Ahh, thanks Tim. Now I do actually understand.
Thanks greg, sorry if I (unwittingly) pissed you off.
It’s All a Preference
- Thread starter mep
- Start date
Ahh, thanks Tim. Now I do actually understand.
Thanks greg, sorry if I (unwittingly) pissed you off.
Hello Soundminded
I am surprised you can’t find anything?? Lot’s of flavors out there
. Which really is the root of this discussion. I own several different speaker types from small monitors 6 ½ monitors to full active set-ups with pro drivers and horns. I am partial to modern horns and waveguides as my preference but I enjoy all of them. They all have different strengths and weakneses. So there is actually quite a bit out there I like. The only similarity between them all is relatively flat on axis response which I have always preferred.Surprised you EQ everything. I am the other extreme I quess. I don’t even have tone controls in my main 2 channel rig. Different strokes so they say. I use EQ in my HT set-up but sparingly as a last resort. It’s set and forget EQ that never gets changed. House curve kinda thing using cut only Urei 539’s 1/3 octave analog EQ’s. What do you use for EQ?? You use parametric like a Urei Little Dipper of the more traditional 1/10 or 1/3 graphics?? Analog or Digital??
Rob
Tim,
Sorry , in this forum subjectivists are written subjectivists. As far as I know this forum has no voice facilities.
I love the way you now endorse F. Toole methods and science but did not accept the basics of his formulation of sound reproduction sometime ago. Sounds like the guy who does not appreciate the food, but loves the cutlery.
As a general remark to this thread, may be I should remember that Amir is patiently presenting Part 2 of the book but did not go through Part 1 Understanding the Principles. I doubt that any one can fully appreciate his writings without knowing F. Toole views expressed in the first part. Unhappily people seem more interested in finding ammunition for their camp than really debating. I quote Laurence Dickie (formerly from B&W, now with Vivid Audio) from the link Amir supplied about loudspeaker design: http://www.soundstagehifi.com/index.php?option=com_content&view=article&id=228:high-performance-loudspeakers-how-theyre-designed&catid=62:monthly-column&Itemid=3
We recognize that ours is certainly not the only approach to achieving these goals, and indeed, there will be circumstances and environments in which others will give better results -- and we never lose sight of the fact that there are probably as many correct speaker designs as there are pairs of ears.
I hope that everyone sees that the last sentence in a hyperbolic assumption, not to be taken literally. IMHO, we often forget it.
He also says this before he gets there: " horizontal off-axis responses are to be smooth and monotonic, "
That is what we are discussing here. The issue with other schools of thought is that we lack listening tests to prove efficacy. They may be good indeed. But without data, their case is less defensible.
F. Toole is neither the author of double-blind listening testing, nor the only person who believes smooth, even FR on and off axis is a good characteristic for a loudspeaker (and that is the subject of the thread, the findings of the Harman research and what I agree with), but jog my memory; what is Toole's basic formulation of sound reproduction that I rejected?
Tim
Wide uniform high frequency dispersion was Edgar Villchur's and Roy Allison's idea too. That's why Edgar Villchur invented the dome tweeter which made its first appearance in AR2a and shortly after in AR3 over 50 years ago. Dispersion was further improved by Allison at AR with AR3a, then AR LST and at his own company with his own speakers. His tweeters were unique in the way they extended angular propagation to nearly 90 degrees off axis by driving them partway towards the center and incorporating a flexible section that allowed the outer section to vibrate laterally even with an axial drive motion. It is likely the single widest dispersion tweeter ever and the widest I know of.
Amir, what does monotonic off axis response mean?
Monotonic is a generic technical term meaning that if the source increases, the output also increases. It doesn't have to be proportional but if the source goes up, the output must too. Here are examples of two monotonic responses:
This one is not:
So what we like to see here is an off-axis which follows on-axis in a monotonic manner. When it all of a sudden has a dip before one driver picks up the load from the larger one, then we lose monotonicity.
