Jump to content

The goals for an "ideal speaker..."


Recommended Posts

"...You've long since lost #1, Howard, and now attempt to bolster the significance of #2 according to your own prescription. I'm not buying it, and suggest that Toole & Olive have a far better handle on this than you could ever hope to achieve.... "


I've extracted a statement you made from another thread that was recently ended because it got off topic. I think it is very interesting that you made that statement, and what it tells me is that you appear to have lost confidence in your position. I think you make some valid points and reasonable arguments at times; but when the going gets tough, you tend to resort to personal attacks and negative comments about Howard's arguments or rationale, and you frequently take cheap shots against his position. That might be all well and good in "The Kitchen," I suppose, but as I look back over Howard's comments in his posts, I see a more forthright, rational approach to the debate than what I've seen from you. He may or may not be correct, but he does seem a bit smarter than you.


• Identify yourself to the group

• End the personal attacks

I think if you did these two things alone, this entire group would put more credibility in what you have to say.

--Tom Tyson

Link to comment
Share on other sites

I think it is very interesting that you made that statement, and what it tells me is that you appear to have lost confidence in your position.

Lost confidence because I steadfastly object to such a contrivance being imposed as definitive over the objective data? It's classic subjectivist mantra, and no diehard empiricist such as myself would ever submit. It is simply irrelevant to the subject at hand; we are discussing speakers here, not ourselves.... ;)

Link to comment
Share on other sites

Wow, I've never seen so many misquotes of Toole's positions in one letter. I thought you read his book? Let me tackle a few of them

At the very least, dozens and probably hundreds of axial measurements would be required, and even Toole did not try that approach. Basically, a few measurements are made and then the so-called workable data is used to extrapolate how a speaker would (or should I instead say "should") perform in any given space.

Harman currently uses 70 curves for the hemispherical and full space averages. You could probably make do with less as the average forms very quickly when you start using curves from far off axis. As for needing dozens of curve, aren't you reducing your curves down to a single curve, an average of 5? Who has more information about the speaker, someone with 5 direct field curves around a typical listening window or someone with a single spatially averaged curve?

Ironically, this hearkens back to Villchur's (and Allison's) individual driver curves, and yet now even those men would now admit that there is more to speaker performance than that. (After all, Allison went on to do reverberant-field curves when he did is analytical work.) But at least single-plane isolated driver curves, because of the circular design of most drivers, give us a 360-degree view of their outputs out to whatever final angle is chosen. This does not work at all with fully functional speaker systems, because you cannot just do a few one-meter readings and expect to not have problems with missing the consequences of the driver interactions and crossover artifacts. Toole fills his book with direct-field, limited-plot diagrams, and then expects us to believe that such sketchy data will tell us about how speakers sound in real-world listening rooms.

Again, 70 curves from every 10 degrees in a horizontal and vertical circumference. Regarding "how speakers sound in real-world listening rooms" you once again confuse reverberent field steady state measurements with perceived balance. We don't need to have that argument again since Roy has deemed that the ear responds to a 30ms time window for perception of balance. As such the live room measurement absorbs a lot of extraneous info and obscures the true response variations that occur at various listening angles.

So, you have to do a LOT of full-system measurements if you are dealing with just the direct field, and if you do not do that you have learned practically nothing.
A few measurements can sample the likely listening window and reveal the trends quite nicely. On axis, +- 10 degrees vertically and +- 20 degrees laterally give a good starting point.
There is also more to speaker performance than the systems themselves in isolation from room artifacts. For Toole (and I suppose you, too, since you lionize the guy) a pair of speakers should behave as a big pair of headphones, delivering sound directly to the ear with minimal room artifacts.
Reread Toole. He is actually a bigger fan of wider dispersion than you would give him credit for. Try the 1985 paper Loudspeaker Measurements: Part 2: "This form of (multi)directional sound output will energize the reverberant sound field but it will also, especially in the listening room used here, produce very energetic early reflections from the room boundaries. It is believed that these are responsible for the enhanced sense of space, while maintaining adequate stereo image quality." (p343) A whole chapter in his book deals with room reflections and finds them generally beneficial. An early study found the Quad ESL63 downgraded in listening tests, especially monophonic tests, for sounding too dry.
Here is the interesting part. Since Toole does not think the room should be involved with speaker evaluating, he goes one step further and determines that the best speaker will be one that excites as few room artifacts as possible. The room is bad, so we should measure speakers outside of room effects, and this leads to the conclusion that the best speakers are those that excite as few room artifacts as possible. The tail essentially is promoted to wagging the dog.
Where did you pick up that Toole isn't interested in the interactions between a system and the room. Recent papers have used direct field response measured at various angles and room characteristics to predict an accurate in-room curve.
One proof of this is that in his book Toole never did do any measurement (or even listening-comparison work) with good super-wide-dispersing loudspeakers. He basically dismissed them (based, almost for sure, on experiences with inferior models, since he never mentions any of the good ones, or even alludes to them), just as he dismissed the Villchur LvR concerts, because their existence undermined virtually all of his major premises. Admittedly, he did use a preliminary drawing of the basic Allison One model on page 194 (taken from an Allison paper on boundary cancellation), but then went on to note that while it was a "thoughtful design," went on to note that they "have no comprehensive measurement data on it." That was the only reference to a wide-dispersion approach (although in that section the discussion involved boundary-cancellation artifacts in the midbass) with main-channel speakers, with his only other mention of the concept involving surround speaker performance. Hey, he at least "alluded" to the Allison Model One!
If you insist on tossing about the term "super-wide-dispersing" loudspeakers, you need to define what dispersion you are talking about and convince us that it is "super" in comparison to the mediocre supposed dispersion of other speakers. Show us a d.i. curve of the Allison products. There are plenty of other systems with drivers of a similar size to the Allison products. I would take some convincing that they have sigificantly different dispersion than other similarly sized drivers. (Not that it matters when you mount them on divergent baffles. A pair of directional horns mounted 90 degrees apart would disperse sound quite similarly, perhaps with less lobing due to overlap?) There are plenty of 3-way systems using dome mids and tweeters that would have dispersion essentially the same as an AR3a. They are certainly not in a class apart.
Toole actually offers up a photo of his own listening room (in the year 2003) on page 29, and it is easy to see why first-arrival, direct-field signals are so important to him. The area is so spacious (with a huge open area to the right) that any wide-off-axis signals would be almost lost in space. (No shorter than 30 ms reflections in that room!) No wonder he favors speakers with a strong and tight directional behavior.
So a cluttered sound room is no good. Similarly a spacious room is useless...
Well, at least Toole did show us a picture of his listening room, which helps to clue us in on his belief system. Some of his early Canadian research facility articles also contained photos, and in at least one of them the primary comparison room was set up almost in an LEDE fashion, which would again put super-wide dispersing speakers at a disadvantage. No wonder he thought (on page 340) that the AR-3 sounded "dull." It was performing in a borderline padded cell.
...and it couldn't be due to the considerably rolled off HF response due to an early, heavy phenolic dome tweeter? I've been to the NRC facilities and wouldn't call them padded cells by a long shot.
Link to comment
Share on other sites


This topic is now archived and is closed to further replies.

This topic is now closed to further replies.
  • Create New...