This is indeed one way of looking at Toole’s findings. And that is also how Toole intends these findings to be taken.
However, you can also look at this by turning things around and adopt a subjectivist perspective. If indeed listeners can easily detect differences between the four speakers just by listening, what is the point of measuring them? What Toole has (also) demonstrated, against the intent, is that in the case of these four speakers measurements are redundant.
Unless we choose to opt to believe in an alternate world, ruled by different physical laws that obtain exclusively in the realm of audio gear, we must assume that measurements translate into audible characteristics.
Toole’s set-up is meant to test a particular hypothesis, and towards that aim, I assume, the four speakers were selected. That’s difficult to find fault with (leaving out the testing protocol implementation about which we learn next to nothing). But the result is also to be expected. Speakers that are deficient in one way or another should sound worse. Whether they are promoted as high-end speakers and celebrated by reviewers doesn’t matter for what interests me here.
I wish Toole had selected four speakers that are expected to perform really well according to measurements, and then see if listeners have a preference for one or the other. Or will they all sound the same to them? Should we expect that result? (By the way, Toole’s graph doesn’t tell us how many people actually preferred speakers B or M – I take it he gives us averages).
Here’s a link to an objectivist/ measurement review found on ASR. It seems that measurements can get things seriously wrong – at least in this case the supposed strong correlation between measurements and sound impression doesn’t hold up. What are we to make of this? Oh horror, does that suggest that measurements don’t tell the full story?
How on earth do you think a speaker is designed? No measurements? The measured profile will deliberately reflect all sorts of engineering compromises and marketing requirements. Toole’s point was that with the vast majority of “middle range” speakers subjective evaluation isn’t much use if the objective is for one listener to communicate their subjective impressions to another listener. Why would I care what your impressions of an item of equipment was? If your impressions only matter to you, they only matter to you. There are just too many variables/preferences//program material/room response in play.
I am aware of the ASR review of the Wilson Tot. Amir’s main conclusion was that the bass and midrange both measured an artificial “lift” and the off-axis response measured flatter than the on-axis response. In other words Wilson deliberately engineered design compromises for this “entry-level” model (in Wilson terms) that satisfied engineering/cost/marketing objectives. His opinion was that the reason for these compromises was most likely that the speaker would appeal to well-heeled customers in a less-controlled showroom environment than in the more controlled environments where it’s higher end products are sold.
Amir actually always combines measurement with “subjective” listening tests. The Wilson Tot was no exception. Amir found that “subjectively” the measured artificially exaggerated bass response was “overwhelming” in his home environment so he had to tame it with roon parametric equalization.
I have personally had a very similar experience with a pair of Buchardt A500’s. During lockdown we have been spending most of our time out of town in a small countryside cottage. I needed a small pair of active speakers with an internal DAC that were more on the scale of the Buchardt’s in this living environment compared with our main home. Because of lock-down I was unable to evaluate the speakers personally and just ordered them on-line on the basis of generally favourable “subjective” reviews. A complete disaster, the speakers out of the box sounded absolutely terrible with especially the overwhelming “showroom” bass of which Amir speaks about the Wilson’s. After a lot of mucking about with filters and room correction DSP (and roon parametric equalization) I measured a flat gently sloping in-room response which sounds great. I would have preferred if the speakers were just delivered, measuring and tuned with a flat gently sloping response rather than the “showroom” sound that seems to appeal to the type of audiophile that prefers listening to their equipment rather than to music.
Absolutely. These engineering compromises, however, also mean that we cannot predict whether the actual sound will be to our liking or not based on measurements and graphs. The measurements may indicate one thing, but the result might be surprising.
This is after all what happened with the Wilson Tot.
Here’s Amir: “If I were to just go by the measurements, the Tunetot would not get good marks.” At least he’s being honest, although one wonders how much his subjective opinion is actually worth. He mentions his “rather reflective room”, which surely is not the best environment in which to test speakers. And then, how good are his ears?
There’s the rather absurd situation with ASR where we have measurements (the scientific angle) coupled with Amir’s subjective impressions. What we get in the vast majority of cases is but the illusion of scientific rigour. Amir takes the measurements, Amir verifies/ validates through listening, and Amir draws the conclusion that his measurements predict the outcome. That’s almost a perfect example of how not to go about things. Amir is clearly biased in favour of measurements, so what else should one expect but a conclusion that validates his measurements? Rather than having Amir three times, we’d need at the very least Amir 1 doing the measuring and Amir 2 (a totally different person) listening to the product and giving his impression. That will never happen with ASR, because that would place the subjective factor right at the centre of things. Where it is, anyway, given that Amir does listen, but sort of camouflaged behind the colourful graphs and scientific lexis.
