C'mon guys, one thing should be obvious by now. All those "experts" don't know anything more than we do about how the "ideal directivity pattern for stereo speakers" should look like. They all just articulate their very own opinion formed by subjective listening experience in a single room. I don't know of any speaker designer that systematically compared different rooms or did a thorough investigation on the effects of early reflections in stereo reproduction.
As much as I like a lot of what he has to say about speakers in general, I think Linkwitz is just as prone as any of us here to trying something new out, deciding he likes the sound of it, (Eureka!) and then following a long and tortuous path of attempting to rationalize why it must be better, (when there really isn't any evidence one way or another) rather than admitting that he just likes the way it sounds
I get the impression from reading that article that as someone who is very firmly in favour of dipoles, it was really bothering him that not only did the Pluto sound really good despite being a monopole, but that it also sounded extremely similar to the Orion, and in some ways even surpassed it slightly.
How can this be, when he so firmly believes dipoles are better than monopoles ? In this regard Orion++ almost seems an attempt to retrofit the dipole design to make it sound more like the monopole
Apparently all is well again now though, as the dipole has once again taken the crown with the Orion++ design with rear firing tweeter Whilst most of his speaker theory on the website is grounded in reality and facts and makes a pretty good primer course on speaker design and theory, the articles specifically on the Orion and Pluto, and particularly the Orion++ one linked above are absolutely chock full of subjective opinions, unproven hypotheses, and a large dose of trying to rationalize why something that he likes the sound of is also technically correct.
I will give him credit where it is due though - unlike some "experts" that tend to get set in their beliefs after a certain number of years, he is always willing to re-examine his previous beliefs, sometimes tossing them out entirely or doing a reversal of his position, if new findings contradict previously cherished beliefs.
Well first of all I've never heard in-head localization from directional speakers - and I've listened to some very directional horn systems over the years. Floating in the air just in front of the speakers yes, never in head, nor "too near". Have you ever heard in-head localization from very directional speakers before, or are you just passing on anecdotes ? (Genuinely interested to know)
Secondly, we could just as easily draw the conclusion that listening to stereo in an anechoic chamber is not satisfying because a traditional +/- 30 degree stereo triangle cannot provide sounds from certain key angles - such as +/- 60 degrees, where a lot of the sense of "envelopment" apparently comes from.
The room reflections do a half hearted attempt at providing the missing "wider" source angles that stimulate a sense of envelopment, but they are only faking it with time delayed versions of the same signal which is arriving from +/- 30 degrees. The brain is sort of fooled, but not really.
The obvious solution is additional "wide" channels/speakers at +/- 60 degrees like those provided by Auddisey, but there aren't really any recordings that have any genuine discrete channels/content at those angles, so you end up just simulating them from the front channels anyway, effectively duplicating the role of the wall reflection electronically.
Last edited:
As much as I like a lot of what he has to say about speakers in general, I think Linkwitz is just as prone as any of us here to trying something new out, deciding he likes the sound of it, (Eureka!) and then following a long and tortuous path of attempting to rationalize why it must be better, (when there really isn't any evidence one way or another) rather than admitting that he just likes the way it sounds
I get the impression from reading that article that as someone who is very firmly in favour of dipoles, it was really bothering him that not only did the Pluto sound really good despite being a monopole, but that it also sounded extremely similar to the Orion, and in some ways even surpassed it slightly.
How can this be, when he so firmly believes dipoles are better than monopoles ? In this regard Orion++ almost seems an attempt to retrofit the dipole design to make it sound more like the monopole
Whilst most of his speaker theory on the website is grounded in reality and facts and makes a pretty good primer course on speaker design and theory, the articles specifically on the Orion and Pluto, and particularly the Orion++ one linked above are absolutely chock full of subjective opinions, unproven hypotheses, and a large dose of trying to rationalize why something that he likes the sound of is also technically correct.
I will give him credit where it is due though - unlike some "experts" that tend to get set in their beliefs after a certain number of years, he is always willing to re-examine his previous beliefs, sometimes tossing them out entirely or doing a reversal of his position, if new findings contradict previously cherished beliefs.
I agree on this
. One should be critical of his findings, but he does keep an open mind. And in the end, if he likes the sound, why bother trying to rationalize what's theoretically wrong with the speaker Well, this point was more or less a theoretical one: in-head localisation may occur in anechoic situations. Even too near localisation is probably rarely a problem in a normal room.
Nonetheless, I have experienced the stereo image sliding slightly towards me after closing my eyes, when I listened to erjee's El Cuerno. We compared it to keyser's Unbaffled Dipole a few months ago. It is a very good speaker, sharper imaging but a little less spaciousness/envelopment than the Unbaffled Dipole. It also sounded a little more 'clean' on high volume levels, but we were unsure why.
