 at 23:48:50
Satie
Audiophile

Posts: 5430
Joined: July 6, 2002
 Don't get me wrong, the Utopias are just one step below the ribbons and ESLs in the top decade, and the mids are among the best to ever come out of a box, but they are an MTM with steep LR4 slopes on M/T XO (the Cremonas OTOH are time coherent with low order XO) and that makes for a little too much phase shift withing the bottom treble and top midrange - that and the wide baffle (er.. a baffling design choice) eat the soundstage outside the speakers - I should add that an excelent acoustic engineer worked the placement for 2 days - so I don't think he missed any opportunity to get that extended soundstage out of the speakers. There may be a way to get creeping waves out of the way as the speaker does not appear to be designed to guide them to the back the way the Cremonas are (they shoot them to the back at about 30 degrees from axis from about 1/4 of the way to the center of the back of the speaker).

BTW the utopia bass is slot loaded but tight as any sealed sub i've heard. That long throw 13" must be some piece of engineering. The B/M XO is shallow - probably 1st order.

I can add that my impression of the Utopias was powerful enough to start rebalancing my XO to your prefered shelved down top octave and the Utopia's warm balance, I had not dared depart so much from the flat before, but the benefits are now obvious - even in organizing the depth dimension more correctly. Things are so very much easier to listen to and bass blooms nicely and the extension at the bottom octave is more pronounced with the added bass level. dropping the preamp did a little good on that count, as did a better power cord for the Marchand XO. Piano's bottom octave is also far more present and impactful.

When I first heard the Utopias I asked "so this is a legitimate freq balance?" The dealer said JV liked it enough to instantly buy a pair.

Now I am thinking how to fix this Rampal Vivaldi on CBS that seems to suffer some tonal pallor. Perhaps more volume - up 4 db, there we go... bet you a chamber orchestra doesn't actually sound like this unless you are in it (which I was for 3 years) whaddayaknow, harpsicord with gravitas.

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 11, 2012 at 00:10:30
Satie
Audiophile

Posts: 5430
Joined: July 6, 2002
 But those slow decays and powerfully slow rise times on the contrabass don't get enhanced by making them sound like they are played in a bathtub. The loose bass has the equivalent sound effect of talking into an empty coffee can. Outside of porting woofers, you get that from current starved preamps, sagging power supply voltage in power amps, and most MC cartridges and small motor turntables.

There is also "wooly bass" from undersized wiring and overtaxed tube amp trannies. Also caused by ageing power cord contacts - they should be pulled out and replugged at least anually and better yet more often than that.

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 11, 2012 at 06:37:38
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 Perhaps the shallow crossover has something to do with the good integration. The ear seems to be more tolerant of gradual changes in sonic character with respect to frequency than sudden ones. There are apparently theoretical reasons why a ported sub can't be as "tight" as a sealed one, but in many ported woofers I think the performance is limited by design choices before the intrinsic limitations of the technology come into play.

From what you say, it sounds like the Utopias have a poor DI due to the baffle size. Still, I'd expect that to affect depth more than width beyond the speakers, width seems to be affected by lateral reflections.

Since you seem to need 4-6 dB of downtilt for works performed in a large venue, and none for works performed in a small venue, I think the only real solution to the flat or not conundrum is some kind of switchable EQ. Maybe you could rig two sets of components in your PLLXO's? (The real solution, as J. Gordon Holt pointed out, is ambiance channels, since it's the HF-attenuated frequency balance of the hall reverb that causes most of the falloff in highs.

Funny, I was listening to Glen Wilson's recording of the Well-Tempered Clavier yesterday, and noticed that it had a roaring, almost piano-like bass. Was wondering if this was accomplished with EQ, because I've never heard a harpsichord that does that. (I gather Landowska's did, but it had a metal frame -- now looked down upon by the HIP purists, but IMO a sensible approach to performances in large modern halls. I remember in particular a performance in Carnegie Hall in which the harpsichord continuo couldn't be heard at all! Beats me how a silent continuo is historical.)

