It has been quite a few years since I have looked at the math part of engineering, probably something like 10 years, because I had not been involved with software development mechanization for a long time. But I will try to understand it. I do know that for drivers that go up above 10KHz, the mass roll off starts at a much higher frequency which would be out of the range you are trying to simulate. So removing the low pass filter should not make a difference unless there is something else in the model.Kolbrek said:Seems like yet another post is required
Yes, since here the driver mass rolloff comes in, and the SPL will drop with frequency.
Did you read the explanation of the "Matched Source" concept? The matched source behaves like a driver with zero mass and infinite compliance.
So no EQ is applied in the polar maps (and certainly not EQ specifically designed for the OS), but the effect of driver mass rolloff is removed. In other words, no highpass filter is added, but a lowpass filter is removed.
Bjørn
Personally, I think for engineering purposes, it is best to model as close to the real thing as possible. Simply because in reality, the driver and guide/horn need to be considered as a total system. Of course, in the process of trying to resolve certain technical issues, it is beneficial to look at the guide/horn from different aspects, but unless the graphs and explanations are presented together, people that do not understand the concept of a "Matched Source" can easily be mislead.
No, there is no equalization of any kind, since none is applied. Not necessary for what goal? Even if EQ were applied, it would still not be CD since EQ cannot control directivity. Redefining a universally understood term to suit your personal preferences serves no purpose and does not change the results. If CD is the goal, then in the end, a LeCleach horn would be totally unsuitable for the purpose from everything I see about it. Why do you consistently avoid addressing that aspect, the fundamental goal? The implication of your statements is that the on-axis is everything. Get that flat, no EQ, that's it, one can ignore the off-axis.
Let me ask directly and see if we can get a direct, unambiguous, un-meandering response, keep it simple. Can a LeCleach horn approximate a CD pattern, EQ or no? If yes, how. If not, then just say no.
Dave
Hello dlr,
CD is something, as you should know, that I feel useless in Hifi listening and a goal I never wanted to reach in the design of the Le Cléac'h horn.
I cannot undesrtand -nor admit- why you are still using that parameter to denigrate the Le Cléac'h horn (did you ever listened to one?).
A very smooth pressure field is more important to those ears.
Smoothness is the only pertinent parameter you should compare between the polar of the OS waveguide and the polar of the Le Cléac'h horn (even if it was useless to apply +6dB equalization with that one) published at Earl's demand.
Earl admitted that CD cannot be obtained with a smooth polar and this is clearly visible from the graphs.
Best regards from Paris, France
Jean-Michel Le Cléac'h
CD is something, as you should know, that I feel useless in Hifi listening and a goal I never wanted to reach in the design of the Le Cléac'h horn.
I cannot undesrtand -nor admit- why you are still using that parameter to denigrate the Le Cléac'h horn (did you ever listened to one?).
A very smooth pressure field is more important to those ears.
Smoothness is the only pertinent parameter you should compare between the polar of the OS waveguide and the polar of the Le Cléac'h horn (even if it was useless to apply +6dB equalization with that one) published at Earl's demand.
Earl admitted that CD cannot be obtained with a smooth polar and this is clearly visible from the graphs.
Best regards from Paris, France
Jean-Michel Le Cléac'h
dlr said:
Jmmlc said:
I am curious as to why you would think that describing, accurately I believe, the technical aspects of one vs. the other is denigration? Either a goal is met by the product or it is not. If CD is the goal, then approximating CD must be the result. If it cannot approximate the goal, it is unsuitable. That is not denigration, that is stating the facts as they exist.
Requiring perfection in attaining the goal does not alter that. If your goal is not CD and it sounds to me that this is the case, then by all means seek your goal. In your post you state that smoothness is the only pertinent parameter. Fine. That has nothing to do with being CD, however. In fact, from all evidence, smoothness in the way that you prefer is counter-productive to being CD, so again, your design is unsuitable for the goal of CD, not just "not perfect".
In that sense, I continue to be a bit perplexed by your attitude and positions on this topic. You do not want a CD product. Then what's your interest in a thread based on a CD product? If you believe CD has no importance, then I'm even more perplexed, as it seems to be that you're more interested in scoring points in the areas that don't relate to CD.
I see two primary goals for CD. One, CD within physical limits, of course. Two, reducing HOM to the extent possible when coupled with the attempt to achieve CD. Your goal seems to be completely contrary to this. So why take issue when that is pointed out?
Dave
Jmmlc said:
Jean-Michel
Then you are completely in the minority since everyone from Linkwitz to Toole agree that CD is highly desirable. I think it quite convenient that you alone find something that your designs are incapable of doing unimportant.
At any rate, I completely agree with DLR. If your goals are so opposite to mine then why do you come here to argue about design approaches? If you want to argue about the goals thats fine, I'd love to get into why CD is so important, but not in this thread. Perhaps start one called "To CD or not CD - that is the question". Now there is an interesting topic.
Hello Earl,
You reply seems weird to me: it is you NOT me who asked Bjørn Kolbrek to publish the polar map of the Le Cléac'h horn in order to compare it to the OS waveguide.
