CeramicMan said:
No need to these, P18's FR curve in that region is noticeably smoother than L18, you just need to know what to look for. The CSD results are no suprise, I posted on this above with the details. That is a reason why CSD can be useful, it is visually easier to see those. But they are also already in the FR of the drivers.
Re: Re: Re: There's little agreement as to which is best
You've spent a lot of words there stating that drivers are minimum phase, yeah I know that, most are. I only stated it as a requirement to be complete in my statements. I don't think I ever said anything but that it was a requirement.
The claim at the start was that the L18 with 7 peaks in it's response could be made to perform as well, even better than the P18 through linear EQ. You don't need to provide any proof that you can EQ the FR, I get that, it is very basic. My point is that you cannot simply treat just the linear case without also considering the non-linear aspects of the driver.
I'm not concerned with where the distortion originates in metal cone drivers, the fact is that it is significantly higher and no amount of linear EQ in the crossover will correct it, and make it the same as a driver with less distortion which is what was suggested at the start of this thread.
What I take issue with regarding your view of the matter is that you think you can just look at the linear aspects of a system and ignore the non-linear.
What I see many saying here is that if you let us pretend that the system is linear then this theorem is valid, but the fact is that the system is not very linear and so you cannot make such forceful and rigorous claims. I'm bowing out now thanks.
Pete B.
[/QUOTE]dlr said:
You've spent a lot of words there stating that drivers are minimum phase, yeah I know that, most are. I only stated it as a requirement to be complete in my statements. I don't think I ever said anything but that it was a requirement.
The claim at the start was that the L18 with 7 peaks in it's response could be made to perform as well, even better than the P18 through linear EQ. You don't need to provide any proof that you can EQ the FR, I get that, it is very basic. My point is that you cannot simply treat just the linear case without also considering the non-linear aspects of the driver.
I'm not concerned with where the distortion originates in metal cone drivers, the fact is that it is significantly higher and no amount of linear EQ in the crossover will correct it, and make it the same as a driver with less distortion which is what was suggested at the start of this thread.
What I take issue with regarding your view of the matter is that you think you can just look at the linear aspects of a system and ignore the non-linear.
What I see many saying here is that if you let us pretend that the system is linear then this theorem is valid, but the fact is that the system is not very linear and so you cannot make such forceful and rigorous claims. I'm bowing out now thanks.
Pete B.
Re: Re: Re: Re: There's little agreement as to which is best
What I see many saying here is that if you let us pretend that the system is linear then this theorem is valid, but the fact is that the system is not very linear and so you cannot make such forceful and rigorous claims. I'm bowing out now thanks.
Pete B. [/B][/QUOTE]
Sorry but the start of the thread was about the CSD differences of two drivers, and a response to it saying the CSD is directly related to the FR of the drivers. When talking about FR, CSD, it is already assumed that the driver is linear. Whether or not it is really linear, or how much linear is irrelevant, because these already presume that the system is linear. And since this is already presumed, there is nothing wrong bringing the point that if FR irregularities are corrected the CSD and time response will be correct as well. Your point that drivers are nonlinear so they can't be corrected completely is valid, but irrelevant in this discussion.
PB2 said:
What I see many saying here is that if you let us pretend that the system is linear then this theorem is valid, but the fact is that the system is not very linear and so you cannot make such forceful and rigorous claims. I'm bowing out now thanks.
Pete B. [/B][/QUOTE]
Sorry but the start of the thread was about the CSD differences of two drivers, and a response to it saying the CSD is directly related to the FR of the drivers. When talking about FR, CSD, it is already assumed that the driver is linear. Whether or not it is really linear, or how much linear is irrelevant, because these already presume that the system is linear. And since this is already presumed, there is nothing wrong bringing the point that if FR irregularities are corrected the CSD and time response will be correct as well. Your point that drivers are nonlinear so they can't be corrected completely is valid, but irrelevant in this discussion.
Hi Pete,
I know we seem to be talking past each other. If so, then my comments are for others.
But you're having trouble clearly distinguishing between linear and nonlinear distortion.
Linear distortion is just amplitude distortion and can be corrected.
Nonlinear distortion is an entirely different phenomena and is dominated by the motor and excursion. Nonlinear distortion cannot be equalized, or at least not easily. (servo feedback, etc. There are ways of eq'ing or mitigating NL distortion)
However, directly connecting the amplitude of cone breakup and nonlinear distortion is incorrect.
