Multiple Small Subs - Geddes Approach

[soongsc]

Originally posted by gedlee


Soongsc

Sorry you are taking such heat, but you are well known for speaking about things with an air of authority that is apparently unjustified. Questioning might be a better tact in your situation that making such bold statements.

Time alignment is important in that it has a major effect on the polar pattern at the crossover. One finds that obtaining a good polar response at crossover necessitates a near time alignment of the drivers to come out right. But the need is for a good polar response and not for some belief that phase matching or phase itself is an audible characteristic. What is audible is the poor polar response that one gets when the two drivers do not have the minimum time differential between their signals.

I understand your point concerning time alignment.

I don't mind taking heat. Most people don't have the knowledge to prove me wrong, thus the conflict. I think any new point of view will face this kind of rejection. The ultimate decision is made by customers, optinions of which I always spend time to try to understand rather than reject. I do communicate with audiophiles, and in many cases, they find similar trends that I predict. Some of which are related with horns/guides.

Anyway, anyone with an open mind will learn faster and have more fun than those that don't. I think we will be making some news a few months from now.

[gedlee]

Your analysis is lost on me - sorry. I don't follow it at all. And some phrases are not even sentences like:

"As a first example lets get te misleading term "decay on resonance frequency" right. "

I never used the phrase "decay on resonance frequency" so I have no idea what the quotes mean. The rest of the discussion is just as illusive. If you want to have an intelligent discussion about the math thats fine, but you at least need to write in complete sentences and in a comprehensive manner.

[gedlee]

Originally posted by soongsc
Most people don't have the knowledge to prove me wrong, thus the conflict.

I do. I have. You ignore it. That's the conflict.

[xpert]

Originally posted by gedlee
Your analysis is lost on me - sorry. I don't follow it at all. And some phrases are not even sentences like:

"As a first example lets get te misleading term "decay on resonance frequency" right. "

I never used the phrase "decay on resonance frequency" so I have no idea what the quotes mean. The rest of the discussion is just as illusive. If you want to have an intelligent discussion about the math thats fine, but you at least need to write in complete sentences and in a comprehensive manner.

Earl,

You wrote:

"For example, lets say we have a mode at 51 Hz and we drive it at 49 Hz. In the steady state the enclosure has 49 Hz in the sound field, but when I turn off this source, the sound field slowly (relative to the decay rate) changes from 49 Hz to 51 Hz. No electronic system can do this and to call this kind of phenomina "minimum phase" is certainly not correct. Acoustics is not electronics and behaves quite differently."

Sorry for being insistently. What You wrote above is in no way supported by contemporary knowledge (remember my degree in physics?). Neither is the maths of mechanical systems in any way different from that of electronic systems in this scope. Nor does the decay mechanism drift from one frequency to an other. Neither abrupt nor gardually. If so it could never be called linear, that simple.

Again, there is no relation to any phase, especially no so called minimum phase.

You have to admit a misstake. Anybody is prone to such anytime, nearly. So, as You are a native speaker please forgive my incomplete whackey so called sentences. I'll try to live up to Your expectations. I wish You would live up to mine.

thanks a lot

[gedlee]

Originally posted by xpert



Earl,

You wrote:

"For example, lets say we have a mode at 51 Hz and we drive it at 49 Hz. In the steady state the enclosure has 49 Hz in the sound field, but when I turn off this source, the sound field slowly (relative to the decay rate) changes from 49 Hz to 51 Hz. No electronic system can do this and to call this kind of phenomina "minimum phase" is certainly not correct. Acoustics is not electronics and behaves quite differently."

Sorry for being insistently. What You wrote above is in no way supported by contemporary knowledge (remember my degree in physics?). Neither is the maths of mechanical systems in any way different from that of electronic systems in this scope. Nor does the decay mechanism drift from one frequency to an other. Neither abrupt nor gardually. If so it could never be called linear, that simple.

Again, there is no relation to any phase, especially no so called minimum phase.

You have to admit a misstake. Anybody is prone to such anytime, nearly. So, as You are a native speaker please forgive my incomplete whackey so called sentences. I'll try to live up to Your expectations. I wish You would live up to mine.

thanks a lot

Lets forget about the minimum phase comment, thats a whole other can of worms.

