There actually is some technical logic, albeit it is weak. If you look at how one models a dipole in a room simulation, it is done with two point sources, out of phase, and separated by some distance (called the dipole moment). Hence a single monopole is one source, a stereo pair of monopoles is two sources, and a stereo pair of dipoles is four sources. Coincidentally four sources is what I have found works best. The dipole pair is more restricted than the four monopole subs, but it is clearly superior to two fixed monopoles.
One feature of large and directive bass radiator in stereo system is that LF sound focuses easily to center. This could cause some image traveling on the stage depending on frequency/played note. This is why distance between bass boxes is longer than distance between front horns in KS-1804mk2. Normally this not an issue, but large+wide floor standing bass may require some offset.
Holy C**P, seems like an agreement was reached!! It shed some light over my situation also perhaps. I have the best bass/lower mids with dipoles so far. Room is 55 sqm with opening to hall way and kitchen. More like 70 sqm…..
However, I still have a 10 db peak in 35-40Hz with dipoles also as in monopoles…
Cheers,
Peter
However, I still have a 10 db peak in 35-40Hz with dipoles also as in monopoles…
Cheers,
Peter
Thank you guys for the discussion! Very educative and slightly amusing too!
I have been listening a pair of highly (and starting low) directive speakers for more than a year now. I don't know about maths, but this works in my 20m2 living room!
One thing that has bothred me in this discussion was that Earl was talking about multisubs and directivity above 700Hz - it leaves 2-3 octaves between Schroeder and 700Hz! This range really (physically) sets some challenge for achieving directivity!
Eg. kimmostos large dual cardioids don't have sharp directivity but lack backwave, still the outcome of perceived sterophonic sound in a small room obviously is good. Dipoles and large horns give sharp directivity but in totally different manner (bacwards!) and they must sound different in a room too.
ps. I have not heard Summas or KS-1804
I have been listening a pair of highly (and starting low) directive speakers for more than a year now. I don't know about maths, but this works in my 20m2 living room!
One thing that has bothred me in this discussion was that Earl was talking about multisubs and directivity above 700Hz - it leaves 2-3 octaves between Schroeder and 700Hz! This range really (physically) sets some challenge for achieving directivity!
Eg. kimmostos large dual cardioids don't have sharp directivity but lack backwave, still the outcome of perceived sterophonic sound in a small room obviously is good. Dipoles and large horns give sharp directivity but in totally different manner (bacwards!) and they must sound different in a room too.
ps. I have not heard Summas or KS-1804
Yes it is very difficult to get high directivity in this region. It is easy enough to get a smooth response from the sources so the direct field will be smooth and reflections in this region (in a small room) are not a major influence on the frequency response - except the floor and ceiling - so is it such a bad compromise to make in a "reasonable" system design? I cannot think of another region which I would so willingly compromise the performance.
It goes without saying that high directivity all the way down to the modal region would be ideal, but only if there are no other trade-offs. I have not found that to be the case.
is it more a question of the non beaming of those frequencies or the non possibility of a (too big) horn design here ?
Unity horn ? Or a planar turned at 90° because less horizontal directivity(350 to 700 with a NEO 10... just one octave, certainly impossible) ?
Unity horn ? Or a planar turned at 90° because less horizontal directivity(350 to 700 with a NEO 10... just one octave, certainly impossible) ?
I think that the merits and/or disadvantages of a unity horn are beyond our scope here and have been well covered elsewhere.
It is a question of size. Give me enough size and I can get high controlled directivity down to any frequency. But if you want to limit the size then you have to accept compromises. How one chooses those compromises is the question.
It is a question of size. Give me enough size and I can get high controlled directivity down to any frequency. But if you want to limit the size then you have to accept compromises. How one chooses those compromises is the question.
It is challenge if one is requiring DI much higher than 6 dB all the way down to Schroeder. 6-7 dB is piece of cake e.g. with hyper-cardioid WWMTMWW. That is well compatible with HF radiator DI=15...18 dB. There will be some lateral reflections for sure but in practice it sounds at least good without need of perfect room acoustics. I think (don't know for sure) one reason is that shape of DI spectrum is typical and natural - something like with small speaker, skull of homo sapiens and many music instruments - supporting different kind of recordings and genres.