Cone breakup may be responsible for some nonlinearities and in fact this can be measured. But it is not the dominant factor in nonlinear distortion.
If you're talking about nonlinear phenomena, you cannot make the assumption that "soft breakup" of a paper dome over a wide frequency range has less nonlinear distortion than "hard breakup" associated with a metal cone. You are mixing the amplitude distortion (read, linear distortion) with any nonlinear phenomena.
There has been no experimental data on nonlinear distortion and cone breakup presented anywhere in this thread. I've actually been trying to use some close thirds measurements to see if I can find any experimental correlation of breakup and nonlinear distortion. This would not show in a CSD. More on that someday, when I have time. Still, it seems that the low end/excursion dominates.
It is an error to assume more severe linear distortion implies more severe nonlinear distortion.
Many metal cones have clearly better nonlinear distortion than their paper counterparts. Look at the Seas datasheets for the metal and non metal seas drivers. It is quite clear that the metal cones have better nonlinear numbers, regardless of their fr/csd.
edit-didn't see Feyz' post. He said the same more succinctly...
I know we seem to be talking past each other. If so, then my comments are for others.
But you're having trouble clearly distinguishing between linear and nonlinear distortion.
Linear distortion is just amplitude distortion and can be corrected.
Nonlinear distortion is an entirely different phenomena and is dominated by the motor and excursion. Nonlinear distortion cannot be equalized, or at least not easily. (servo feedback, etc. There are ways of eq'ing or mitigating NL distortion)
However, directly connecting the amplitude of cone breakup and nonlinear distortion is incorrect.
Cone breakup may be responsible for some nonlinearities and in fact this can be measured. But it is not the dominant factor in nonlinear distortion.
If you're talking about nonlinear phenomena, you cannot make the assumption that "soft breakup" of a paper dome over a wide frequency range has less nonlinear distortion than "hard breakup" associated with a metal cone. You are mixing the amplitude distortion (read, linear distortion) with any nonlinear phenomena.
There has been no experimental data on nonlinear distortion and cone breakup presented anywhere in this thread. I've actually been trying to use some close thirds measurements to see if I can find any experimental correlation of breakup and nonlinear distortion. This would not show in a CSD. More on that someday, when I have time. Still, it seems that the low end/excursion dominates.
It is an error to assume more severe linear distortion implies more severe nonlinear distortion.
Many metal cones have clearly better nonlinear distortion than their paper counterparts. Look at the Seas datasheets for the metal and non metal seas drivers. It is quite clear that the metal cones have better nonlinear numbers, regardless of their fr/csd.
edit-didn't see Feyz' post. He said the same more succinctly...
Some perspective
To give some perspective, assume the driver distorts 20% on AVERAGE (which is a very bad driver) while playing music. 20% distortion means the distortion is -14 db lower than the signal generated. -14db error while doing frequency response correction or shaping is acceptable. I know this is a simplification, but at least gives some perspective, that attempting to correct FR anamolies of drivers are not a futile attempt just because the drivers are nonlinear. I would be more worried about messing up the off axis response while correcting the on axis than the nonlinearity of the driver messing up the equalization.
To give some perspective, assume the driver distorts 20% on AVERAGE (which is a very bad driver) while playing music. 20% distortion means the distortion is -14 db lower than the signal generated. -14db error while doing frequency response correction or shaping is acceptable. I know this is a simplification, but at least gives some perspective, that attempting to correct FR anamolies of drivers are not a futile attempt just because the drivers are nonlinear. I would be more worried about messing up the off axis response while correcting the on axis than the nonlinearity of the driver messing up the equalization.
Feyz said:
Well, there must be some form of curve fitting used in the graph, otherwise there would be large regions of empty space between isolated dots in the vicinity of 1500, 1800, 2100, 2400, 2700... etc Hz.
Both CSD graphs display a response increase when approaching 15kHz~20kHz, which is not present in the original FR plots. This suggests either a problem with shape of the sliding windows, or the total sampling period was too short, or both. The sliding windows could also have different widths, making the L18 response appear to be "sustained" longer for no apparent reason.