I am surely not alone to think that CD cannot be separated for other characteristics as smoothness of the polar distribution in judging the directivity properties of a waveguide or a horn.
The conclusions from Toole's studies (when he was at the National Research Council not after...) were largely oversimplified.
If one read carefully Toole's study he'll see that Toole said that directivity increasing with frequency can be positive in order to obtain a more precise image. Many loudspeakers in the best noted class in his study were not constant directivity. (I read Toole's paper one more time 2 days ago)
He admitted that the large and deep soundstage many people apreciate is obtained at the expense of a lower precision in the image. The most explicit example being omnidirectionnal loudspeakers that produce a most surrounding reverberated field around the listener but is unable to give a precise image of the location , dimensions, etc... of the sources -instruments, voices...).
I belong to the category of audiophiles for which the precision of the image is more important to the width of the soundstage.
Best regards from Paris, France
Jean-Michel Le Cléac'h
You reply seems weird to me: it is you NOT me who asked Bjørn Kolbrek to publish the polar map of the Le Cléac'h horn in order to compare it to the OS waveguide.
I am surely not alone to think that CD cannot be separated for other characteristics as smoothness of the polar distribution in judging the directivity properties of a waveguide or a horn.
The conclusions from Toole's studies (when he was at the National Research Council not after...) were largely oversimplified.
If one read carefully Toole's study he'll see that Toole said that directivity increasing with frequency can be positive in order to obtain a more precise image. Many loudspeakers in the best noted class in his study were not constant directivity. (I read Toole's paper one more time 2 days ago)
He admitted that the large and deep soundstage many people apreciate is obtained at the expense of a lower precision in the image. The most explicit example being omnidirectionnal loudspeakers that produce a most surrounding reverberated field around the listener but is unable to give a precise image of the location , dimensions, etc... of the sources -instruments, voices...).
I belong to the category of audiophiles for which the precision of the image is more important to the width of the soundstage.
Best regards from Paris, France
Jean-Michel Le Cléac'h
gedlee said:
One cannot simply have a single item in a metrics of design factors. For example, you feel the designing for on axis frequency response is not as important as off axis response. Weighing of all design factors like these contemplate the tradeoffs that drive the final design. If you have only CD left for the argument. Of course you can stick with it. Since there is not quantifiable way of clearly determining the definition of CD, the technology must be way in the stone age. Nowadays, Kona coffee has better definition than that.gedlee said:
This seems quite consistent with the various sims I ran. But it does seem to vary less as the distance is further away from the source. So as long as we know we will be listening at a distance it really seems quite okay looking at the data. It seems that this variation with distance could be also related with the size of the mouth. A rough guess would be distance of 10x diameter of mouth or more might be desireable.Jmmlc said:
Jean-Michel
Your implication that CD does not provide for precise imaging is contradicted by the science and the data - the subjective response of the auditions of my designs. Subjective response is the only way to quantify imaging since it has no quantifiable measure, although we do know what impacts imaging. People universally feel that my CD waveguides have very precise imaging. Imaging is a direct field controlled phenomina and when the Very Early Reflections are properly handled through directivity then the imaging will be good if the direct field is flat and smooth. But later reflections have a strong impact on coloration because the ear integrates the sound from all locations when it evaluates the tone color and spatiousness. For this reason CD is required for low coloration, and directivity is required for good imaging. Your designs achieve directivity so they will image well, but they do not provide a flat reverberant field so they will not give a good sense of color or room spatiousness.
Toole is adamant that the reflections must have the same general response as the direct sound. How can this be possible with a source that has an ever narrowing directivity. You are picking and choosing Tooles work to suite your claims and not looking at it in its entirity.
While both the direct field and the reverberabt field are important one is free to place different priorities on them, as you and I do. But achieveing both can only be viewed as a preferable situation and when one of them is ignored then this can only be considered as sub-optimal. My designs achieve a flat direct field AND a flat reverberant field. That is the goal and that is what is achieved.
Once in the far field the response is not significantly affected by the distance from any source. Thats simply the physics of the situation. 3 m is going to be in the far field so the responses will not change with distance.
You never posted the complete polar maps of the devices, you only picked those aspects that suited you. I asked Bjorn to post ALL of the data simply because you were doing this. As you know, he agreed with me and hence we finally see all of the results.
Your implication that CD does not provide for precise imaging is contradicted by the science and the data - the subjective response of the auditions of my designs. Subjective response is the only way to quantify imaging since it has no quantifiable measure, although we do know what impacts imaging. People universally feel that my CD waveguides have very precise imaging. Imaging is a direct field controlled phenomina and when the Very Early Reflections are properly handled through directivity then the imaging will be good if the direct field is flat and smooth. But later reflections have a strong impact on coloration because the ear integrates the sound from all locations when it evaluates the tone color and spatiousness. For this reason CD is required for low coloration, and directivity is required for good imaging. Your designs achieve directivity so they will image well, but they do not provide a flat reverberant field so they will not give a good sense of color or room spatiousness.
