Every transducer has characteristic directivity/frequency (narrowing to upper end) - match them, harmonize with the baffle. This is what I meant.
Waveguides and horns are different in this respect, they stretch the comfort zone
Typically a speaker's low end "usability" is limited by
- Fs
- distortion
- Xmax
and high end usability is limited by
- narrowing dispersion
- cone breakup
- response raggedness due to many factors
Basic stuff, perhaps too basic for you Earl
A 2-way is a multi-way in this respect.
Actually my post was to experss some pessimism about soongsc's quest for the Holy Grail of transducers...
Waveguides and horns are different in this respect, they stretch the comfort zone
Typically a speaker's low end "usability" is limited by
- Fs
- distortion
- Xmax
and high end usability is limited by
- narrowing dispersion
- cone breakup
- response raggedness due to many factors
Basic stuff, perhaps too basic for you Earl
A 2-way is a multi-way in this respect.Actually my post was to experss some pessimism about soongsc's quest for the Holy Grail of transducers...
Last edited:
That's not actually "dealing with directivity", its just accepting that pistons always have a narrowing directivity and just moving on. Basically abandoning constant directivity as a design goal.
Too often I hear this kind of claim and I wanted to point out the fallacy of it. Multiple drivers do not achieve constant directivity and never can. They do deal with the other factors that you mention, but not directivity.
I will admit that using some form of waveguide in the baffle, like Genelec does, can achieve CD, but it is a fairly wide CD. Its not a high DI CD, which is my goal. A flat DI with frequency (CD) is very desirable, but a flat DI with frequency at a higher DI level is even more desirable. Maybe I should call my goal HDICD for High Directivity Index Constant Directivity. Or CHD for Constant High Directivity. CD is just too broad.
Too often I hear this kind of claim and I wanted to point out the fallacy of it. Multiple drivers do not achieve constant directivity and never can. They do deal with the other factors that you mention, but not directivity.
I will admit that using some form of waveguide in the baffle, like Genelec does, can achieve CD, but it is a fairly wide CD. Its not a high DI CD, which is my goal. A flat DI with frequency (CD) is very desirable, but a flat DI with frequency at a higher DI level is even more desirable. Maybe I should call my goal HDICD for High Directivity Index Constant Directivity. Or CHD for Constant High Directivity. CD is just too broad.
Last edited:
Constant directivity through the whole passband is something that we don't see today. Controlled directivity can be achieved by many ways and to various grades of DI.
Here is a remarkable 3-way with cone drivers. No wg, no horns, no dipoles - just a 3-way with a curved wide cabinet. Not bad at all.
KS-483 from here DIY archive of Kimmo Saunisto
Gedlee Summa uses a horn and cone bass to achieve controlled directivity. Sonogram from Earls white paper vs. my AINOgradient (dipole and a small horn) - both achieve very similar directivity as KS-483 by different "mechanisms". DI is much higher that KS-483. Both sonogarams are not-normalized.
Notice - different scales and measurement conditions - not exactly comparable!
Here is a remarkable 3-way with cone drivers. No wg, no horns, no dipoles - just a 3-way with a curved wide cabinet. Not bad at all.
KS-483 from here DIY archive of Kimmo Saunisto
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Gedlee Summa uses a horn and cone bass to achieve controlled directivity. Sonogram from Earls white paper vs. my AINOgradient (dipole and a small horn) - both achieve very similar directivity as KS-483 by different "mechanisms". DI is much higher that KS-483. Both sonogarams are not-normalized.
Notice - different scales and measurement conditions - not exactly comparable!
Attachments
Last edited:
I would agree with that.
If you have raw data in a reasonable form I can plot it in the same format as the Summa.
Last edited:
The cone n dome multiway has collapsing dispersion in the highs- so the highs will never sound 'just right', but either overbright or muted. You can only balance the power response and axial response in a CD design, and even then you're limited on bandwidth.
Still seems quite well-done though.
Still seems quite well-done though.
What you've achieved so far is HDICDBOAHF – High Directivity Index Constant Directivity But Only At Higher Frequencies
Markus
More like Mid freqs as the DI is fairly constant above about 700 Hz. Not much has a constant DI below that and as I have said so often, I don't think that it matters as much at the lower freqs.
Always the perfectionist.
(I need to update that white-paper, that curve is so out-of-date.)
More like Mid freqs as the DI is fairly constant above about 700 Hz. Not much has a constant DI below that and as I have said so often, I don't think that it matters as much at the lower freqs.
Always the perfectionist.
(I need to update that white-paper, that curve is so out-of-date.)
Thank you Earl, but that outdoor measurement is not up to date any more. I will soon be able to publish measurements with Neo8-PDR. According to proto measurements it will make directivity even smoother between 800-4000Hz. Because of the weather I can't do outdoor measurements now.
Markus is actually speaking for dipole or cardioid woofers (or huge horns?). They have virtues and measure "better" below say 300Hz than large driver-monopoles. However we have this battle of the meaning of that. I can say that yes a well done dipole bass in a proper room and placement can sound excellent. But that includes many ifs
I assume you are not talking about arrays, that can be designed with weighting profiles to achieve any arbitrary wavefront and directivity. At McIntosh I designed a sinx / x array with very steep angular cuttoff. With DSP it would be easy to adjust that profile (or any arbitrary profile, for that matter) to grow and shrink in proportion to wavelength to give very constant directivity. Look up the work of 1 Limited who have done very intereting work with this.
Even the simple symmetrical arrays as I did at Snell used a careful transition between 1 inner element and 2 outer elements to create lobe free transitions and parallel response at multiple vertical angles. I was designing with a target of parallel response for 0 and 40 degres (+- vertical ) and achieving more response uniformity than I recall seeing from your horns.