Re: Re: Re: Re: Re: There's little agreement as to which is best
Look try to be certain that what you write and claim is accurate, a few people concentrated on the CSD, and you seem fixated on it, and the fact that once again you think you can assume a linear system just because it is "presumed". So reality doesn't matter just that we all stick to you're chosen presumptions which make you correct? I was trying to give the group a reality check!
This is the first post and how the thread started, much more than CSD:
"Curious about ZAPH's designs. Seas L18 and P18RNXP Post #1
Having NO knowledge in speaker design, I'm doing alot of reading and trying to get a better understanding of what makes a good driver (among other things). I was looking at Zaph's L18 design and looking at his 6.5" measurements and comparison... And can't help but wonder why did he choose the L18 over the P18RNXP?
If I look at the responce curve and the harmonic distortion, They are very comparable (to me). But when I look at the SCD graph, the P18RNXP looks like a clear winner. What am I not understanding or overlooking?"
Seems he asked about a lot of things and wants a basic understanding.
Pete B.
Feyz said:
Look try to be certain that what you write and claim is accurate, a few people concentrated on the CSD, and you seem fixated on it, and the fact that once again you think you can assume a linear system just because it is "presumed". So reality doesn't matter just that we all stick to you're chosen presumptions which make you correct? I was trying to give the group a reality check!
This is the first post and how the thread started, much more than CSD:
"Curious about ZAPH's designs. Seas L18 and P18RNXP Post #1
Having NO knowledge in speaker design, I'm doing alot of reading and trying to get a better understanding of what makes a good driver (among other things). I was looking at Zaph's L18 design and looking at his 6.5" measurements and comparison... And can't help but wonder why did he choose the L18 over the P18RNXP?
If I look at the responce curve and the harmonic distortion, They are very comparable (to me). But when I look at the SCD graph, the P18RNXP looks like a clear winner. What am I not understanding or overlooking?"
Seems he asked about a lot of things and wants a basic understanding.
Pete B.
Hi,
Well the above answers the basic question, pistonic cone behaviour
to well beyond the c/o point, which cannot be done with "soft" drivers.
"Soft" drivers of this size enter break up mode between 1KHz and 2kHz.
You pay your money and you take your choice : "perfect" midband
but some emphasis of distortion, to "very good" midband with no
emphasis of distortion.
Peoples opinions of the above may vary, but its Zaph's perogative
to build the speakers the way he wants to, you go along with his
methodology or not. It is not a philosophical question, the L18
design is one of the best all metal cone designs available, if you
do not like metal cone designs leave it, if you do consider it ....
I do not think a particular metal cone design should have to defend
the whole metal vs plastic vs paper vs textile debate. Seas do not
bother with such questions either, they simply give the customer
pretty good versions of what they want for various technologies.
/sreten.CeramicMan said:
The space between those points are my guess is filled by using a longer FFT size.
But, the point is just by looking at the two FR curves of the two drivers, to my eyes, I can easily say that P18 is a better choice in terms of linear distortion (FR and time) in the region below 3Khz. Because L18's FR curve is more wiggley (for the lack of a better term) at that region and P18's is smoother in contrast. And CSD just relays that information also.
You are right the CSD window used at start must be different than the ones used at FR generation, and that makes CSD more suspect.
What I am really trying to say is, there is no magical energy storage property of the drivers that doesn't already show themselves on a high resolution FR graph. What tools, graphs one uses to detect or display them is a choice.
Re: Re: Re: Re: There's little agreement as to which is best
Don't presume that every word written at a board is for your edification alone. Lack of precision and thoroughness tends to leave things such that misinterpretation results. That may be the case here to some degree and I think that I may see why we're going around in circles.
I have not addressed the question directly about the two drivers. Others have, I see no node for me to do so. I have not heard them, others have, zaph in particular. His first post was probably the most direct, correct and useful post in this whole thread as a response to the original poster. Unfortunately, it took the downturn with other posts that had not only misleading but bad information. That should be abundantly clear now.
No, you are now stating your conjecture. Based on those who have made extensive measurements as well as the manufacturers data shows that for the same motor, the nonlinear distortion will actually be higher for softer diaphragms in the passband, though it can vary due to different amplification from the linear distortion characteristic differences between cones. What this means and what has remained unchanged is that it is up to the designer to select an appropriate driver for the task. The impact of where the distortion enters into must be known for an optimum design, since optimum is all one can approach and what's optimum for one person is not necessary the same for the next person.