Much higher DI at lower midrange will block room out and may lead to use HF radiator with very high DI (which is usually trade-off). Some people may like it no matter what and where they listen, but some degree of artificiality will be encountered.
Gentlemen, I'd like to agree, but I can't blindly believe it (to quote Dr. Geddes, it would stretch my imagination). Evidence supporting that, please ?
So far we have following facts:
1. papers by Griesinger concluding that >700 Hz D/R ratio is key for localization of (real) sources in (large) concert halls. That >700 Hz is important is beyond questioning.
2. papers by the same Griesinger concluding that < 700Hz (and even <400Hz) is easily and sharply located with broadband speech signal (something not uncommonly found in music)
3. same Griesinger concludes that LF vs. HF stereo localization/imaging is screwed up by the room (bass towards the center of the stage and sometimes stretching HF apart) and suggesting one can improve that by EQ-ing the L-R and L+R signals to bring HF and LF localization in coincidence.
What is actually missing is quantitative data about:
- what happens to localization with a broadband signal covering _both_ <700 and >700 Hz. So far, no. 3 above suggests both ranges are equally important.
- if/how much a higher DI <700Hz influences LF separation/imaging. Another Griesinger paper suggest a higher D/R ratio is better for localization(makes sense), which favors the use of directional source < 700Hz. This is especially relevant as the room gets smaller and the D/R decreases significantly (see Griesinger for proof: "Build up and decay in a hall half the size of BSH. The RT decreases to about 1.1s, from a value of 2.2s in the full size BSH model. Note as expected the onset time delay has also decreased a factor of two. But notice the D/R has decreased, from about minus 6dB to minus 8.5dB. " )
Last edited:
2. What is "NEO 10"?
1. I know one guy that have 90° turned ribbons. He was very 😱 surprised by the fact that I listen music not at one point.
And hiend speakers with 90° turned horns (H40xV90) exists..
Room is not only thing that causes LF centering and HF stretching - as mentioned earlier. Increasing LF directivity and radiator size may increase LF centering, narrow band image traveling and wide band image stretching. Decreasing LF directivity or increasing number of LF sources by dipoles or multiple monopoles hide direction information and HF radiator start to rule image perception (but may cause unequal resolution and pronounced bling-bling). This may increase wide band image stretching as well.
Because this problem is partly built in 2ch stereo system and human hearing, I prefer not to make trade-offs which favor watching of sound stage images over accuracy of pressure signal. Good image localization will follow automatically, and traveling+stretching is possible to reduce at least tolerable without destroying more important properties.
This is just my preference & policy but I don't shame to recommend it.
So you basically are saying it's better to remove/scramble localization cues at LF and leave only HF dictate localization. In this case, you could simple make LF mono (note: we're talking about the above-modal-region LF here)
I can understand your point, however this goes against Griesinger's findings on spaciousness, who argues that much of "spaciousness" is given by lateral LF energy (L-R), therefore he is proposing to boost the difference (L-R) signal at LF and increase L/R separation, rather than the opposite. I personally also prefer bass that is more "stereo", as it gives a subtle, but clear perceptional touch.
Why is that ? To me, it sounds counter intuitive. More LF directivity should mean more L/R separation (also due to higher D/R) and thus less collapse to center.
Sorry: what is bling-bling ?
I can understand your point, however this goes against Griesinger's findings on spaciousness, who argues that much of "spaciousness" is given by lateral LF energy (L-R), therefore he is proposing to boost the difference (L-R) signal at LF and increase L/R separation, rather than the opposite. I personally also prefer bass that is more "stereo", as it gives a subtle, but clear perceptional touch.
Why is that ? To me, it sounds counter intuitive. More LF directivity should mean more L/R separation (also due to higher D/R) and thus less collapse to center.
Sorry: what is bling-bling ?
