I think I mentioned this on another thread: the best data on this can be found in Toole's 1986 paper "Loudspeaker Measurements and Their Relationship to Listener Preferences: Part2" where he rank ordered 20 speakers in well controlled listening tests and then showed a number of measurements of each speaker. When you look at the speakers in order of preference you can see that wider and flatter axial or near-on-axis response is clearly an attribute of the higher ranked speakers. Yet his fig. 12 shows the power response of the same speakers and it is clear that a good power response curve isn't essential. Some systems with smooth and flat power responses are poorly ranked, others are top ranked with clear holes in their power response. (Most systems that follow Linkwitz/Reilly crossover thinking will have significant holes at the crossover points.)
If you agree that the ear takes primary clues from the early sound (direct sound and a few important reflections) then it is easy to accept that the total power response, which includes far off axis response that will arrive later, can be an unimportant part of what we use to judge frequency response.
I'm not sure there is an exact d.i. that is ideal. Less directional and you will hear more room sound and less direct sound. Some will like the spaciousness of that. More directional and you hear less of the room and more of the speaker directly. Others will prefer that. Hard to argue absolutes when these are matters of preference. As for radiation patterns, again hard to generalize. Can the ear hear radiation patterns?
That may well be, but if the only other better option were to prove to be a waveguide, OS or other, from the DIY side the dipole is far easier to fabricate. I find the concept of the OS waveguide intriguing as is a near-full range dipole or cardiode optimized for controlled directivity, but I'm in no position to DIY a waveguide. I also am not able to re-design my main listening room, my basement, so I live with what I have. I suspect that the best fit for fixed circumstances may indeed be a dipole. My classical 3-ways have sounded good in that environment and the occasional test upstairs, but I do always look for improvements, at least those that I can construct myself.
Dave
I'm actually a fan of controlled directivity, but more because it is likely to lead to smoother and flatter response over the probable range of listener positions (I think many listeners are significantly off axis, to some degree, in their typical listening). Beyond that, if you are shooting for some particular frequency response for reflections, I don't believe research supports that. Lipshitz and Vanderkooy suggest that reflections with holes in their response are fine, reflections with peaks might not be. Bech shows that the floor bounce is frequently audible. It arrives early enough with enough level (and at an angle where binaural hearing prevents separation) to fall within thresholds of perception. But its particular response isn't the issue, more the combined response of it and the direct sound.
I think the Toole and Olive vacillate on their conclusions on the subject. In the end their strongest statement is that smoothness of power response is a good revealer of significant resonances. In Olive's paper where he shows the correlation numbers between factors and ranking, power response smoothness is important, but not flatness. As repeated above, Toole's data shows no correlation between power response shape (or general d.i.) and perceived accuracy.
David
I would say that we do.
Remember that directivity and the room are tradeoffs between Spatiousness and imaging, as you say above. But what if I don't have to trade off those things, what if I can have both? Thats an improvement right? Its only a matter of opinion when there are tradeoffs involved because very few will not like the situation where you can have both.
So you use narrow directivity for good imaging and room reflections for good spatiousness. Of course you can't leave the room alone in this scenario, it must be a part of the solution and not a "given".
Did you read Sean's latest work on room equalizers? This is what I am refering too, Floyd was not part of this. He states very clearly that listeners prefered a room EQ which corrected a null in the loudspeakers power response. This is a distinct contradiction to your position.
Everybody agrees with "controlled directivity" (even if they don't achieve it), but a monopole is "controlled directivity". It a high DI CD that I feel is the right target and that is much harder to achieve.
I think that floor bounce and ceiling bounce are both serious issues that need to be corrected. Both of these can be dealt with at this first reflection point. Of course all reflections can be dealt with in this same way, but then we end up with a very dead room, and a undesirable tradeoff.
Last edited:
A well-debated subject in these pages, floor and ceiling reflections can also be managed via narrow vertical dispersion, i.e., axi-asymmetric waveguides.... 
Ref: Parham
Ref: Parham
When people are polite and show respect for each other the debates are fine. Its when this doesn't happen that they stop.
The subject of NON-axisymmetric waveguides is very interesting. I have a prototype, but no time, it seems. But there are tradeoffs, not the least of which is a dramatic increase in cost. Is it worth it? Only time will tell, but for now a throw rug and ceiling difusser work wonders.
The subject of NON-axisymmetric waveguides is very interesting. I have a prototype, but no time, it seems. But there are tradeoffs, not the least of which is a dramatic increase in cost. Is it worth it? Only time will tell, but for now a throw rug and ceiling difusser work wonders.
I'd ask "why in the heck would a person like that read or post in a measurements thread?" Please know Cal that I'm just joking. Of course humor always has a touch of reality in it.😉 😕Masochists.