Perceptually, we can tolerate some amount of this. But the wider the dip, the more audible it can become. We can see these variations in the previous graphs posted:
The thing that shows this is at the bottom of the graphs which shows the deviation from on-axis response. You want a smooth response there and not broad variations that are not monotonic.
Amir, I'm still having a problem understanding this. Are you talking about ITDG and the Haas effect? There will always be a delay between the arrival of the direct sound and the first reflection depending on the added travel distance and the speed of sound.
Amir,
For me, the main question is that your formulation is exclusive - if a speaker does not fulfill this criteria you immediately consider that it should sound inferior. I am not yet persuaded of it.
I am not an expert, but I can easily accept that there are other aspects in loudspeaker design that are more important than these ones for sound quality or even that this simple model is not the best or unique.
I have said if before - IMHO sound reproduction as an whole is more an engineering business than a science debate. One can win the science debate due to lack of adversaries, but this does not assure supremacy.
You mean with respect to monotonacity? If so, no, it has nothing to do with Haas effect. Here is the deal:
What we hear = direct sound + indirect sound of the speaker
If you change the frequency response of the indirect sound of the speaker, it will also change the timbre of what we hear total. It has no choice but to do that.
The Haas effect tells us that we don't hear these reflections as distinct echos. It doesn't get into the explanation of the point here.
Go find a pair of Martin Logan hybrids. put on some music, I don't even care if you're totally unfamiliar with it. Sit down in front of the speakers and listen for a minute until you get a feel for the tonality. Now stand up, slowly, paying attention to that tonality as you do. Depending on which ML hybrids you're listening to and how their tilt is set, that may do it by itself. If not, now shuffle slowly to one side, in a gentle arc that takes you closer to the front of the room, to the side of one speaker as you change position, still facing the speakers. Pay attention to what happens to the tonality, not the stereo balance, but tonality, as you move off axis.
You've just demonstrated to yourself how completely different the amplitude response of the first reflections will be compared to the direct sound of the speakers when it comes back and mixes with that direct sound a split second later, regardless of the material they reflect off of. Think, for a moment, about the effect that will have on the overall tonallity of the playback. If you remain unconvinced of the importance of smooth, even off-axis response, I'm afraid you'll stay there. But consider this:
The speaker manufacturer cannot anticipate what material the off-axis sound will be reflecting off of. He cannot anticipate how far away that material will be or when those first reflections will mix with the direct sound. All he can do is minimize the distortion that comes out of his product, or minimize the volume (maximizing the distortion) of the distorted, off-axis sound relative to the more accurate, on-axis sound, and call it "controlled directivity." I think minimizing the distortion, and the amplitued change, is the better strategy and the listeners in the Harman test agreed.
If you performed the listening test I just described, I think you would come over into the light where speakers off-axis sound very much like they do out front.
By the way, I'm just picking on the MLs because they are the example in this thread and are particularly bad in this regard. I also find it fairly disappointing that they are dipoles, that are meant to use reflected sound by design, and have this problem. But the bottom line is you can find this problem in lots of box speakers and most horns. Klipsch, at all price levels have it in spades.
Tim
Last edited:
First, let's be clear that this is not my formulation
. I am simply reporting what Harman research shows. And personal experience sitting through a bit of the testing they have done. The research shows that we can highly predict what the listener prefers. It doesn't mean we can guarantee it but highly predict. Therefore, going against the grain there is really betting on weak odds.But don't they have to show that other than saying they use Kevlar of this, and foobar for that? Do we just accept technical language just because it sounds right? Take "time alignment" of drivers. I asked Harman folks why they don't stagger their drivers to get them to align this way. Answer: they used to believe in that. But then came one of the top researchers in this field and built a black box and simulated that time differential and it simply was not audible. So they no longer believe. Who would you say is right? The guy who uses our common sense that drivers need aligning or the test of the hypothesis?