As to very high directivity, what is the justification of this? I assume a cleaner impulse response in a live room will result, but is that necessarily a good thing? (Toole disagrees in his book.) If we can achieve a fairly clean early response, especially with regard to local boundaries and vertical reflections, why do we need extra deadness beyond that?
Finally, since I believe it is your standard practice to listen off axis to your horns (to get past some axial issues) what is the effective d.i. under those conditions?
David
Dave
You always say "Well mine was better" so I'll let those comments slide.
Yes I was not talking about shaded arrays, but still they would find it difficult to achieve a narrow directivity. Better would be to use the approach that Genelec uses - simpler and more cost effective.
Narrow directivity avoids the early reflections and yes this is precisely where Toole and I disagree. Greisinger agrees with me by the way.
Since my systems are constant directivity the DI is independent of listening angle. And, by the way, I toe in the speakers not to avoid "axial issues" but to yield a wider sweet spot.
You always say "Well mine was better" so I'll let those comments slide.
Yes I was not talking about shaded arrays, but still they would find it difficult to achieve a narrow directivity. Better would be to use the approach that Genelec uses - simpler and more cost effective.
Narrow directivity avoids the early reflections and yes this is precisely where Toole and I disagree. Greisinger agrees with me by the way.
Since my systems are constant directivity the DI is independent of listening angle. And, by the way, I toe in the speakers not to avoid "axial issues" but to yield a wider sweet spot.
keyser, Kimmo is a very sharp guy and careful with what he shows. I believe that it was done indoors and with variable gating up to 300ms in the low end (I don't know about lsp) He has said that he was really surprised when he saw the results and the speaker works extraordinarily well in a room! I believe that the magic lies in the wide bafle and wisely chosen xos. Directivity control is good but DI is lowish though.
Generally this thread is bubbling about what kind of form(slope) and level the DI/Fq curve should be. The right way to measure and plot is important if we do comparisons, but for us amateurs we must accept differencies in methods and scales.
And then, DI doesn't tell much alone. A dipole has high DI but strong radiation bacwards unlike horns. They sound different in a room even with exactly same DI curve.
Generally this thread is bubbling about what kind of form(slope) and level the DI/Fq curve should be. The right way to measure and plot is important if we do comparisons, but for us amateurs we must accept differencies in methods and scales.
And then, DI doesn't tell much alone. A dipole has high DI but strong radiation bacwards unlike horns. They sound different in a room even with exactly same DI curve.
Last edited:
Did you realize the lowest line is 120° (240° pattern)? That data plugged into most of the sonogram-style graphs around this thread would just be a box of red.
Earl,
I wasn't saying in any way that "mine was better". You made a very unambiguous statement that "Multiple drivers do not achieve constant directivity and never can" and I was pointing out that that is patently untrue.
Shaded arrays typicaly give even narrower directivity than the typical horn. The potential for directivity is related to source size so any longish line array will be bigger than the largest horn and surpass it in directivity, if that is really your objective. The most impressive demonstration I have heard of directivity was the 1 Limited array designed for Pioneer. It could send out pencils of sound that could bounce off of various room surfaces to create the multiple channels of a 5.1 system. The on beam/off beam discrimination was stronger than any other demonstration of directivity that I have experienced.
I wasn't aware that your systems were constant directivity. All the curves I have seen show a smoothly rising directivity. That isn't a bad thing and I don't want to diminish it as an achievement, but I do think the term "Constant Directivity" gets brandished around a lot these days. (Especially when we are at best tacking a CD horn on a rising directivity woofer).
The primary application of d.i. is in setting the relationship between axial level and reverberent level. As d.i. is defined as the relationship between the level on the primary lobe vs. the spherically integrated response, there is a different effective d.i. if we listen off the primary lobe. That is, if the speaker has a d.i. of 10 but we listen off axis where level has dropped 3dB, then the effective d.i. is really only 7.
I think this applies in your recommended installation.
David S.
I wasn't saying in any way that "mine was better". You made a very unambiguous statement that "Multiple drivers do not achieve constant directivity and never can" and I was pointing out that that is patently untrue.
Shaded arrays typicaly give even narrower directivity than the typical horn. The potential for directivity is related to source size so any longish line array will be bigger than the largest horn and surpass it in directivity, if that is really your objective. The most impressive demonstration I have heard of directivity was the 1 Limited array designed for Pioneer. It could send out pencils of sound that could bounce off of various room surfaces to create the multiple channels of a 5.1 system. The on beam/off beam discrimination was stronger than any other demonstration of directivity that I have experienced.
I wasn't aware that your systems were constant directivity. All the curves I have seen show a smoothly rising directivity. That isn't a bad thing and I don't want to diminish it as an achievement, but I do think the term "Constant Directivity" gets brandished around a lot these days. (Especially when we are at best tacking a CD horn on a rising directivity woofer).
The primary application of d.i. is in setting the relationship between axial level and reverberent level. As d.i. is defined as the relationship between the level on the primary lobe vs. the spherically integrated response, there is a different effective d.i. if we listen off the primary lobe. That is, if the speaker has a d.i. of 10 but we listen off axis where level has dropped 3dB, then the effective d.i. is really only 7.
I think this applies in your recommended installation.
David S.
Last edited:
I'm not sure that is a proper paraphrasing of Toole. I think the consensus is pretty universal that strong early reflections, especially if they arrive in the vertical plane of the speakers are a bad thing.
Where Toole seems to differ with the general consensus is his belief that a high level of late energy is not only acceptable but a good thing. He refers to the case of the Quad ESL and that it was marked down strongly in mono listening tests.
Toole argues that a fairly low d.i. is a good thing and that the contribution from the room is not to be feared. The Soren Bech papers were pretty clear on that too. I wouldn't take it to an extreme but I don't generally disagree with it.
David S.