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 11, 2012 at 12:44:09
Satie
Audiophile

Posts: 5430
Joined: July 6, 2002
 I think the issue of reflections having anything positive to do with the soundstage width and the images in it beyond the speakers is still in question.

My experience is that it has something to do with phase and that reflections are pretty much all bad for it. I have managed to get it routinely out of 1st order slopes, that in Thiels, Vandersteens, the mod Tympani and heard it from the SF Cremona Ms, other than that, from acoustats, ML CLS, and from drivers being played without an XO, the Neo 8 being particularly spectacular at it on their own - even just set on the floor against the power amp.

The steens do it with 6+ feet from sidewalls and 5+ feet from the front. I helped our former Thiel dealer pull out the speakers from their demo spot to show him that it was not a myth.

My friend's Cremonas did not have quite that much space, but were 6'+ from the front wall and at least 4' from the side, with the engineered sandwich walls and absorbers taking the room's r60 to 0.28 sec. They did have a problem with the creeping wave discharge having missed the absorber and making it to the wall where it reflected and collapsed the soundstage expansion beyond the speakers. Once we found the reflection spot ( not hard since the beaming of the creeping waves can be felt on the hand just as it leaves the speaker's back ridge) we dampened it and the soundstage extension appeared on the right just as it was on the left.

The original instruments guys need to play in original spaces - where they can get paid original salaries since those barely seat 100 ticket buying folks. How does $20 per gig sound to you? The music was not written to be unheard with a cat gut violin being drowned out by your own breathing at row 32, not speak of anything behind it.

Original instruments and ensemble sizes should be confined to recording studios and banished from the modern concert hall along with wood harpsicords and clavicords.

I have that adjustment at hand since I have a level control for the tweeter, so increasing its level lowers the defacto XO point and shelves up the remaining freq range.

 

Excellent post Josh!, posted on March 11, 2012 at 19:03:17
CometCKO
Audiophile

Posts: 873
Joined: August 9, 2002
 The language & practice of statistics are not likely to be too helpful here at an individual level. I've done a lot of work in sensory testing of food. As with audio, a lot is known about the actual physiology of taste and we can calibrate individual sensitivity to sweet, bitter, sour and salt, which are partly a function of age (young people have more taste receptors on their tongues) and culture/kinds of foods people grow up tasting.

One has to be careful to separate discrimination from preference. As you point out, experts know more what to look for. Plus our expert tasters have all been calibrated for their relative sensitivity to taste components.

Yet when it comes to preference, the experts do not agree. Furthermore, they only broadly agree with the general public in preferences. Companies for whom I have worked have made some very expensive mistakes by relying too much on the expert tasters.

As with audio, there is also a certain amount of adaptation that occurs over time. An unfamiliar taste/food combination might be rejected upon first tasting, but as experience with it grows it might become a preferred food. Most food companies have learned that it is necessary to assess consumer experiences over an extended usage time period before making a decision of whether or not to introduce a new food product into the market. (sound familiar?)

The challenge comes when trying to "fix" a taste profile that is broken. For instance, when developing a new chocolate bar, one can systematically vary the amounts of sugar, milk, cocoa butter, chocolate liquer, and the roasting/conching to produce a wide variety of different-tasting and different-looking chocolate products. Yet some combinations are MUCH more appealing than others. Optimizing the exact "best" combination might involve testing/evaluating thousands of prototypes among a broad cross-section of consumers (usually a statistical experimental design reduces the actual number of alternatives that need to be tested). What seems to be a fairly simple problem -- making a preferred chocolate bar -- turns out to be exceptionally complicated due to interactions between the variables. And frankly, that is a very easy challenge compared to the number of variables inherent in evaluating audio.

Subjective evaluation of "goodness" of a particular system might yield some agreement across a sample of audiophiles. Yet there will be substantial outliers, people who do NOT agree with the average. The likelihood of being able to predict and assign a confidence interval to the individual preferences of a single audiophile is very, very low.