Consider a midrange. Nonlinear distortion is largely related to displacement. If I want to minimize that component, one way is to use a somewhat higher highpass Fc to lower the displacement in the final design. For a woofer I could do the same, sacrificing a bit of extension. But in the end it's all a choice of the designer. You cannot tell me with any credibility what the final distortion characteristics will be present in any design until that design is complete. The linear distortion corrections made by the crossover have a direct and significant impact on that result. There is no absolute with raw drivers because of this.
Some drivers may have very good low levels of distortion until they have significant displacement. If one does it right, even small amounts of Fc change could have a strong impact on the final distortion characteristics. So the linear and the nonlinear tied in that fashion, but they are separate in others, one being the ability to use CAD software to control the linear aspect directly.
I never said that and if you had paid more attention to my earlier posts you would know my position on this. I separated the two because the original thread veered off into parts with one being on the absolutes of a CSD vs. FR. Even within that sub-thread, so-to-speak, I pointed out specifically the issue of the nonlinear distortion impact, but did not care to go down that path as the discussion was focused on the linear distortions.
We are not pretending. We are discussing the same as those who originally postulated the case for computer modeling and then implemented software that, as others here have explained quite clearly, must rely on the system as being LTI. The nonlinear aspect is then examined through measurements that cannot be determined by the CSD nor the FR. The designer then tries to use his own brain to use both sets of data to make his decisions on choice of drivers, then on where and how to cross. Some may to it vice-versa, that doesn't matter. In the end, the final distortion characteristics can still only be known through subsequent measurements.
The drivers are far more linear than you realize. Were that not the case, none of the CAD software used for many years would not work as they rely on that fact. They do not address directly the nonlinear influences in the drivers, that part is true, but that's what the designer does separately. It's a process and to some degree still an art. That's part of what makes speaker design an interesting hobby for many.
Dave
[/QUOTE]
Don't presume that every word written at a board is for your edification alone. Lack of precision and thoroughness tends to leave things such that misinterpretation results. That may be the case here to some degree and I think that I may see why we're going around in circles.
I have not addressed the question directly about the two drivers. Others have, I see no node for me to do so. I have not heard them, others have, zaph in particular. His first post was probably the most direct, correct and useful post in this whole thread as a response to the original poster. Unfortunately, it took the downturn with other posts that had not only misleading but bad information. That should be abundantly clear now.
No, you are now stating your conjecture. Based on those who have made extensive measurements as well as the manufacturers data shows that for the same motor, the nonlinear distortion will actually be higher for softer diaphragms in the passband, though it can vary due to different amplification from the linear distortion characteristic differences between cones. What this means and what has remained unchanged is that it is up to the designer to select an appropriate driver for the task. The impact of where the distortion enters into must be known for an optimum design, since optimum is all one can approach and what's optimum for one person is not necessary the same for the next person.
Consider a midrange. Nonlinear distortion is largely related to displacement. If I want to minimize that component, one way is to use a somewhat higher highpass Fc to lower the displacement in the final design. For a woofer I could do the same, sacrificing a bit of extension. But in the end it's all a choice of the designer. You cannot tell me with any credibility what the final distortion characteristics will be present in any design until that design is complete. The linear distortion corrections made by the crossover have a direct and significant impact on that result. There is no absolute with raw drivers because of this.
Some drivers may have very good low levels of distortion until they have significant displacement. If one does it right, even small amounts of Fc change could have a strong impact on the final distortion characteristics. So the linear and the nonlinear tied in that fashion, but they are separate in others, one being the ability to use CAD software to control the linear aspect directly.
I never said that and if you had paid more attention to my earlier posts you would know my position on this. I separated the two because the original thread veered off into parts with one being on the absolutes of a CSD vs. FR. Even within that sub-thread, so-to-speak, I pointed out specifically the issue of the nonlinear distortion impact, but did not care to go down that path as the discussion was focused on the linear distortions.