Noup, I prefer pure stereo and pressure accuracy by directive bass.
Better L/R separation does not guarantee anything because ears will separate HF signal in stereo more easily than LF when listening triangle is (too) wide.
Google me
Thanks earl, for an enthusiast like me, the best order trade offs list is ceratinly one of the most difficult thing to understand. For sure we don't want two huge big totem in our living room... at least most of us I surmise !
Is a flat phase line from a speaker in free fields (without room) is equal to the time presentation than a wide bench of instruments can makes in real life ? In another words : is a speaker & more a standalone driver between its lower & higher frequency range is not out of phase by essence ? Does any passive or active design can resolve this and is a flat phase line result from this work ?
This is what I'm looking for to conceptualy understand before any reflexion about rooms and directivity changed by design! This thread is very hard to follow, many arguments of many fellows seems good to the brain !
Is a flat phase line from a speaker in free fields (without room) is equal to the time presentation than a wide bench of instruments can makes in real life ? In another words : is a speaker & more a standalone driver between its lower & higher frequency range is not out of phase by essence ? Does any passive or active design can resolve this and is a flat phase line result from this work ?
This is what I'm looking for to conceptualy understand before any reflexion about rooms and directivity changed by design! This thread is very hard to follow, many arguments of many fellows seems good to the brain !
Except that no. 1 clearly states that is not the case. All Griesinger says is that <700 Hz is a factor. He clearly states that >700 Hz is dominant. Your interpretation of no. 3 goes completely against no. 1. I do not think Griesinger wrote these papers as a contradiction - I think that it is your interpretation that is in error.
It sounds like you are asking if linear phase is a good thing. Yes, for the most part it is, but more important is a smooth phase response which will always result if one has a smooth frequency response (minimum phase system assumed, which most are.) As Griesinger points out in no. 1 (above) phase coherence at frequencies > 700 Hz is critical for good imaging.
Here's some food for thought: by using a driver which spans the 1-4khz range you are increasing intermodulation distortion through 1-4KHz.
That said, I still consider a 6" woofer, 2" dome mid, 1" tweeter crossed at 800hz and 4KHz, to be a better configuration than a 6" woofer and 1" tweeter crossed at 2KHz. Generally power response is better, harmonic distortion is lower, frequency response is smoother and power handling is higher.
That said, I still consider a 6" woofer, 2" dome mid, 1" tweeter crossed at 800hz and 4KHz, to be a better configuration than a 6" woofer and 1" tweeter crossed at 2KHz. Generally power response is better, harmonic distortion is lower, frequency response is smoother and power handling is higher.
Using a single driver over 1-4KHz will increase intermodulation distortion between frequencies within that range.
If you had a 2-way speaker where both the woofer and tweeter had excellent performance from 1-4khz and you were free to crossover anywhere in that range, crossing somewhere in the middle of that range would produce the lowest intermodulation distortion.
That however is not always the case in 2-way setups. Rarely do you have the freedom to choose a crossover point over a 3khz range. The woofer probably has rising harmonic distortion above 4khz and the tweeter rising harmonic distortion below 1khz. So you choose to cross somewhere around 2-3khz to minimise harmonic distortion.
If you add a midrange driver which has very low distortion from 1-4khz (i.e. a 2" dome midrange driver) the intermodulation distortion will probably be lower than the woofer or tweeter would produce anyway. Therefore 3-way is generally a better setup.
Last edited:
This is just 2octaves. Most drivers will be operating over 4 to 6octaves.
Why should a driver covering only 2octaves have more intermodulation distortion than a driver operating over 5octaves?
Is the range 1kHz to 4kHz much more sensitive to intermodulation distortion?
What happens when a driver that is operated from 1kHz to 20kHz? Is that less susceptible to intermodulation distortion.
What if that driver were HiPass rolled off at 2kHz, would it's operation over the range 2kHz to 20kHz have less, or more intermodulation than if it were rolled off from the other end, i.e. no treble and now operating from 2kHz to 8kHz, still 2octaves.
- Status
- Not open for further replies.