I think you already got the point
Last edited:
To me the speaker directivity is not a value in itself, but a means to achieve certain goals. I use the radiation pattern of my dipoles mainly to control the first reflections. In my room the speakers are roughly 1 m from the front and side wall, with a stereo base of 2 m. When listening in the 60° triangle I get the first horizontal reflections as shown in the picture below. Dipoles are toed in 40-45°.
I almost achieve the ideal dipole pattern with -3 dB at 45°, -6 dB at 60° and -12 dB at 75° (second picture). This leads to the attenuations depicted in the first picture. Note how the attenuation is inversely proportional to the delay: earliest horizontal reflections get most attenuated.
I almost achieve the ideal dipole pattern with -3 dB at 45°, -6 dB at 60° and -12 dB at 75° (second picture). This leads to the attenuations depicted in the first picture. Note how the attenuation is inversely proportional to the delay: earliest horizontal reflections get most attenuated.
Attachments
Quite right but we don't know what the goal is nor are we able to define it in a meaningful way. Not enough facts nor enough standardization in music production.
I doubt that you started by drawing reflection paths and then decided to build a dipole
This is what I think is "ideal" about a dipole; two nulls can be pointed at the first reflection points. The next question is if even stronger directivity would be beneficial for stereo. My opinion on this is that it has advantages and disadvantages. Sharper imaging and higher tolerance of high SPL's but more listening "through a window" and and less spaciousness.
Like Toole says, on some recordings it's nice and on some it's not. Stereo is just a little limited in its capabilities and the recordings are too diverse.
Last edited:
Perhaps not, I'm not very experienced at this loudspeaker measurement thingy.
DB Mandrake said:
Interestingly, if you think of it from an image model point of view it is the same with only the number of images making the difference. That is, the image model would say picture an infinite number of stacked rooms each with a loudspeaker in the same floor position. The boundaries between rooms are somewhat transparent only restricting energy in proportion to their absorption coefficient. At t=0 each is given an identical impulse. The impulses arrive at the listener with a transit time based on total distance and a loss proportional to the number of floor ceiling boundaries they have crossed.
If we calculate only the direct arrival plus the arrival of the image in the first room below then we have a measure of the floor bounce. If we calculate a large quantity of the stacked images we will have the standing wave response. The only difference is the measuring time window.
Gieven that boundary absorption will diminish the effect of later reflections, I'd bet you would see the standing wave response in a small number of reflections.
David
If we want to reduce reflections in line with the speaker and maximize lateral reflection then a dipole is less than ideal at the usual orrientation. However, if you take Rudolf's diagram and turn it another 30 degrees then you can totally kill the back wall bounce and strongly increase the side wall bounce. This isn't a typical orrientation but it looks like it should work very well.
Who has a pair of dipoles to try?
David
If you're relying on a high degree of rejection of the ipsilateral wall reflection due to it being at or near the 90 degree notch of the dipole, what happens for listeners who are at a different distance from the speakers ?
Surely this makes the forwards/backwards seating location very critical, due to the sharpness of the notch in the polar pattern of a dipole. Sit at just the right spot and you'll get almost complete attenuation of that first ipsilateral reflection, (maybe a good thing, if your tastes are to minimize side-wall reflection) but sit a couple of feet closer or further back and now you'll get far less attenuation and the overall presentation will change, perhaps dramatically.
Not only that, as you move further forward from the side-wall reflection "dead zone" the reflections will receive an additional 180 degree phase shift in addition to the path length time delay, while if you move back from the "dead spot" they will be in phase, meaning that being slightly ahead or behind the ideal spot will sound very different to each other.
Compare that to a more conventional well designed "directional" or CD design, where there are no sharp notches in the polar pattern, and no abrupt change or reversal in phase of the reflections.
I'm struggling to accept that a very deep and narrow notch in the polar response of a speaker is a desirable thing, especially if you are trying to increase sweet spot seating location radius, not reduce it.
The fact that there is a constant phase difference of 180 degrees at all frequencies on either side of the notch is worrisome to me too - nobody has yet provided a convincing (or any) explanation why we should not be concerned about the fact that some reflections depending on incident angle are getting an additional 180 degree phase rotation - something that all the diagrams with nice little arrows gloss over.
Maybe at high frequencies there is enough random phase decorrelation provided by the room geometry and path length that the treble reflections from the room corners will come back sufficiently scrambled in phase that it doesn't matter, but at midrange frequencies, no, there will be in and out of phase coherency issues.
