I've had a look at some of JBL's products and I can't help but agree with Dave. The directivity plots look very respectable. Earl, maybe you should try diffraction slots? 
EDIT: I had opened the previous page of this thread several hours ago, and before posting hadn't seen the last seven or so posts. So maybe I'm a bit late in the day. I'll go and read them now!
EDIT: I had opened the previous page of this thread several hours ago, and before posting hadn't seen the last seven or so posts. So maybe I'm a bit late in the day. I'll go and read them now!
Last edited:
With all due respect I don't think you understood my point. As you move at different angles toward or away from the speaker axis, the woofer sound field remains constant, the reflected sound field remains constant, but the direct tweeter produced field varies considerably. This is true at the highest frequencies for a conventional tweeter but is true at all frequencies within the range of a cd tweeter.
I had asked a question about why the gedlee speaker must be toed inward and this appears to be in part the answer. As you move off left or right, you are moving away from a speaker in distance but toward an angle where the tweeter output increases across its entire bandwidth. I pointed out that the optimal angle for these two phenomena to compensate each other precicely depends on how far away you are from the speakers.
You've got a significant error in concept here- woofers become directional with increasing frequency, with directivity at the XO (in a speaker like the Summa or old 4430) matching that of the horn.
The loudness of the woofers including both their direct and reverberant fields will not change much as you move laterally from left to right at the same distance but at different angles to them. The reflected sound level (reverberant field) level of the tweeter will not change by much either. But the direct sound field level will. That is what the graphs show. It will change as a shelved response getting louder as you move closer to the axis, softer as you move away from it. The relative FR of the tweeter will not change but the absolute level compared to the woofer will. By toeing the speakers inward, as you get further away you are moving towards the axis. At a given distance, there is a toe in angle where the offset of moving away in distance exaclty compensates for moving towards the axis. But it is different for each distance. The best choice is for the distance to a preferred seat or the average between the speakers and the wall behind the listener.
I am glad you brought up this ambiguity about the definition of a constant directivity speaker. My original understanding was that the speaker would have a constant output in all directions up to a cutoff angle. This made it an excellent choice for a public address system or sound reinforcement system in a large space like a sports arena because you could predictably seamlessly match multiple units to obtain the most uniform coverage and the highest overall gain before feedback. It's use as a high fidelity loudspeaker seems to be entirely different.
Completely wrong. Woofer beamwidth narrows with increasing frequency.
You do realize that the reverberant field of the tweeter is dependent upon the direct sound? It's what excites the reverberant sound!
A constant directivity speaker will act this way, not a 'normal' speaker. The woofer response does the same thing in the frequencies near the crossover, but yes, the shifting of a listening position will alter the SPL of the "omni mode" woofer range, while the "directional mode" will track to that of the tweeter, in a directivity matched CD speaker.
Well, naturally there are more and less preferred positions, but the two balanced effects (axial position vs. distance) do tend to compensate over a fairly large area, when properly set up.
Hello Soundmind
Why's that? Just scale things back. They sure sound good in a home setting. I have a used both CD and non CD designs and IMO the CD designs seem to sound and work the best, at least in my room.
Rob
Horn speaker technology was originally develped for the motion picture industry. In that application a large room, a motion picture theater had to be filled with a lot of sound from very low powered amplifiers, 5 or 10 watts. These types of speakers were originally developed mostly on the West coast. Wealthy people with large homes were able to acquire them and use them for home sound systems. Eventually they were dressed up and scaled down for home use. But these types of speakers. were never considered high accuracy by those for whom fidelity really mattered, lovers of classical music and opera. The best of these in the 1950s and 1960s like A7-500, Klipschorn, JBL S-7, S-8 were awful for this use. But they found a home with rock freaks and in applications with sound reinforcement consultants and contractors because of their efficiency and power handling capability.
They have been scaled down yet again for domestic use and more powerful magnets and higher temperature wire insulation has made it possible to make them even more efficient.
That being said, other technologies have not lived up to their expectations once imagined for them either. Over 100 years after Edison and Sabine, neither the technology to reproduce acoustic instruments from commercial recordings with high accuracy satisfactory to critical listeners nor the science to understand how have been achieved yet. What's more, the problem is not particularly interesting anymore to those with the potential to solve it. That is because there are far better things for the best scientists and engineers to spend their time and best efforts on. There's not much money in it for them either, just ask Dr Gedlee. For me it's just a hobby.
