Hi,
besides the questionability of listenability of distortion values, I just can´t see what and where the nonlinear distortions of the dipole woofer are? To measure a cheap driver which is inherently nonlinear...pushing that driver with amounts of equing in a configuration which it never was intended to be put in is asking for distortion. So, which part of the distortion -of which a great part probabely stems from the measurement setup itself- is driver related and which part is dipole related? No comparison is made to a CB- or BR-system which could at least give a hint of what could be going on here.
A correct measurement isn´t as simple as implied from the threadstarter. Painting nice curves on the other hand is one of the easiest things with modern measurement equipment. The difficult thing is to measure right and to get hard and serious numbers that contain real information.
Since a correct reliable measurement isn´t easy and two measurements with two different setups are hardly comparable at all and since it showed that the distortion values don´t tell much valueable about the sound of a system, it´s no wonder that theses measurements don´t play any important role in teh design of a speaker.
jauu
Calvin
besides the questionability of listenability of distortion values, I just can´t see what and where the nonlinear distortions of the dipole woofer are? To measure a cheap driver which is inherently nonlinear...pushing that driver with amounts of equing in a configuration which it never was intended to be put in is asking for distortion. So, which part of the distortion -of which a great part probabely stems from the measurement setup itself- is driver related and which part is dipole related? No comparison is made to a CB- or BR-system which could at least give a hint of what could be going on here.
A correct measurement isn´t as simple as implied from the threadstarter. Painting nice curves on the other hand is one of the easiest things with modern measurement equipment. The difficult thing is to measure right and to get hard and serious numbers that contain real information.
Since a correct reliable measurement isn´t easy and two measurements with two different setups are hardly comparable at all and since it showed that the distortion values don´t tell much valueable about the sound of a system, it´s no wonder that theses measurements don´t play any important role in teh design of a speaker.
jauu
Calvin
A church organ concert shows that even 16 Hz are audible, if they are loud enough.gainphile said:20Hz shouldnt be audible
Human hearing is very insensitive to very low and very high frequencies. The Phon aka dB(A) curve shows that the listening threshold of 3 dB(A) at 1kHz is actually 3 dB, while at 20 Hz it is already above 70 dB ( but still 3 dB(A) ). Few speakers can deliver that amount of SPL at 20 Hz although they are visibly moving, and that is probably the reason, why many people think 20 Hz are inaudible.
It takes 95-100 dB to reach the 50-60 dB(A) level that is supposed to be conversational SPL. How many 15"-woofers and how much equalizing does that require?
Concert levels are somewhere above 120 dB for 20 Hz. Impressive, but still bearable, while 100 dB at 1 kHz already physically hurt your ears.
And all that is the reason, why the human ear is also less sensitive to distortion at low freqeuncies.
dantheman said:Ooops. I've got way off topic here and I appologize.
However, again directed at Dr. Geddes or anyone else who knows more than I: so a system Q of say 0.5 won't tighter or faster than a Qtc of 2.5? Maybe I'm confusing terms.
Thanks again,
Dan
The rise time of a transient is virtually independent of the Q, but the higher the Q the more the signal will ring out. But to call this "tighter" or "faster", I don't see the connection. And you have to remember that subs are always closely coupled to the room so discussing subs without discussing room modes is pointless. In actuality, the "Q" of a sub gets almost completely lost in the "Q"s of the room modes. The room will dominate the situation completely. So its not the Q of the subs that matters but how the subs work with the room that matters. As far as subs go, I have a slight preference for bandpass because they have an acoustic LP function and this tends to help keep down the harmonics that can make the sub localizable. But beyond that I have no preference in size, type, Q, resonance, etc. Only how many are used, where are they placed and how are they adjusted.
Hi Earl,
I agree and disagree with what you have stated. When we start looking at bass in the free field, there is little different in the transient response between a Q =1.0 and a Q = 0.5 system when looking at a bass transient. I say this because the additional ringing of the Q = 1 system is far less that the "ringing" of say a kick drum or plucked bass. The biggest difference is that the Q = 0.5 system will be down 6dB at Fc and the Q = 1 will be flat at Fc with a slight rise above Fc. If we consider the kick drum or plucked bass as a high Q system, like Q = 10 or 20 (or even higher) then the natural decay of the bass note will be so much longer in duration than a Q = 1 decay that the decay of the reproduced note can be considered as quasi-steady behavior relative to the woofer system. Thus, for a note close to the Fc of the woofer the major difference (in free space) is the amplitude of the note, as affected by the woofer system Q. As an example see the figure below. Top is a burst at Fc for a q = 1 system and the lower is the same burst for a Q = 0.5 system.
Yes, there is a difference at the end of the burst due to the difference in decay of the Q = 0.5 and Q = 1 systems, but it is quite insignificant compared to the difference in amplitude while the input is applied.
IMO, this difference in amplitude around Fc is what is interpreted as "fast" bass. IMO, fast bass is noting more than bass with reduced amplitude of the fundamental.