As I've stated before, I don't think you can place a dipole on a continuum between "less directional" and "more directional", and extrapolate the directivity performance characteristics of a dipole to a "more directional" monopole. A dipole is a special case which must be treated as such, and can't be approximated in directivity by any monopole, equal DI figures notwithstanding.
A dipole with a DI of 4.8 dB does not behave the same way as a monopole with a DI of 4.8dB, so a monopole with a DI of say 8dB is not just the same thing further along a "more directional" linear scale...
Last edited:
I don't but let me guess, the presentation will be very wide with added "realism"
Surely not.I doubt that you started by drawing reflection paths and then decided to build a dipole
But when I went the route to a deeper notch at 90°, it was with the intention to use it in loudspeaker positioning.
I turned the dipoles 30° more inward. The "stage" widens well into the side walls. But the definition of the stage between the drivers gets worse - which was to be expected. Centered instruments double in width, partners singing in duets are dragged apart. This is much worse than the strong, but well delayed side wall reflection as shown in my diagram. My layout is not the same situation as in Toole, p.297, fig. 15.8 b, but it helps to widen the stage beyond the speakers - into a pleasant 90° panorama for many recordings. This without blurring images between the speakers to a comparable extend.
In fact I don't think that (the quality of) imaging as such has much to do with the loudspeaker directivity pattern (as such). But is has massively to do with how the directivity of the speaker illuminates the adjacent walls. By turning the dipoles from 0° (along the listening axis) to 90° (facing each other) I can dial in quite different listening experiences - in the same room and without changing anything along the walls.
Rudolf
No? That's probably the advantage of being an evil speaker manager that has people doing it for him.
Actually, I think it probably is floor bounce. Sure, it could be from the back wall or a room mode but I commonly see this kind of measurement from a speaker measurement done outdoors, if the source is placed on a stand. It's actually very common.
This is one of the things I've been most focused on, the (deep and wide) notches that commonly occur towards the upper end of the modal range. Whether they are from boundary reflection self-interference or room modes matters very little to me, because both can be mitigated with multiple sound sources, much like the multisub configuration but at a different acoustic scale.
I have and I've just given it a try. My room is only 2.5 m (8 ft) wide and highly reverberant though. The dipoles are also 2,5 m away from the back wall and I can't put them closer.
The side wall reflections are extremely early, about 2 ms. With the normal toe-in that reflection is down > 9 dB, which is definitely a good thing for a reflection that early I would guess.
I've already tried "toeing out" the speakers, but then the image collapses into two pieces and phantom sources are split.
With an 80 degree toe-in, the ipsilateral reflections still come from about 60 degrees off the rear-axis of the dipole (because the side wall is so close), so they are still relatively weak. The image is still reasonably well defined but further backwards and there is no significant broadening.
With a 90 degree toe-in, the image broadens more and sounds more spacious and diffuse, the sound really fills the room. The image is now vague and somewhat ambiguous for some instruments.
Last edited:
Wayne, haven't we alread discussed this in a similar thread or even this one?
Yes, we do seem to keep revisiting the same things over and over again on mesageboards. But I think probably some people visit a site and/or thread rarely enough that's probably OK. What's old hat to some of us is a new revelation to others.
I'm also not sure that we all are in complete agreement about what to do to mitigate the problems in the transition region, near the Schroeder frequency. Some advocate damping, others suggest multiple sound sources. Probably best if both methods are employed, although I do think damping is more of an architectural problem, in that the only thing that really works are panel absorbers. If you can't rebuild the walls or add large false wall panel absorbers, then the flanking sub or helper woofer is probably the only good solution because I don't see a "thick rug" doing much at 150Hz.
The Toole/Olive study with 4 speakers in 4 different rooms was informative in the subject. The detrimental earlier reflection(if they can be called that) didn't influence which speaker was preferred in every room negatively or positively. I think we are essentially overanalyzing this of course. They have effects..good and bad, but overall is the speaker measures smooth/flat and set up with some care, it's going to make an enjoyable noise with enjoyable material.
Oh, no offense intended Dave! I thought it sounded funny so I had to write it. Probably should have thrown a smiley on there, but the deadpan can be more effective sometime.
Someone mentioned accuracy a few pages ago. We will never be close to accurate for the original event, but we can get close to the recording. If there were one "ideal" pattern, we know what we are stuck with still.
Dan
Oh, no offense intended Dave! I thought it sounded funny so I had to write it. Probably should have thrown a smiley on there, but the deadpan can be more effective sometime.
Dan
Oh, no offense taken, but between you and Markus I'm just not feeling much love, sniff!
David
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.