If you dod not recall I too have a PhD in Physics, in Theoretical Physics to be precise, and my math skills are not rudimentary.

I stand by my statements as correct as shown by Morse in said text and they are not wrong simply because YOU say that they are wrong. The burdon of proof is on you since I have attached my proof from the work that I described.

[gedlee]

The system is not letting me post the page, but it can be seen here www.gedlee.com/downloads/morse_pg429.pdf

[xpert]

Originally posted by gedlee



Lets forget about the minimum phase comment, thats a whole other can of worms.

I stand by my statements as correct as shown by Morse in said text and they are not wrong simply because YOU say that they are wrong. The burdon of proof is on you since I have attached my proof from the work that I described.

Hi,

You cited Morse.

http://www.gedlee.com/downloads/morse_pg429.pdf

I read it. Nothing on that page supports this:

"I turn off this source, the sound field slowly (relative to the decay rate) changes from 49 Hz to 51 Hz."

The citation lacks all definitions one has to know to understand what it is about. What is it aiming for? Example given, how are thoses series defined? Your proof is none. It is Your turn to show explicitely how Your claims can be derived from Morse (if Morse is an authority, which he is).

I'm quite sure that You didn't mean that decay MAKES an other frequency. Decay changes proportions between frequencies that have ever been there. Or in case of switching off a source, are part of the signal. To switch off a source is a signal.

Morse himself mentions a temporal symmetry between switching on and off. Woud You turn on the source, the sound field slowly changes from 51Hz to 49Hz? I'm quite sure all the argument is about some sloppy wording.

so long

[soongsc]

Originally posted by gedlee


I do. I have. You ignore it. That's the conflict.

I think in areas that I did find myself incorrect, I did openly admit it, not face saving remarks. If I diagnose my own data which show different trends, of course I beleive my data as you would believe yours. We just never got down the the exact conditions we each used. But bear in mind that I am constantly generating new data an looking at much more different test conditions than others have made public. The lack of comparible tests always leaves room for interpretation and opinion.
One thing that I always try to stick to is to focus on the issue, not on the person. Therefore, I accept the fact that disagreements exists, which is very common, and a fact in life.
Having worked with a few large companies in the US, many opinions of mine initially get ignored, but normally it gets respected a year or to down the road. So as you can see, I have no problem with disagreements.

[john k...]

My point is that MP has nothing to do with how the system acts after the signal is removed. MP is just a mathematical relationship with response the the transfer function. I believe form the many discussions with Earl on this subject that Earl disagrees with the MP concept because in acoustics the transfer function is a four-dimensional function of position and frequency, x,z,y,f, whereas for an electrical system the transfer function is a one dimensional function of F at the output. My point has always been that for a given position, x,y,z, the resulting spatial transfer function is either MP or not MP. The TF changes with position, and relative to that position the response again, may or may not be MP. The point is that the position, x,y,z defines the transfer function. At that point the transfer function then can be determined to be MP or not.

Does it matter? That depends on your outlook. From the system point of view, the only thing MP does is to allow us to recognize the amplitude correction implies phase /time correction. If the system is non MP that doesn't hold.

The other point about MP is that if two tweeters which measure on axis exactly the same and are MP, then substitution of either tweeter in a system which was designed for flat on axis response with one of the tweeters will have flat response when the other is substituted. If the MP relationship continues to hold for off axis positions and both tweeters have the same polar response, then the system will also have the same polar response with either tweeter. If MP doesn't hold, you would have no idea what the system would look like with the second tweeter, on or off axis. Now it is not necessary to for CAD software to require MP or even consider it. But it is inherent in the design process because it is inherrent that for a dirver the response, T(x,y,z,f) for any position x,y,z is typically MP, or MP plus some constant time delay.

But when we talk about in room response, low frequency or otherwise MP really isn't the issue in any case. We are stuck with an 4-dimensional transfer function, T(x,y,z,f) and there is no way to make T(x,y,z,f) have some predefined characteristic by applying a single equalization function Eq(f). T(x,y,z,f) * Eq(f) is still a three dimensional function of space. Even with multiple subs and multiple eq, T(x,y,z,f) would equal T1(x,y,z,f) * Eq1(f) +T2(x,y,z,f) * Eq2(f) +...... + Tn(x,y,z,f) * Eqn(f) where there are up to n woofers and equalization functions. There may be ways to optimize the position of each woofer and its corresponding eq to have relatively smooth (low frequency) response over some range of positions, but the only true way to have uniformly smooth response over a wide regions if to eliminate the x,y,z dependences in the room transfer function(s).