Dan🙂
I would really like to measure your speakers. Bet the voicing has the energy center (pitch center) really near 632Hz (maybe error writing 682, oops sorry) so the balance sounds right. "Breaking away" from something that does not work is surely the right thing to do. Would not call that breaking away as much as fixing it. When designing products for others, by far the biggest complaint was caused when the pitch center was off. "To bright," "to thin," "not enough bass," "to much bass," (if nothing else is wrong) is often the pitch center being off.
In my DSP test system there is a program used to measure pitch centers. Split the passband into 4.
Believe it or not, the final calibration after all the meters and graphs say things are right is the ear. That is the "real" test as opposed to any measured estimate. We know the system is close because all the meters and graphs are showing right. Many attempts to tighten this correlation have been made (Toole and so on) but it is still the ear hear and not the gear. As stated many times in this thread the correlation between any measurement and experience is not as good as many of us wish. So, suppose I do the same thing as you. At some point the gear must be abandoned in favor of what sounds right, just like you. (Did I say that twice? haha) The gear guides us but is not the answer and is not the ears.
Would like to hear your speaks also.🙂
Anyone remember the post about dynamic matching of drivers? This could have something to do with the ear and the gear not agreeing. I use that mono amp with the driver in the feedback loop for the midrange and the tweeter on a simple amp which improves the dynamics on the mid because the tweet is way more dynamic than the mid. This use of different amps helps match them. Broadband dynamic range testing is quite difficult for gear and as nothing for the ear.
=SUM
By "energy center (pitch center)" do you mean a frequency at which the energy is equal above and below it? If so, then ~600Hz does seem to be the center for pink noise. Some music is lower - a lot of blues is centered around 400Hz.
I think you can have spatiousness and imaging if you can put your reflections in the right time range. Certainly, concert hall design focuses on that. Harder to do in a home listenng room.
With regard to hearing a polar response, the polars of the speaker and the room itself will determine the profile of reflections, their strength and direction. Isn't that what you really hear? Do you perceive that a speaker had strong response in a certain direction, rather than the room having a strong reflection?
If two speakers have a similar general directivity and create a similar IACC, would it matter that their polar curves were different?
SUM, are you using or designing for Mackie then? The driver in the feedback loop is what makes me question.
Dan
Dan
I did see it and corresponded with him on it. It falls into the class of DSP EQ that equalizes the steady state response to a target room curve.
This is a totally unreliable approach. If his listeners liked the end results it means, in my oppinion, that he got lucky on the d.i. curve of the sytem being equalized. All the papers that I have read on the subject strongly suggest that you'll want to EQ the early sound to flat and let the late sound fall where it may. Equalizing to a room curve can't guarantee that. I have considerable first hand experience with this with the Snell RCS 1000 (predecessor to the Tact), and the Audesey. You find that the technicians doing it are constantly changing the room target curve to get a good result.
Along similar lines are the cinema X curve which has changed over time and now has a different target for rooms of every size (clearly a moving target).
Finally, the room response being equalized isn't the system power response but some blend between the direct response, power response, and room absorption curve. His results show that, of 4 systems tried with 2 clearly performing poorly, the Harman unit was prefered. I don't think this proves that equalizing power response is correct..
David
Attachments
Dave - everything that I say here is in the context of small rooms, if you keep that in mind my comments might be clearer, because "Harder to do in a listening room" is the whole point.
Your room reflection arguments sound kind of like a chicken and egg argument. Do we hear the reflection of the polar response from the speaker at the angle of the reflection? Well Ya!
I can't answer your last question because "similar" is too ambiguous. If they are similar enough then we aren't disagreeing on anything.
Again Dave, I think that you missed my point. I agree that EQing the room curve is not the way to go and that you should onl;y ever EQ the direct field. And if what you are saying is that the whole test was flawed because they did this, well then that's a reasonable answer, but the fact remains that his data does contradict your argument.
Probably not making myself clear. By similar I meant similar d.i. rather than polar curve.
The original post was asking what the best polar shape was: cardiod, dipole, etc. I was trying to say that a "best" polar shape is hard to define, we may notice the more general properties of direct to reflected sound ratios and sometimes how different polars interact with the room, but I wouldn't expect one particular polar pattern to be superior to all others.
David
That gets to the basic question of whether there are absolutes in audio? Can we definitively say "this speaker is good, that speaker is bad"? The first step to answering that would be "do people describe the sound of a speaker in similar terms"? In my experience people do hear the same, at least however they do hear, they have gotten used to the acoustical world around them, and a speaker either replicates it or it doesn't.
Of course, when we are talking about flawed systems then everyone can pick what flaws they can live with and which they can't.
I consider myself a strong objectivist, but I'll admit that I listen and make minor adjustments (by ear) to every system I design. Can't let the meters have all the fun!
David
Last edited:
- Status
- Not open for further replies.