Now, move into a room and the same thing can occur. Assuming that Fc is above the room fundamental, then the difference s between the Q = 0.5 and Q = 1 woofer is that modes which are excited around the Fc of the woofer system are excited to a much lesser degree by the Q = 0.5 system since its output power is 1/4 that of the Q = 1.0 system (note I said power, not SPL).
Now, if the woofer system is in a closed room I believe there are advantages to woofer with 2nd order Q = 0.5 alignment and Fc close to the room fundamental. This is because the roll off of the Q =0.5 woofer will compliment the rise in SPL due to room pressurization more naturally.
I agree and disagree with what you have stated. When we start looking at bass in the free field, there is little different in the transient response between a Q =1.0 and a Q = 0.5 system when looking at a bass transient. I say this because the additional ringing of the Q = 1 system is far less that the "ringing" of say a kick drum or plucked bass. The biggest difference is that the Q = 0.5 system will be down 6dB at Fc and the Q = 1 will be flat at Fc with a slight rise above Fc. If we consider the kick drum or plucked bass as a high Q system, like Q = 10 or 20 (or even higher) then the natural decay of the bass note will be so much longer in duration than a Q = 1 decay that the decay of the reproduced note can be considered as quasi-steady behavior relative to the woofer system. Thus, for a note close to the Fc of the woofer the major difference (in free space) is the amplitude of the note, as affected by the woofer system Q. As an example see the figure below. Top is a burst at Fc for a q = 1 system and the lower is the same burst for a Q = 0.5 system.
Yes, there is a difference at the end of the burst due to the difference in decay of the Q = 0.5 and Q = 1 systems, but it is quite insignificant compared to the difference in amplitude while the input is applied.
IMO, this difference in amplitude around Fc is what is interpreted as "fast" bass. IMO, fast bass is noting more than bass with reduced amplitude of the fundamental.
Now, move into a room and the same thing can occur. Assuming that Fc is above the room fundamental, then the difference s between the Q = 0.5 and Q = 1 woofer is that modes which are excited around the Fc of the woofer system are excited to a much lesser degree by the Q = 0.5 system since its output power is 1/4 that of the Q = 1.0 system (note I said power, not SPL).
Now, if the woofer system is in a closed room I believe there are advantages to woofer with 2nd order Q = 0.5 alignment and Fc close to the room fundamental. This is because the roll off of the Q =0.5 woofer will compliment the rise in SPL due to room pressurization more naturally.
Calvin said:
I agree, we need more rigorous examination of these sorts of things. But criticism alone isn't enough. We have to start somewhere, and the above measurements count as 'somewhere'.
Attached is a good article on distortion of test microphones, particularly the wm61a. Their distortion isn't negligible. A lot of amps have appreciable distortion in the lower frequencies, so they too have to be factored in.
Test conditions are pretty important, and without them, there is no way for other people to use the results. But that doesn't mean the measurements aren't useful for the person who made them. The main reason Zaph's measurement are so useful for the rest of us is their all being done under the same conditions.
I hope people keep exploring this. As far as I can tell, loudspeakers are the largest source of linear and non-linear distortion in the reproduction chain, and if we want to improve reproduction, loudspeaker distortion is the place to look.
John K.
I don't disagree with what you are saying but I think that it is an oversimplification of how LFs work in real rooms and how our hearing works at LF. Our hearing is simply incapable of responding to a signal like an oscilloscope does so I don't see any relavence between the oscilloscope pictures and hearing, especially in a small room with discrete modes.
One can simply not remove the room and our hearing limitations from the problem at LFs. We detect LF very slowly, usually many periods of the signal. And the modes in the room interact with the woofer in complex ways that depend on so many factors about the room and the LF source layout that making generalized statements aren't going to be very useful.
To answer Marcus comment and to further the discussion, typically modes in a room are not so sharply tuned that there are "gaps" between them. When you see nulls in the response this can only be because more than one mode is being effected because single modes cannot create nulls, only the summations of two or more modes can do this. And if you did have a room where the modes are so sharply tuned that large "gaps" exist, then all bets are off because this room is going to be a disaster no matter what you do. I always assume some level of damping such that the modes basically interact with one another. Only at the very first room mode might this not be true. But our hearing acuity at 30 Hz. is nothing short of almost nonexistant. I usually start to think about and design for LFs at about the econd mode or so because its the 50 - 100 Hz octave that is critical to LF perception. 25 - 50 Hz should be there, but details about the system at these frequencies are not going to be very relevent.
But if you do want to discuss a room with a single low damped mode in this region, then I would contend that a higher Q woofer with an Fs in this region would act as a second mode which may well be the best choice. Again, I'd have to look at the specific problem not a generalization. But in general, I don't worry so much about 25 - 50 Hz as I do 50 Hz. - 200 Hz.
You simply cannot get away from the fact that at LFs in small rooms there is a modal field excited by one or more sources and only this total combination of effects can be considered. In this context the "Q" of the woofer is not going to be significant given the various Q's and resonances of the rooms modes.
