The interaction with the room is certainly different but I'm not sure if it's generally better. As soon as wavelengths reach room dimensions a dipole stops working as a dipole. The behavior is probably better described as two "out of phase" monopoles with virtually no distance between them. Is such a configuration generally better at delivering low frequency large room behavior that might be preserved in certain recordings?
Last edited:
Two opposite phase monopoles with infinitesimally small distance between them is called ideal dipole, by the way.
- Elias
This is too much generalization for me.
The rear wave is as good as any other wave direction. What we really don't want is a few specific reflections: the ones reaching the ears too early. And in those reflections we only need to eliminate those frequencies which give wrong clues. The reflection points and frequencies in question are easily detectable. It is probably less than 5 % of the dipole energy reflected from the front wall which can be called harmfull at all.
Even with the rear wave included a true dipole will radiate much less energy than the monopole.
Which allows us - and should encourage us - to eliminate only those small portions of energy in selected frequency bands which really do give bad spatial clues.
To come back to the LX521: SL has to consider that many builders of a LX521 will not have the ability or means to find and eliminate bad front wall reflections. For those people it might make sense to omit the rear tweeter completely.
Rudolf
This is too much generalization for me.
The rear wave is as good as any other wave direction. What we really don't want is a few specific reflections: the ones reaching the ears too early
The research of Bech indicates otherwise. Strong reflections < 5 ms are harmful indeed. But reflections from the front wall, floor and ceiling can cause audible colourations because of comb filtering. For lateral reflections, the IACC is lowered compared to frontal reflections (there is a larger difference between the 'comb' at both ears) and the comb filtering is more easily 'filtered out'. Lateral reflections are less likely to cause colouration and more likely to add spaciousness.
This is certainly true but I would not suggest that for a typical dipole woofer (or U-frame cardioid) that the distance is virtually zero. I've done a number of simulation and measurements that clearly show that with the 18" separation between front and rear of my Note dipole woofers the response of each source is. One in the modal region, dipole or otherwise, woofers behave as multiple sources with varying phase (and perhaps amplitude). Additionally, it is often said that dipole are better because they don't excite modes perpendicular to the dipole axis (provided the dipole axis is aligned with a room axis). However, with a stereo pair of woofers it should be realized that the dipole can excite the same front to back axial modes as a monopole pair and in the cross room direction, with symmetric placement, the monopoles will excite odd order modes to the same degree but with opposite phase. Additionally careful placement can also prevent excitement of at least the 2nd order mode in the cross room direction.
The most simplistic way to look at woofer systems is to recognize that in the modal region the response at a listening position depends on 1) the near filed response of the source(s), the position of the listener and source(s), and the room transfer function between each source and the listener. Free field response means nothing.
When I put the Note II together I initially omitted the rear tweeter because I could not really hear a difference with or without it. I only decided to keep it at the last minute because if it should be a problem in some rooms it's easier to turn it off than to add it later. My tweeters come in about 6k Hz.
I believe your findings are exactly in line with my personal experience and what Toole found (fig. 5.2):
Image shift and source broadening is nothing you repair at 6 kHz plus.
My dipole tweeters come in at 2 kHz though.
Rudolf
A friend of mine had tried to make the reverberation time of his room quite low and as frequency independent as possible. On the way to that goal he had to get rid of many reflections in his room. When he had achieved his goal, the first floor reflection stood out more colouring than ever before.
Why? Because he had lost all the other (later) comb filters from the side and rear walls, ceiling etc., which would have (to a certain degree) compensated the influence of the floor reflection.
There should be only two direct front wall reflections from a dipole (one at probably ~1/3 of the front width and one from the room corner). Absorb those above 1 kHz and let the rest of the wall in peace.
I believe your findings are exactly in line with my personal experience and what Toole found (fig. 5.2):
Image shift and source broadening is nothing you repair at 6 kHz plus.
My dipole tweeters come in at 2 kHz though.
Rudolf
The tweeters on the NaO II come in at 2.2k. Turning the rear tweeter on and off makes a significant difference. They actually sound rather dull without the rear tweeter, except on some rock recordings. With the Note II the difference is insginificant at best, in my room. FWIF, my room's RT60 is about 350 msec from 250 to 1k Hz. 300 msec at 2k, 288 at 4k and 250 at 8k.