[xpert]

Originally posted by john k... ... the only true way to have uniformly smooth response over a wide regions if to eliminate the x,y,z dependences in the room transfer function(s). [/B]

Hi John,

You do not aim for that, do You? The ansatz of Earl Geddes seems quite convincing to me. Take what is affordable (3 subs + 2 mains = 5 sources) and follow an algorithm to yield at least a second best solution. To some extent the second best is equivalent to the best, when the limited human hearing skills are taken into consideration.

This is straight forward engineering. Some kind of sound field generation would be more satisfying to the scientist. Alas it is not practical due to limited budget, limited time, limited knowledge.

cheers

[gedlee]

Originally posted by xpert

You cited Morse.

http://www.gedlee.com/downloads/morse_pg429.pdf

I read it. Nothing on that page supports this:

"I turn off this source, the sound field slowly (relative to the decay rate) changes from 49 Hz to 51 Hz."

I'll stand by my statement and claim that it is supported by Morse. If you need more information to understand it then LOOK IT UP.

[gedlee]

John

I don't disagree with a anything that you say, except maybe "MP has nothing to do with how the system acts after the signal is removed". Decay is part of the system response. Since the steady state response defines the impulse response and the impulse response defines the response for all time and all signals, this statement does not seem like it could be correct.

At any rate you understand very clearly the limitations of the concepts for acoustics and why I don't bother with it. If it works for you fine.

[xpert]

Originally posted by gedlee


I'll stand by my statement and claim that it is supported by Morse. If you need more information to understand it then LOOK IT UP.

Earl,

sorry. You are the one to understand things. I would like to help You. Please show us how You could come to that forementioned misconceptions.

It's your turn now.

[xpert]

Originally posted by gedlee
John

I don't disagree with a anything that you say, except maybe "MP has nothing to do with how the system acts after the signal is removed". Decay is part of the system response. Since the steady state response defines the impulse response and the impulse response defines the response for all time and all signals, this statement does not seem like it could be correct.

At any rate you understand very clearly the limitations of the concepts for acoustics and why I don't bother with it. If it works for you fine.

Quatsch,

The term "the system acts after the signal is removed" is an unlucky wording in this context. This wording is from Morse whom You brought in as an authority. You shouldn't take john_k into account for that.

If You are talking about a transfer function there surely is NO difference betweeen steady state and any other state. This very statement is from me, it was not Yours. Please don't use my argumentation against john_k, whilst You try to make him resonsible for Your own misleading wording.


As far as I have understood Morse from Your very (brief) citation, he does NOT support any of Your statements. First of all, what is decay within a cavity described by eigenmodes (may be) and some sort of damping on the reflective surfaces? Do You try to derive Your claims from an understanding of that situation?

What is different compared to, example given a one dimensional oscillator mass/compliance type, electric circuit bild up with reactive impedances or much simpler an organ pipe? The concept of eigenmodes encourages to look at the whole systems as a sum of parts that superimpose linearly.

Do You try to tell us, that the eigenmode concept is not valid here? To only flutter with some papers/books is not very kind. Do You think john_k does understand Your objections? If not why not try to explain the basic concept behind Your doubt.

Thank You

[john k...]

Originally posted by gedlee
John

I don't disagree with a anything that you say, except maybe "MP has nothing to do with how the system acts after the signal is removed". Decay is part of the system response. Since the steady state response defines the impulse response and the impulse response defines the response for all time and all signals, this statement does not seem like it could be correct.

At any rate you understand very clearly the limitations of the concepts for acoustics and why I don't bother with it. If it works for you fine.

Earl,

What I was saying about MP having noting to do with decay was in regards to a system undergoing forced oscillation at some frequency and then when the forcing function is removed it will decay with oscillation at its natural frequency if the system is sufficiently underdamped. That characteristic is independent of whether the system is MP or not. That's all. Don't read anyting into it.


xpert:

At this point, what ever satisfies the listener is fine with me. Bass in a small room is never going to be "right" over a wide area. Different methods will achieve more or less satisfying results that other approaches. We can argue the merits of one approach over another, but ever floats your boat is good enough for me.