Using choice modeling, you can improve that probability, by understanding past choices, or by exposing hypothetical equipment/music combinations and assessing choice likelihoods. But statistics? Bah! Basically, you are the only one who can decide what is "good enough" and whether you have achieved an acceptable level of ecstasy (to quote Lyle Lovett).

Hope this is food for thought?


"Knowing what you don't know is, in a sense, omniscience"

 

RE: Excellent post Josh!, posted on March 12, 2012 at 09:55:09
djcxxx
Audiophile

Posts: 73
Location: South Florida
Joined: April 14, 2003
 Great post. What do you supose would be the audio equivalent of "umami"?
:)

 

RE: I am fairly certain, the "ingredient" had to do with finally nailing the lowest octaves., posted on March 13, 2012 at 19:01:38
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 Well, here's what Toole says about it in a paper: "Many audiophiles appear to put value in a soundstage that extends beyond the physical span of the loudspeakers, a phenomenon that can be created through deliberate or accidental binaural effects in recordings, or by lateral reflections from adjacent walls in the reproduction space."

In an anechoic chamber, the sound is said to extend in a narrow slit between the speakers.

A few weeks ago, I suggested to someone who couldn't get imaging beyond his speakers that he remove the absorption he had at the lateral first reflection points. He reported back that the problem was solved.

And that's pretty much all I know about it, except that there are theoretical reasons why you shouldn't be able to hear beyond the speakers without reflections or crosstalk cancellation, specifically that the interaural time delay is limited by the physical location of the speakers (the interaural crosstalk). A sufficiently loud reflection within the Haas window is known to produce image shift, which I think is why sidewall reflections can override the ITD.

(Well, almost, there's also a phenomenon called "apparent source width" or ASW that is valued by concert hall acousticians. The idea is to make the instruments come from a space that is wider than the stage. This is always achieved at the expense of lateral imaging precision and AFAIK is also done with lateral reflections.)

Don't really know about the effects of group delay on image width. I'm thinking that the improvement you're hearing may be a result not of the group delay, but of improved lateral lobing behavior in the crossover region. This would follow from the empirical observation made by Linkwitz and Toole that reflections have to be spectrally similar to the direct sound for the brain to interpret them as spatial enhancement.

I suspect that the second reflections from the oblique bounce at the corners influences lateral spread as well. I'm guessing that they're within the window.

 

RE: Excellent post Josh!, posted on March 17, 2012 at 16:34:15
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 Thanks, it is indeed. And I learned something about chocolate bars! I confess I assumed they were formulated by elves or something.

Agree that audio preferences are multivariate, however, there is one criterion that audio has that food doesn't (well, natural food anyway). Despite what some people believe, studies find that people generally prefer natural sound reproduction. And so while taste will always play a role (a loudspeaker is an imperfect device, so there are tradeoffs to be made), there's actually a *high* correlation between subjective rankings of loudspeakers made in controlled double-blind tests by various individuals with a wide range of audio expertise.

The one group that doesn't agree -- and that doesn't agree even with itself, when repeating the blind ranking test -- is people with hearing loss below 1000 Hz. That's a surprisingly large percentage of the population. Those people's choices are idiosyncratic. They can't hear what we do. (Fortunately, for those of us who are middle aged or older, age-related high frequency hearing loss doesn't have this effect, though I suppose it would if it were extreme and you couldn't hear the tweeter!)

If you're interested, I suggest you read Floyd Toole's excellent book, Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms, or some of the papers by Olive and Toole which can be found online for free. It puts to rest quite a few shibboleths, such as the assertion that people from different countries prefer a different kind of "national sound," and while there are some holes in their methodology, it produces some very useful correlations between measurement and listener preference. In particular, frequency response in *all* directions correlates very closely with preferences -- not just on-axis, but polar response as well.

 

RE: Excellent post Josh!, posted on March 17, 2012 at 16:35:20
josh358
Industry Professional

Posts: 12376
Joined: February 9, 2010
 Midbass boost, it's sort of the MSG of loudspeaker design. :-)

 

Page processed in 0.040 seconds.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2025, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]