I appreciate Toole's knowledge and experience, but David Griesinger of Lexicon suggests that envelopment actually goes all the way down in frequency. From about 1kHZ on down you've got inter-aural crosstalk as well as the room modes damaging the stereo effect and image location (apparent separation). The question might be what to do about that. One of the Griesinger papers shows a circuit that cancels L+R more and more as you go down in frequency below 800HZ (or 300HZ - it's switchable). I'm using that (frequency selective L-XR), the inter-aural crosstalk cancellation method (modified Carver for my main system and Polk method on my stereo soundbar), and aiming my woofers toward the side walls rather than directly toward the sweetspot.
Back in about 1985 I noticed that the top of the line Snell speaker (not OB) had a rear firing tweeter, that according to their literature only operated above about 10kHZ (if I remember correctly). I suspect that they tried several cutoff frequencies before settling on that one. This was a speaker that was calibrated to be flat +/- 1dB. With what I know today, 1kHZ - 6kHZ is all about amplitude comparisons for image location, and above 6kHZ horizontal image location becomes rather vague. Above 6kHZ is apprently responsible for up/down imaging cues, due to the shape of the ear and ear canal, 8kHZ allegely being a key frequency for that. Perhaps Snell wanted to put air in the high treble (cymbals and such) without messing with uppermid imaging (1kHZ - 6kHZ) or the up/down imaging around 8kHZ (?).
One of the people on this forum made the comment that the 1 inch domes coming in at 1.4kHZ in the Orion speaker were tedious to listen to over time. He seemed to blame it on the 1 inch dome. I figure it was because of the off axis response having a jump when the 8 inch driver got directional at 1.4kHZ. This issue would perhaps be doubled by the presence of the same thing happening on the back side of the Orion speaker. Going from an 8 inch to a 1 inch is never going to be optimal.
In my OB system I used a vertical line array of four 5 inch drivers instead of the 8 inch, to minimize this issue (and others). In the Orion, the off axis response is still a big issue even if you're directly in front of the speakers, since the room will react to it, and color the final perception. I think the Seas millenium 1 inch dome is very good sounding down to 1.4 kHZ (w. 4th order active crossover). I think the problem was that it picked up where an 8 inch left off.
One of the people on this forum made the comment that the 1 inch domes coming in at 1.4kHZ in the Orion speaker were tedious to listen to over time. He seemed to blame it on the 1 inch dome. I figure it was because of the off axis response having a jump when the 8 inch driver got directional at 1.4kHZ. This issue would perhaps be doubled by the presence of the same thing happening on the back side of the Orion speaker. Going from an 8 inch to a 1 inch is never going to be optimal.
In my OB system I used a vertical line array of four 5 inch drivers instead of the 8 inch, to minimize this issue (and others). In the Orion, the off axis response is still a big issue even if you're directly in front of the speakers, since the room will react to it, and color the final perception. I think the Seas millenium 1 inch dome is very good sounding down to 1.4 kHZ (w. 4th order active crossover). I think the problem was that it picked up where an 8 inch left off.
that's right, SL choose the 8" to prioritize the low mids register and avoid using two 5" or 6".
Note 1: Since these tests were done Seas introduced the W22EX001. Like the W18EX001 it has very low distortion, but with 1.75 times the cone area it can move more air. This driver is my first choice for any new open baffle speaker design
The LX521 solves both issues, directivity of the 8" and power of the 1", for the ones who were not totally happy with it.
Note 1: Since these tests were done Seas introduced the W22EX001. Like the W18EX001 it has very low distortion, but with 1.75 times the cone area it can move more air. This driver is my first choice for any new open baffle speaker design
The LX521 solves both issues, directivity of the 8" and power of the 1", for the ones who were not totally happy with it.
The 8" is not the primary issue. It is the tweeter blooming that causes extra off-axis SPL and might become audible depending on the room and placement.
If the 8" would have been the main issue then he would have avoided it in the LX. But since the issue were the tweeters, it is solved by the 4" on a narrow baffle, which avoids blooming at the cost of an additional driver.
If the 8" would have been the main issue then he would have avoided it in the LX. But since the issue were the tweeters, it is solved by the 4" on a narrow baffle, which avoids blooming at the cost of an additional driver.
Last edited:
I'm not sure what you mean by "blooming"? What does that mean technically? If you mean that it causes a sudden increase in dispersion at 1.4kHZ, then we're saying the same thing (?).