We are not pretending. We are discussing the same as those who originally postulated the case for computer modeling and then implemented software that, as others here have explained quite clearly, must rely on the system as being LTI. The nonlinear aspect is then examined through measurements that cannot be determined by the CSD nor the FR. The designer then tries to use his own brain to use both sets of data to make his decisions on choice of drivers, then on where and how to cross. Some may to it vice-versa, that doesn't matter. In the end, the final distortion characteristics can still only be known through subsequent measurements.
The drivers are far more linear than you realize. Were that not the case, none of the CAD software used for many years would not work as they rely on that fact. They do not address directly the nonlinear influences in the drivers, that part is true, but that's what the designer does separately. It's a process and to some degree still an art. That's part of what makes speaker design an interesting hobby for many.
Dave
[/QUOTE]
Re: Re: Re: Re: Re: There's little agreement as to which is best
You quote me, make a direct reply, suggest that I need to be edified (LOL) and lack precision and think you alone hold the rigorous proofs and facts. Then claim that you're statements are not directed at me. If that is the case then don't quote me and don't try to spin what you say. Say what you mean and mean what you say. I find your comments offensive and I'm going to stop reading you at this point, it is a waste of time. You go rambling off topic, also claim that the example you choose to discuss is not the subject of this thread. My point is not about specific examples anyway, it is about the theory that you demand is rigorously true as you apply it, when it is not. You are the one who needs to read and learn thanks, rather than suggesting it to me. I joined the AES over 30 years ago, yeah I read the article thanks, and have spoken directly to Witold.
Several here spout a linear systems theorem, pound their fist and say this is the rigorous truth, when in fact if they opened their eyes they might notice that it is only an approximation. I've said it enough times, several here seem to think they are the only ones with the facts and offer offensive comments to those who offer a more accurate view.
Pete B.
[/QUOTE]dlr said:
You quote me, make a direct reply, suggest that I need to be edified (LOL) and lack precision and think you alone hold the rigorous proofs and facts. Then claim that you're statements are not directed at me. If that is the case then don't quote me and don't try to spin what you say. Say what you mean and mean what you say. I find your comments offensive and I'm going to stop reading you at this point, it is a waste of time. You go rambling off topic, also claim that the example you choose to discuss is not the subject of this thread. My point is not about specific examples anyway, it is about the theory that you demand is rigorously true as you apply it, when it is not. You are the one who needs to read and learn thanks, rather than suggesting it to me. I joined the AES over 30 years ago, yeah I read the article thanks, and have spoken directly to Witold.
Several here spout a linear systems theorem, pound their fist and say this is the rigorous truth, when in fact if they opened their eyes they might notice that it is only an approximation. I've said it enough times, several here seem to think they are the only ones with the facts and offer offensive comments to those who offer a more accurate view.
Pete B.
Re: Re: Re: Re: Re: Re: There's little agreement as to which is best
I was not intending to be offensive and since you took it as such, for that I apologize. At times my writing is not as clear as I would like it to be. I do still have disagreements with your take on some aspects as I explained them, but we will have to agree to disagree on those. My intent on writing as I did was to say that there are others reading as well and at times it is helpful to "speak to the group" even though it's in a direct reply to a single person.
I do think that I made statements in earlier posts that explained my position such as the one on the nonlinear aspect being a separate issue from my focus of discussion in the thread.
To quote myself that was in response to Danny, not you:
As for the topic, I have focused on a more narrow aspect, primarily the CSD and the FR in general. If you would re-read my first post on page 1, you'll see what drew me into it in the first place. It was not based on the first posters questions, it was as I just said:
I'm not trying to spin. This is what I said and I tried to explain why.
Dave
PB2 said:
I was not intending to be offensive and since you took it as such, for that I apologize. At times my writing is not as clear as I would like it to be. I do still have disagreements with your take on some aspects as I explained them, but we will have to agree to disagree on those. My intent on writing as I did was to say that there are others reading as well and at times it is helpful to "speak to the group" even though it's in a direct reply to a single person.
I do think that I made statements in earlier posts that explained my position such as the one on the nonlinear aspect being a separate issue from my focus of discussion in the thread.
To quote myself that was in response to Danny, not you:
As for the topic, I have focused on a more narrow aspect, primarily the CSD and the FR in general. If you would re-read my first post on page 1, you'll see what drew me into it in the first place. It was not based on the first posters questions, it was as I just said:
I'm not trying to spin. This is what I said and I tried to explain why.