With all due respect your native language is not English? Is that correct? You have a very unusual interpretation of straightforward, clearly written English which you use to contradict writers, not just on this thread.
You even manage to misinterpret the results of the Stradivarius test with which you have opened this thread. The test clearly says that the concert violinists did not prefer the sound of the modern violins. They preferred the “playability” of the modern violins. They couldn’t distinguish the sound of the Strads vs. the modern instruments. But the modern instruments were easier to play. In other words they preferred the feature set of the modern violins.
Amir is equally clear in his reviews.
With analogue equipment like speakers his “subjective” impressions match the “objective” measured response. If he “measures” an artificial bass lift or an off-axis response that has been tweaked for customers moving around a showroom, that is exactly what he “subjectively” hears.
I personally find that sort of review useful.
With modern digital equipment on the other hand he often “objectively” measures distortion that is below the threshold of human hearing. “Subjectively”, that is also what he hears. Nothing.
I also find that sort of review useful.
Generally I find the sort of “subjectivist” reviews you get in high end audiophile magazines driven by their advertisers pretty much useless and have not read any of them in a very long time. I also find most modern internet and youtube “subjectivist” reviews equally useless as well (although I do like ASR) but I am still entertained by them and often watch them. I have very little time for new-age magical “anti-science” and I do not think that there are any “harmless” environments. That includes entirely “subjective” audio equipment evaluation that is not grounded in objective measurement.
Maybe, but not because such a correlation does not exist, but, rather, because there a general lack of a principled, quantitative understanding of the relationship between design and performance, between electrical performance and acoustic performance, and, finally, between between acoustic performance and listeners’ subjective listening experience, the splendid research of AES researchers notwithstanding.
I think that the objectivist/subjectivist divide in the music lover/audiophile community is unnecessary and leads to a debate that is, often, not particularly productive. Neither approach, by itself, offers a sufficient account of the audiophile listening experience. Schiit understand the dichotomy very well and try to straddle the divide by releasing products that appeal to both “camps.”
I am UK born, English mother tongue based in Denmark with my Danish wife. Normally, in a pre-lock down world I would be 50% based in Dublin where I spend a great deal of time arguing with postmodernist “subjectivist” social scientists.
As I said, my question had absolutely nothing to do with the way you write. Just curiosity. Thanks a lot for replying to my (admittedly) off-topic question.
Why has my post been flagged? The same goes for a post by tripleCrotchet. I don’t have any problem with his post. So why should anyone else? We may be having a rather vigorous discussion, but I don’t think either of us has been guilty of incivility.
2 Likes
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
156
Happy to see we’re finally talking about speakers, the area where subjective impressions actually matter. Though, not other people’s. And measurements matter, too.
Here’s my thinking.
Speakers couple a stereo system to a room. How well any pair of speakers work will depend to a greater or lesser degree on the room they are placed in. How any pair of speakers sound will depend on the room in which you listen to them. So it’s imperative to audition speakers in the actual space you want to use them in.
How to know which speakers to audition? Here’s where measurements can help. There are zillions of possible speakers out there. Manufacturers rarely provide much objective information about speakers other than weight and price, neither of which is necessarily correlated with the sound in a particular room – more on that in a minute. However, speakers with particular measurements may have a better chance of fitting in. So I’d look at measurements first to decide which speakers to listen to. Of course, measurements are properly done in a controlled environment like an anechoic chamber, and as @Archimago points out in his discussion of Toole’s work, nobody listens to music in an anechoic chamber. [He also lists three common methods of measuring speakers.] Interpreting the results for a different listening space takes some care – I’m not sure whether I know how to do it properly. But having the data is useful, just as knowing the gauge of a piece of wire will tell you some things about its electrical current capacity, but not everything.
Let’s talk about price for a minute. To engineer a speaker for a very specific space, you can throw away consideration of other spaces, and not have to make your speaker pan-room. This can make it much less expensive, if done properly. On the other hand, if you have to engineer your speaker to fit a wide variety of spaces, you have to support many characteristics which a particular user may never need. This tends to raise the price. So I would not be surprised to find out that expensive speakers rate well in many subjective listening tests in random rooms, just as I would not be surprised to find out that inexpensive speakers do not do as well in those tests. But that doesn’t mean that the inexpensive room-specific speaker would not do as well as the more expensive speaker when listened to in the room type it was designed for.