He didn't mention frequency (AFAIK) so if we are talking about full-space radiation he would be right.
That's a pretty big stretch there yogamaster
, considering we're talking about directivity across varying frequency ranges. IF that was what he meant it would have been qualified by "at LF" or "At 100Hz", rather than the general terms used which any reasonable interpretation has as across an operational bandwidth.
It's Dr. "Geddes" and thanks for the history lesson, but what does that have to do with anything? We - audio engineers - have learned a lot in the past few decades.
And I think that you are assuming the above. What I have said is that making speakers is not as profitable for me as consulting. But my consulting fees are pretty high.
What I think can be said is that the desire for big speakers to achieve high performance is a limited market. There are two reasons for this. First, many people have somehow become convinced that speakers don't need to be big to be good. Thats simply not true. The bigger they are the better I can make them. Second, people just don't want big speakers even if they do understand the tradeoffs. They will sacrifice sound quality for size in most cases.
I think that it is more meaningful to define CD as a frequency response that is constant with angle. That it changes in level is not important. And, of course, this can only be true above some frequency. For the example that I quoted it was true above 900 Hz. There has been a great deal of discussion about how low in frequency CD is required. Certainly at very LFs (in a small room) it is not require at all, there is no advantage what-so-ever. I would stand by the fact that the ear processes sound very slowly below 500 Hz as a reason that CD is not required below 500 Hz, because the integration time is too long and our sensitivity is falling. Where it lies above 500 Hz is an open issue and I would claim that a smooth transition from where it starts to be conatsnt to where it is omni-directional is more important than where this occurs, as long as there is CD above 1 kHz.
Above 1kHz our hearing is extremely accute and we can and do detect things like non-CD and group delay and a whole range of things that we cannot detect at lower frequencies. This is why I would never put a crossover above 1 kHz - there is no such thing as a "good" crossover, only ones that are better or worse.
You forgot reason number three - most people (of the general population) just don't care about sound quality. Sad but true.
Part of that is lack of exposure to really good quality systems to "calibrate" their expectations against, (you can't miss what you've never had) and part of it is just a general indifference.
If a small box can make a loud noise, that's enough for most people, never mind that the quality will be atrocious at higher SPLs.
Until that changes, both high quality and large speakers will remain a niche market.
That's true of direct radiators where a negligible % of the input is converted to output, but not for horns, where efficiency can be tens of % and bespeaks considerable acoustic coupling.
An example is one of my subs, a Danley DTS-10, whose paralleled drivers are a 2 ohm nominal load in free air but 4 ohms in the box.
And the drivers most definitely unload below the passband.
Others have corrected your misapprehension of the directivity of the woofer at the top end of its range near the crossover, (where its usually designed to match the CD horn's directivity) but lets assume you're talking about bass frequencies below 200Hz or so.
Do you really believe that as you move off axis from the listening position in a room that the bass response remains the same ? Direct to reflected ratio and woofer directivity become meaningless at bass frequencies in a living room sized room due to modal effects below the Schroeder frequency.
At the same time the woofer is omnidirectional, yet the reflections from the walls generate a huge three dimensional array of "virtual woofers" in virtually stacked rooms. This means it is the room the controls the "directivity" at bass frequencies and there is really nothing we can do to the main speakers to alter this.
So saying that the higher frequencies will be reduced as you go off axis but low frequencies will stay the same is simply not true in a room - low frequencies will vary all over the place around the room if you just have a pair of speakers with no distributed sub system.
I think it's pointless to examine the directivity of a speaker below the frequencies where the ear can distinguish between and separate the first arrival and the delayed reflections - below this frequency range (a few hundred Hz) the two merge together and the speakers and room become a single system which creates a summed response, and that summed response is what we perceive. (Not so at high frequencies)
It's also the frequency range where the room causes large deviations in the response that require EQ and/or spatial displacement of subs to correct, so if you are doing that correction anyway on a room by room basis, whether the speaker has some directivity down there or not is a moot point, since you will be equalizing the speaker+room response to get the correct end result.