Dave
ucla88 said:
If your comments are for others then *please* direct them to someone else, I find your comments offensive. You think that I cannot distinguish between linear and non-linear, hmmm no one else here seems to think that in my long thread on amplifiers, or my contributions to other threads, so where do you get off making this claim? The fact is rather, that you seem to think you can divide them, oh let's look at the linear case now, now flip a switch and look at the non-linear. That is not the nature of real systems, looking at the linear case is an approximation. Now I'll stop reading you thanks, go ahead and try to get the last word in, you can have it.
Just let me add that it seems that we are reading different data for the SEAS drivers, I don't see any support of your claims above.
The L18 alone does not support your claim that the cone breakup does not add it's own distortion. It may be at HF, but the fact is that it is there. A few here seem to have trouble understanding how breakup can produce harmonics. Do you think a bell rings as a sine wave, no it has harmonics and so do many cones with break up modes. This is easily seen through careful inspection of the L18 FR and distortion data, and there are many other examples that I've been looking at for over 20 years. Now, I have no interest in an unproductive debate about this, it is my view, since you make discussion difficult I'm going to say take it or leave it. I did not bring it up earlier because it was not required to make my point.
Now I really hope to drop this, I just don't have the time for a back and forth.
Why don't we agree to disagree.
Pete B.
Now that the various (Designer) participants have beaten each other to death, can I make the following points for the average DIY'er
1) Any published Frequency Response curve alone is inadequate to make a choice of driver- the resolution is not good enough, and certainly not adequate to indicate all problem areas.
2) the CSD helps considerably to fill in the missing data, and by showing information in a way that is easy to recognise, so are useful even if some areas are suspect, or duplicate info in the FR..
3)) Impulse testing adds even more information, and we ignorant peasants need all the help we can get
4) ALL published data is suspect, and at best should be considered as useful for comparision purposes when taken by the same person under the same conditions, a point Zaph makes often on his site.
5) All designs are trade-offs, and your choice will probably not be mine
So to all those that pubish information, the more information you provide, (particularly on conditions of testing, the better.
1) Any published Frequency Response curve alone is inadequate to make a choice of driver- the resolution is not good enough, and certainly not adequate to indicate all problem areas.
2) the CSD helps considerably to fill in the missing data, and by showing information in a way that is easy to recognise, so are useful even if some areas are suspect, or duplicate info in the FR..
3)) Impulse testing adds even more information, and we ignorant peasants need all the help we can get
4) ALL published data is suspect, and at best should be considered as useful for comparision purposes when taken by the same person under the same conditions, a point Zaph makes often on his site.
5) All designs are trade-offs, and your choice will probably not be mine
So to all those that pubish information, the more information you provide, (particularly on conditions of testing, the better.
Hi Pete,
I think we can agree to disagree. I made my comments above because I don't think I'm going to change you or Danny's mind. However, a lot more folks read and participate. I throw my view out, not only to beat a dead horse, but so the lurkers can read both viewpoints and make up their own mind.
I've interpreted the data and math one way, you another. And that's ok. Everyone else can read the posts, the references, and seek a deeper understanding on their own.
mark
I think we can agree to disagree. I made my comments above because I don't think I'm going to change you or Danny's mind. However, a lot more folks read and participate. I throw my view out, not only to beat a dead horse, but so the lurkers can read both viewpoints and make up their own mind.
I've interpreted the data and math one way, you another. And that's ok. Everyone else can read the posts, the references, and seek a deeper understanding on their own.
mark
Well, I think you guys have had plenty of time to answer those two simple questions:
It looks like the only accurate prediction was mine:
What's the matter fellows? No one wanted to step up to the plate. Zaph? No attempts?
First it was... We can't tell anything from the graphs because of the third octave smoothing.
So then I post the curves with no smoothing, and still nothing.
Here, I'll give you guys an easy one.
Two 8" woofers. Measured at 1 meter. Both mounted in the same box (9" wide and 14" tall). No smoothing applied.
Obviously one has more stored energy than the other.
One simple question this time.
1) Does the frequency response tell you how much stored energy is present?
It looks like the only accurate prediction was mine:
What's the matter fellows? No one wanted to step up to the plate. Zaph? No attempts?
First it was... We can't tell anything from the graphs because of the third octave smoothing.