Frankly, in my experience, I find a deep null some where between 30 to 50 Hz region most dissatisfying. I'm personally more tolerant of a little boom compared to a complete suck out. For example, a recording of a big bass drum without the low frequency part sound a lot worse that the same thing with a little boom. I'd rather hear it wrong than not at all. :0

[gedlee]

Originally posted by john k...


Earl,

What I was saying about MP having noting to do with decay was in regards to a system undergoing forced oscillation at some frequency and then when the forcing function is removed it will decay with oscillation at its natural frequency if the system is sufficiently underdamped. That characteristic is independent of whether the system is MP or not. That's all.

John

Seemed like you were saying more than that at the time.

xpert:

Your arguument is not concrete enough for me to discuss. I have made my argument by simply pointing out the discussion in Morse. You seem unwilling to go there to understand the points made there so I don't see any point in continuing. If you read Morse, and can find an error, or a point of disgreement, then I would be happy to discuss it with you. But I don't think that it is my job to explain it to you when it is readily available (and something that anyone who wants to argue acoustics should have read already - Morse is a (the) fundamental text on the subject).

[xpert]

Originally posted by gedlee
xpert:
Your arguument is not concrete enough for me to discuss. I have made my argument by simply pointing out the discussion in Morse. [/B]

Sorry Earl,

I won't discuss a clear misunderstanding of Yours without Your collaboration. You've got Morse wrong all to clear. You don't differ to much from audiophiles in using matrial wording. But You avoid to answer to concrete queries on what Your talking is all about. Thus Your thinking remains arcaine. Wellcome to the club.

I will cite myself, and - give up.

s far as I have understood Morse from Your very (brief) citation, he does NOT support any of Your statements. First of all, what is decay within a cavity described by eigenmodes (may be) and some sort of damping on the reflective surfaces? Do You try to derive Your claims from an understanding of that situation? What is different compared to, example given a one dimensional oscillator mass/compliance type, electric circuit bild up with reactive impedances or much simpler an organ pipe? The concept of eigenmodes encourages to look at the whole systems as a sum of parts that superimpose linearly. Do You try to tell us, that the eigenmode concept is not valid here? To only flutter with some papers/books is not very kind. Do You think john_k does understand Your objections? If not why not try to explain the basic concept behind Your doubt.

[JPV]

Originally posted by gedlee


I'll stand by my statement and claim that it is supported by Morse. If you need more information to understand it then LOOK IT UP.

Is it not correct to say that the output of even a simple electrical lumped constants system with resonnance at 51hz will drift from 49 to 51 hz if the 49 hz input will be switched off. The switching off ( multiplication by a step function) is broad band and will contain enough energy at 51 hz to see a 51 hz decaying transient if the resonnance is sharp enough. This drift is not specific to acoustics but is pehaps more important in a system with distributed constants and propagation.


JPV

[gedlee]

Originally posted by JPV


Is it not correct to say that the output of even a simple electrical lumped constants system with resonnance at 51hz will drift from 49 to 51 hz if the 49 hz input will be switched off. The switching off ( multiplication by a step function) is broad band and will contain enough energy at 51 hz to see a 51 hz decaying transient if the resonnance is sharp enough. This drift is not specific to acoustics but is pehaps more important in a system with distributed constants and propagation.


JPV

I thought about this exact example myself and I will say that I believe that you are correct. However I'm not a EE so thats not my expertise. On the other hand, the acoustics I know well and I am 100% confident in my position and won't be goaded into an argument that makes no sense. It seems likely to me that ALL systems would have to decay at their resonant frequencies when the excitation is turned off. The excitation simply become the inital conditions for an unforced problem (i.e. a homogeneous Diff Eq.) All homogeneous Dif EQ solutions decay at the natural frequencies since there is no driven frequency.

[xpert]

Originally posted by JPV


Is it not correct to say that the output of even a simple electrical lumped constants system with resonnance at 51hz will drift from 49 to 51 hz if the 49 hz input will be switched off. The switching off ( multiplication by a step function) is broad band and will contain enough energy at 51 hz to see a 51 hz decaying transient if the resonnance is sharp enough. This drift is not specific to acoustics but is pehaps more important in a system with distributed constants and propagation.