Toole is adamant that the reflections must have the same general response as the direct sound. How can this be possible with a source that has an ever narrowing directivity. You are picking and choosing Tooles work to suite your claims and not looking at it in its entirity.
While both the direct field and the reverberabt field are important one is free to place different priorities on them, as you and I do. But achieveing both can only be viewed as a preferable situation and when one of them is ignored then this can only be considered as sub-optimal. My designs achieve a flat direct field AND a flat reverberant field. That is the goal and that is what is achieved.
Once in the far field the response is not significantly affected by the distance from any source. Thats simply the physics of the situation. 3 m is going to be in the far field so the responses will not change with distance.
You never posted the complete polar maps of the devices, you only picked those aspects that suited you. I asked Bjorn to post ALL of the data simply because you were doing this. As you know, he agreed with me and hence we finally see all of the results.
Hello,
1) Toole's written conclusions are often less affirmative that many say on forums like this one. IMHO he never dismissed one group of audiophiles even a minority.
2) An excellent CSD (derived for an excellent pulse response) is required for an excellent imaging. CSD results published for OSwaveguide are not excellent for the most...
3) You have a definition for flatness, for smoothness and few other Xness that differ from mine.
Best regards from Paris, France.
Jean-Michel Le Cléac'h
1) Toole's written conclusions are often less affirmative that many say on forums like this one. IMHO he never dismissed one group of audiophiles even a minority.
2) An excellent CSD (derived for an excellent pulse response) is required for an excellent imaging. CSD results published for OSwaveguide are not excellent for the most...
3) You have a definition for flatness, for smoothness and few other Xness that differ from mine.
Best regards from Paris, France.
Jean-Michel Le Cléac'h
gedlee said:
It is not necessary for the SPL to be almost the same level within 45 deg. If the off axis response has a similar balance as it rolls off with angle, it would acheive this. Well, if you also call this CD... What is the quantified definition of CD anyway? Nobody seems to know the answer!dlr said:
Or is it just another bag of worms?
Hello Soongsc,
with the help of Bjørn Kolbrek' polar map for the OS waveguide I'll try to give my definition of constant directivity.
First we have to define an attenuation level that should correspond to the limit of the area of the polar inside which the waveguide can be considered as CD. Not easy but you can by example take a value a bit lower than the min relative level recorded on axis on the polar (see attached graph): here the min is -8dB so let's take -10dB.
The question now is how to define the angle range and the frequency range inside which the CD equalized can be defined as CD at -10dB.
We can draw several rectangles inside which the polar doesn't fall below -10dB. For the purpose I traced 4 rectangles. For each of those rectangles we can say "the OS waveguide is CD at -10dB within such frequency range (the width of the rectangle) and such angle (the height that due to symetry along the axis we can have multiplied by 2)". In every 4 rectangles I traced the OS waveguide is CD at -10dB. Now which one should be used in order to be eventually a reference for further comparions.
A solution that's seems correct to me is to take the rectangle having the max area. The area is given (in undefined area units ) by the numbers in black. Here the rectangle having the maximum area is the one having an area of 210 area units.
With that definition one can says that the figured OS waveguide is CD at -10dB between 2kHz and 20kHz and from -45° to 45°.
That's another story if one defines the CD at -6 dB...
For what it's worth... ;-)
Best regards from Paris, France
Jean-Michel Le Cléac'h
with the help of Bjørn Kolbrek' polar map for the OS waveguide I'll try to give my definition of constant directivity.
First we have to define an attenuation level that should correspond to the limit of the area of the polar inside which the waveguide can be considered as CD. Not easy but you can by example take a value a bit lower than the min relative level recorded on axis on the polar (see attached graph): here the min is -8dB so let's take -10dB.
The question now is how to define the angle range and the frequency range inside which the CD equalized can be defined as CD at -10dB.
We can draw several rectangles inside which the polar doesn't fall below -10dB. For the purpose I traced 4 rectangles. For each of those rectangles we can say "the OS waveguide is CD at -10dB within such frequency range (the width of the rectangle) and such angle (the height that due to symetry along the axis we can have multiplied by 2)". In every 4 rectangles I traced the OS waveguide is CD at -10dB. Now which one should be used in order to be eventually a reference for further comparions.
A solution that's seems correct to me is to take the rectangle having the max area. The area is given (in undefined area units ) by the numbers in black. Here the rectangle having the maximum area is the one having an area of 210 area units.
With that definition one can says that the figured OS waveguide is CD at -10dB between 2kHz and 20kHz and from -45° to 45°.
That's another story if one defines the CD at -6 dB...
For what it's worth... ;-)
Best regards from Paris, France
Jean-Michel Le Cléac'h
soongsc said:
It is not necessary for the SPL to be almost the same level within 45 deg. If the off axis response has a similar balance as it rolls off with angle, it would acheive this. Well, if you also call this CD... What is the quantified definition of CD anyway? Nobody seems to know the answer!