Let me give you another example. I was recently reading about the reference lab that Fraunhofer Institute (German government funded audio research group that brought us MP3) was building. They measure and find the floor reflections being strong. So they go and put a thicker rug there and the bounce goes way in the measurements. Then they go on and say that based on listening tests, that had a negative effect! As Dr. Toole is fond of saying, "two ears and a brain" are very different than a mic and a measurement graph. At frequencies above transition (200 to 400 Hz), psychoacostics rules because the two ears hear different things and the brain has to interpret what that data means. As you, I am not a speaker researcher but I am a good student
. And when I read examples like this, I learn to not go by common sense anymore. I want listening test data to make sure psychoacoustics is not forgotten.The problem of course with psychoacoustics is that it is not always 100% precise. But if it converges well to a nice majority, we do have useful data.
I say it is the other way around: real science does exist but unfortunately does not translate into commercial success. Folks have a great gig here. No one is told what the reference really is (i.e. what was being recorded). So just about anything can pass as the highest fidelity. It is only when they are put side by side that we can see deferentially which is closer.
All of this said, yes, in the heat of the argument all of this seems to come across as more definite as it should. We would not bet our lives on such things. But in absence of any compass to show us the north, a magnetized paper clip on a leaf sitting on a pond, is better than wondering around without it
.
Then why on earth build your product line around dipole technology that deliberately utilizes indirect reflections as a feature? Even MLs "entry level" floor standers are dipole, though they sound significantly better off axis than the panels. These speakers are made to create reflections, therefore, by your reckoning they time-smear by design.
Tim
Last edited:
Tim,
Do you remeber?
(I was quoting F. Toole, as I said in next post)
All the listening tests carried by Harman rely ultimately on the preference and pleasantness of the listeners. They find happy when their trained listeners agree with the preference and pleasantness people, not the opposite! They believe that our natural preferences go towards good sound reproduction.
"If you change the frequency response of the indirect sound of the speaker, it will also change the timbre of what we hear total. It has no choice but to do that."
Agreed. Therefore the earliest reflections including lateral reflections that are so desirable should arrive at the listener with flat frequency response and the amplitude should have a linear relationship with the direct sound. Since not only is the off axis response at a given angle for nearly all speakers a function of frequency and all surfaces frequency selective in their reflectivity, unless provisions are engineered into the speaker to compensate for these facts the sound the listener hears will not meet those criteria. What provisions has JBL/Revel/Harman engineered into their speakers other than wide tweeter dispersion to accomplish that? Is their dispersion as wide as say AR3a or LST? AR3a is down 5 db at 15 khz 60 degrees off axis. LST has four of these tweeters, the two outermost at 45 degrees to the front panel increasing lateral dispersion.
I do, and don't feel a conflict between preference and accuracy because I've usually found that I prefer the more accurate reproduction.
And now I know I'm not alone. You keep repeating that Toole quote as if it reveals something. It does not; it is obvious. Of course we must like what we listen to and if what someone likes is valves and vinyl and speakers with very bad off-axis response (and not great on-axis) that's fine with me. That is your preference, enjoy it. I've never said anything else. But if you can't accept that it is just your preference, if you insist on making up weak synonyms for "accuracy" and insisting that your distortions are more natural than a lack of distortion - and this happens in the audiophile community all the time - I will insist that they do not sound more natural to me, nor evidently to the statistical majority when listening blind at Harman laboratories.
And then, of course, we'll enter into the inevitable "but but but but but......"
Tim
If Harman has an inside track to what "the market" likes then why doesn't it have an array of speakers at all price ranges that capture much of whatever market there is the way AR did in the 1960s? It certainly has the distribution channels and brand name recognition. Not only are most speakers on the market for audiophiles as far as I can tell narrow high frequency dispersion beamers, even most audiophiles don't follow Floyd Toole's advice about needing 4 subwoofers. Now there's a real contribution to the science that seems on solid ground yet few audiophiles take his advice. Why is that?
Similar threads