I have flagged you. My post has been flagged as it makes no sense in the absence of your post. To be clear, this is not a “vigorous discussion”. It is a pointless circular argument. I have made my point and will not be repeating myself. I suggest you do the same.
OK – if that’s how you see it. I have a different reading of things. You will have noticed that I also have flagged one of your previous posts – this isn’t some childish act of retaliation, but you call into question 1) my grasp of English 2) erroneously impute a misreading to me to which, as my reply has been flagged, I am now unable to reply. Nor are we going round in circles, as up to this moment you haven’t addressed any of the methodological flaws I have identified re putative correlations between measurements and sound impressions.
I guess we have different cultural standards of what a vigorous discussion involves. (So as not be flagged, let me be clear: nothing wrong with different cultural standards!)
My experience is that a modern generation of integrated active speakers with internal DAC’s and DSP have changed a lot of my presuppositions. For example, at home I have a now 10 year old pair of B&W 802D’s. These were moved from a modern house in Dublin with concrete floors and sloping ceilings (almost ideal acoustics) to a much older Victorian apartment in Copenhagen with conventional ceilings and a sprung wooden floor. A terrible bass-boom response was eventually solved “objectively” by computer measurement and convolution and the system now sounds subjectively similar to what I remember in Dublin.
Things have moved on due to lock-down. We spend much more time in a smaller country “cottage” with low ceilings and a generally more challenging acoustic environment. I use a pair of Buchardt A500’s but with suitable DSP I achieve a subjectively similar performance to the much more expensive (when weighted for inflation and also enormously larger) B&W 802D’s. It is inconceivable to me that the process of “subjectively” coupling these two pairs of speakers to very different acoustic environments could have been achieved without the assistance of highly sophisticated “objective” measuring software and a calibrated mic.
The takeaway for me has been that DSP/convolution etc. has a much more dramatic effect on overall SQ than mucking about with DAC’s, cables, linear power supplies etc. etc. Very high quality sound is now possible at much lower price points than even in the recent past. Some would probably argue that the Buchardt’s are not exactly cheap. It is all relative but I think they are cheap as they are a complete system.
Research by Toole/Harman, Fletcher-Munson etc. shows that people do have a preference for a given frequency response. This is of course for the whole system. Their published curves is of course generalized curves, not for the individual. But that’s why we choose our own loudspeakers, pickups and turn our knobs/EQ.
Given that, keeping a system simple is an easier way to get the preferred sound: Keep the number of variables down.
If the DAC, amp and any component between the source material and loudspeakers/room have a flat (neutral) frequency response, you only have to tweak the loudspeaker/room - with or without the help of DSP. A turntable setup adds variables again.
For years, “matching components” was a big deal. The pickup, tube amp and the speakers all could add up to a wonderful or awful sound (FR). With todays DACs and quality amps, those days are over - at least for me. Of course you still have to match some other specs (W/Ω for amp/speaker combo).
I’m not saying that it’s wrong to have a lot of variables going on. Much of the audiophile hobby is about controlling it all like a philharmonic conductor - from the special mastering/vinyl cut to the carpet in the listening room. But you can also do it simpler and still have a super experience of reproducing sound waves with high fidelity, aka listen to and enjoying music.
3 Likes
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
163
I completely believe that. I think the fully integrated self-correcting speaker system is the wave of the future. DAC, amp, measuring microphones, correction software – all running in the same box that produces the sound. Won’t be as much fun as messing around with separates, but easier to get great SQ.
Don’t worry. The messing about has just been moved out of the hardware domain into the software domain. You still get your audiophile “fix” .
I almost went bonkers with the Buchardt’s when I first plugged them in but after a lot of messing about in the software domain it all came together. They sound great to me now and it is a real relief not to be tempted by hardware upgraditis. It’s probably a confirmation bias of some sort but when you see a nice flat gently sloping room response in the Buchardt DSP system you “know” on some level that is the best they can sound. As my ears are old now I just lift the midrange voice a few DB’s in roon parametric equalisation as the secret “subjective” sauce. There is a learning curve but once you have experienced one of these integrated speaker systems there really is no going back. So much cheaper than old school separates when you consider they are complete systems.
That is what I actually like Sonos - they do this with an Iphone. Would love to have “higher end” speakers like with “everything” inside, and maybe connecting a UMIK-1 or another calibrated microphone to a smartphone - run around the room and… done! Audacity/Dirac is ok, but software like REW is too complicated.
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
166
Do they have built-in microphones or come with a mic? How do you “see” the curves? Are you using the Hanseong hub?
Did you go with Buchardt because you’re in Denmark?
.