JPV

Hi,

It might have been due to my bad English that Mr. Earl Geddes didn't get my objections right. He did get Yours wrong too.

The clue here is, he forgot about the Fourier integral. That not only the "switch" is a broadband signal but also the decay itself has to be described as broadband. Depending on the Q, not only on the switch.

Geddes: "All homogeneous Dif EQ solutions decay at the natural frequencies since there is no driven frequency."

If it does decay, losses are part of the DifEq. Then the solution is a Fourier integral instead of a spot frequency. That easy.

Replace "at" by "around", eliminate the swift idea, replace it by convolution of broadband signals with a transfer function.

For me I found out, that I don't know how to solve the wave equation with lossy walls if the systems is more than 0-dimensional. The losses have to be taken into consideration right from the start because a possible steady states is an equilibrium between losses and source. Losses and source seperated. Simulation is easy, but a handy formula? Any input?

Thank You

[Rudolf]

Originally posted by xpert

Geddes: "All homogeneous Dif EQ solutions decay at the natural frequencies since there is no driven frequency."

If it does decay, losses are part of the DifEq. Then the solution is a Fourier integral instead of a spot frequency. That easy.

Replace "at" by "around", eliminate the swift idea, replace it by convolution of broadband signals with a transfer function.

Xpert,
I have to admit that I don´t fully understand the mathematics behind this decay process, but I even less understand the problem you have with Earl Geddes argument.

If I remember correctly this whole discussion started with some CSD which showed ringing to change from one frequency to another within a short time segment. It goes without saying that this change cannot take place abruptly in the moment, when the excitation has been switched off, but has to take some time. Laws of physics would not allow otherwise. In the light of this your demand to <Replace "at" by "around"> would be purely cosmetic. Everybody following this discussion from the cited CSD on will have been aware of this transitional situation.

If this is not what you are talking about, please just forget about my objection.

Rudolf

[john k...]

I think you are all arguing over semantics and not physics.

[BHTX]

Originally posted by john k...
I think you are all arguing over semantics and not physics.

Exactly. I can't believe I wasted this much time trying to figure out what the hell was going on.

[xpert]

Originally posted by Rudolf


I have to admit that I don´t fully understand the mathematics behind this decay process ... discussion started with some CSD which showed ringing to change from one frequency to another within a short time segment ... Laws of physics would not allow otherwise ...

Sorry to say, that all isn't about cosmetic or semantics. It is about some basics that should be understood in the first place. A thoroughly used wording should reflect that comprehensive understanding.

That said very homogenious DifEq has in case of losses - that is here so by definition - a solution that isn't a single frequency but a spectrum. Given a starting condition that is not null the spectrum will be constant over time, no shifts, period.

Alas, nowbody will gain an insight into signal theory by cutting MDF boards. It isn't meant to be harsh, by the way. I really would appreciate someone competent in acoustics solve the 1-dim wave equation with a source somewhere between lossy boundaries. Morse? Any handy formula?

cheers

[xpert]

Originally posted by BHTX


Exactly. I can't believe I wasted this much time trying to figure out what the hell was going on.

Hi,

The Geddes approach for smooth low bass response is worth it. And he asked for this discussion that You couldn't figure out:

Originally posted by gedlee

Well you are both wrong. Morse is one of the formost Physicists of the 20th century and doesn't get things like this wrong. Read it in his book - Morse, P.M., "Vibration and Sound" - last page of the book pg. 429 Do a "further analysis" if you dare! If you have the capability to do that! Show its wrong or recind your comments!

As I posted before, no offense. But I couldn't rescind. For me the outcome isn't to bad:

To really calculate all that decaying modes within a room is a challange. An algorithm to bettern the acoustic situation by trial is worth a thought. To validate Geddes' algorithm it should be tested in daily life. As far as I saw here it performed very well if not perfectly, not at least budget wise.

Why talk about "phase" issues at all?! From psycho acoustics it can be shown that stuff as such is by far irrelevant here. But IF one talks about it, one should avoid sloppy wording.

so long