So then I post the curves with no smoothing, and still nothing. Here, I'll give you guys an easy one.
Two 8" woofers. Measured at 1 meter. Both mounted in the same box (9" wide and 14" tall). No smoothing applied.
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.
Obviously one has more stored energy than the other.
One simple question this time.
1) Does the frequency response tell you how much stored energy is present?
Unreliable measurements
Danny, this is nothing but attempts at "gotcha". First you post tainted measurements, more than once, now you post measurements that are questionable again. The first set started at almost 0ms going to about 4ms. Now these start at 2.88ms and go to 7.85ms. There's no consistency and no way to believe that these are proper. It's a waste of time debating you. This is characteristic of your tactics. They are not meant to enlighten.
You will not respond to direct question and comments, at times you just start all over again as you have here. You used this tactic previously. I know that I have no confidence in your posted measurements at this point. That's probably one reason why Mark and zaph stop. As I will now as well with this final post.
You find useful data in a CSD. The set of measurements that you posted do not include enough time data to make them useful were I to care about the CSD. They have zero impact on my design decisions because I can see everything I need to see in the FR, as I said. Supplement that with good distortion measurements, such as those that zaph and Mark provide and there's no need for any CSD of any kind.
The CSD has no bearing as much because the CSD of a raw driver is almost immaterial to the final result. Optimizers work on the FR. The designer chooses an the bandpass Fc's to mitigate known distortion issues and FR anomolies. The CSD with a well-designed XO will be nothing like the raw one. I will be making that point evident when I put up my page on the 2-way I'm playing with now. I hope to get something initial this weekend, but if not it will go up soon. It will belie much of what you originally were claiming. And as I said, to satisfy those who just must see them, I'll even be providing CSDs. But those CSDs will probably surprise those who put faith in CSDs as providing some sort of exceptional insight into the driver. They don't if you know what you're doing. That's not meant to offend, it's just how I feel.
This is my last post here. I have too many other things to do to waste more time, which is unfortunately what this now is. Claim whatever victory you want and congratulations.
Dave
Danny, this is nothing but attempts at "gotcha". First you post tainted measurements, more than once, now you post measurements that are questionable again. The first set started at almost 0ms going to about 4ms. Now these start at 2.88ms and go to 7.85ms. There's no consistency and no way to believe that these are proper. It's a waste of time debating you. This is characteristic of your tactics. They are not meant to enlighten.
You will not respond to direct question and comments, at times you just start all over again as you have here. You used this tactic previously. I know that I have no confidence in your posted measurements at this point. That's probably one reason why Mark and zaph stop. As I will now as well with this final post.
You find useful data in a CSD. The set of measurements that you posted do not include enough time data to make them useful were I to care about the CSD. They have zero impact on my design decisions because I can see everything I need to see in the FR, as I said. Supplement that with good distortion measurements, such as those that zaph and Mark provide and there's no need for any CSD of any kind.
The CSD has no bearing as much because the CSD of a raw driver is almost immaterial to the final result. Optimizers work on the FR. The designer chooses an the bandpass Fc's to mitigate known distortion issues and FR anomolies. The CSD with a well-designed XO will be nothing like the raw one. I will be making that point evident when I put up my page on the 2-way I'm playing with now. I hope to get something initial this weekend, but if not it will go up soon. It will belie much of what you originally were claiming. And as I said, to satisfy those who just must see them, I'll even be providing CSDs. But those CSDs will probably surprise those who put faith in CSDs as providing some sort of exceptional insight into the driver. They don't if you know what you're doing. That's not meant to offend, it's just how I feel.
This is my last post here. I have too many other things to do to waste more time, which is unfortunately what this now is. Claim whatever victory you want and congratulations.
Dave
If it shows that you can't back up what you've claimed than for you I guess it is.
Not true again. Measurements posted were just as I intended.
Not true again. Consistent with industry standards.
If you'd have put two and two together when I said they were near field measurements then the gated time window would have made sense to you.
Yep, the gated time window started just before the arrival of the impulse response and ended right before any first reflections from the room. You got a clean 5ms time window to view. Is that not enough for you? It will be for this example.
My measured data will show that not to be true as well.
In some areas, I agree, but not so in other circumstances.
Good luck with all of that Dave. I look forward to hearing it at the next big audio show.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.