I would suggest that for a CD system the absolute lowest frequency that it would ever need to be CD down to would be the point where direct arrivals and reflections can no longer be distinguished and merge together in our sensory perception. (A few hundred Hz) This is an absolute lowest frequency - practically speaking there may be other reasons why it can be quite a lot higher than this with minimal detrimental effect.
I see no reason for CD to be maintained down to bass frequencies, even if it were somehow possible in a confined room.
"I think that it is more meaningful to define CD as a frequency response that is constant with angle. That it changes in level is not important."
It is if you are a sound contractor installing a PA system. I'd bet the overwhelming majority of CD loudspeaker systems are installed in commercial sound reinforcement systems, PA systems and only a tiny fraction of the total used in home hi fi systems. Probably under 1%.
"Do you really believe that as you move off axis from the listening position in a room that the bass response remains the same ? "
Not to nearly the same degree as a tweeter and at the lowest frequencies for entirely different reasons than distance or angle. You can place a box speaker system outdoors and walk around it and when you're behind it all you hear is the woofer if there are no nearby structures to reflect higher frequencies. If the woofer becomes directional it is crossed over at too high a frequency. With these puny little 6" and 8" woofers so popular these days they can go fairly high without beaming much. For systems with a single driver in each range, the most limiting factor insofar as directionality is concerned is invariably the tweeter.
This raises a good point in asking what we mean by "Constant Directivity". Clearly directivity is the comparison of axial response level vs. total power response. It is either measured as Q or d.i. (dB) which is the logarithmic conversion of Q. Constant directivity clearly means that Q and d.i. are constant within a reasonable tolerance. It thereby guarantees that if a speaker's power response is equalized to a particular curve its axial frequency response will match. Conversely a given axial response will be accompanied by the same shape power response, if directivity is constant.
This says nothing about polar curves.
It also says nothing about the interplay of horizontal beamwidth and vertical beamwidth. For example, some measurements of the Altec 604 that I took years ago showed that when the vertical directivity expanded and contracted, the horizontal response did the opposite. The directivity index was fairly flat but the off axis response sucked.
So, clearly, we want more than just well behaved directivity. We want uniformity of the horizontal response and uniformity of the vertical response. As Earl say: frequency response that is constant with angle. This does not define polar response either. Some times polars are broad and level to some beamwidth and then fall off sharply. Other times the polars begin to fall as soon as you move off the central axis. We can probably live with either polar curve shape if it is uniform with frequency. This alone would allow us to equalize response on axis and have the response curve (not response level) stay constant as we move off axis.
PA designers do care about the shape of the polar curve because they have audience members distributed through a seating area at many radiation angles. As mentioned before the right polar shape may allow some compensation between horn angle and audience distance, letting you get more uniform level across a seating area. This has even been specifically exploited in some products such as the JBL J pattern line arrays.
For home audio we don't care as much about this as we typically cater to an audience of one. Still there may be an advantage to having a well defined relationship between axial response and total radiated power and (more important in my mind) well behaved response for a reasonably wide angular variation around the listening axis.
David S.
This says nothing about polar curves.
It also says nothing about the interplay of horizontal beamwidth and vertical beamwidth. For example, some measurements of the Altec 604 that I took years ago showed that when the vertical directivity expanded and contracted, the horizontal response did the opposite. The directivity index was fairly flat but the off axis response sucked.
So, clearly, we want more than just well behaved directivity. We want uniformity of the horizontal response and uniformity of the vertical response. As Earl say: frequency response that is constant with angle. This does not define polar response either. Some times polars are broad and level to some beamwidth and then fall off sharply. Other times the polars begin to fall as soon as you move off the central axis. We can probably live with either polar curve shape if it is uniform with frequency. This alone would allow us to equalize response on axis and have the response curve (not response level) stay constant as we move off axis.
PA designers do care about the shape of the polar curve because they have audience members distributed through a seating area at many radiation angles. As mentioned before the right polar shape may allow some compensation between horn angle and audience distance, letting you get more uniform level across a seating area. This has even been specifically exploited in some products such as the JBL J pattern line arrays.
For home audio we don't care as much about this as we typically cater to an audience of one. Still there may be an advantage to having a well defined relationship between axial response and total radiated power and (more important in my mind) well behaved response for a reasonably wide angular variation around the listening axis.
David S.