I have said this so many times, but I'll say it again. Obsessing about the Q, the type, the "tuning"; all of that is secondary to the number, the location, and the setup. I don't care what kind of subs you give me (within reason), as long as there are at least three and they have gain, phase and LP cutoff control and I will get good bass - as good as anyone else will get worrying about all those other details. They just don't matter.
I don't disagree with what you are saying but I think that it is an oversimplification of how LFs work in real rooms and how our hearing works at LF. Our hearing is simply incapable of responding to a signal like an oscilloscope does so I don't see any relavence between the oscilloscope pictures and hearing, especially in a small room with discrete modes.
One can simply not remove the room and our hearing limitations from the problem at LFs. We detect LF very slowly, usually many periods of the signal. And the modes in the room interact with the woofer in complex ways that depend on so many factors about the room and the LF source layout that making generalized statements aren't going to be very useful.
To answer Marcus comment and to further the discussion, typically modes in a room are not so sharply tuned that there are "gaps" between them. When you see nulls in the response this can only be because more than one mode is being effected because single modes cannot create nulls, only the summations of two or more modes can do this. And if you did have a room where the modes are so sharply tuned that large "gaps" exist, then all bets are off because this room is going to be a disaster no matter what you do. I always assume some level of damping such that the modes basically interact with one another. Only at the very first room mode might this not be true. But our hearing acuity at 30 Hz. is nothing short of almost nonexistant. I usually start to think about and design for LFs at about the econd mode or so because its the 50 - 100 Hz octave that is critical to LF perception. 25 - 50 Hz should be there, but details about the system at these frequencies are not going to be very relevent.
But if you do want to discuss a room with a single low damped mode in this region, then I would contend that a higher Q woofer with an Fs in this region would act as a second mode which may well be the best choice. Again, I'd have to look at the specific problem not a generalization. But in general, I don't worry so much about 25 - 50 Hz as I do 50 Hz. - 200 Hz.
You simply cannot get away from the fact that at LFs in small rooms there is a modal field excited by one or more sources and only this total combination of effects can be considered. In this context the "Q" of the woofer is not going to be significant given the various Q's and resonances of the rooms modes.
I have said this so many times, but I'll say it again. Obsessing about the Q, the type, the "tuning"; all of that is secondary to the number, the location, and the setup. I don't care what kind of subs you give me (within reason), as long as there are at least three and they have gain, phase and LP cutoff control and I will get good bass - as good as anyone else will get worrying about all those other details. They just don't matter.
In EE parlance modes are mathematical poles, not zeros. A single mode could only have gain. But if there are two or more modes then they can cancel and a "zero" is created, i.e. a null. But a solo mode cannot have a null. The degree to which there is a null depends on the phases and amplitudes between the interacting modes. It can be very deep as in complete cancellation or only just a shallow blip.
I think that this also define my meaning of a "mode" and a "null".
I think that this also define my meaning of a "mode" and a "null".
The standing wave is the Eigenmode (the shape) and the Eigenvalue is the frequency of the resonance. These terms come from the math. So basically a "mode" is both the shape and the frequency. A standing wave is a "mode", but modes can also have a traveling component. The part of the mode that is absorbed is traveling and the reactive or nonabsorbed part is "standing". All modes have both parts.
I would have to disagree with Earl in that individual modes have nulls. The transfer function for a given room mode does have zeros. If you mathematically examine the in room SPL it is possible to express the SPL at a point in terms of the room modes.
Each mode contributes to all frequencies, as indicated by the summation but at different levels. Each mode has a transfer function that is basically that of a 2nd order LP filter with Q determined by the room damping characteristics. In addition, the filter will have spatially dependent gain determined by the product of the eigen function for that mode (standing wave) evaluated at the source and listening positions. Thus this LP filter representation of a mode will have a poles at the modal frequency (where k = Kn for Kn real) and will have spatially dependent zeros where egien function is zero. Since the eigen functions are cosine functions (of the form cos(n Pi x /L)) they will be pressure anti-nodes where the cos = +/- 1 and nodes (zeros) where cos = 0.
An externally hosted image should be here but it was not working when we last tested it.
Each mode contributes to all frequencies, as indicated by the summation but at different levels. Each mode has a transfer function that is basically that of a 2nd order LP filter with Q determined by the room damping characteristics. In addition, the filter will have spatially dependent gain determined by the product of the eigen function for that mode (standing wave) evaluated at the source and listening positions. Thus this LP filter representation of a mode will have a poles at the modal frequency (where k = Kn for Kn real) and will have spatially dependent zeros where egien function is zero. Since the eigen functions are cosine functions (of the form cos(n Pi x /L)) they will be pressure anti-nodes where the cos = +/- 1 and nodes (zeros) where cos = 0.
Two points John
First, the equation that you show is that for a 2nd order bandpass (you forgot about the "k" out in front which is in the numerator), not a Highpass.
Second, in general, this equation is complex with complex eigenvalues and the eigenmodes. Only in the idealized case where the eigenvalues are purely real (no absorption) and hence the eigenmodes purely real will the product of the two eigenmodes in the numerator have zeros (real cosines). In an actual case where the room has damping the eigenmodes and eigen frequencies are complex and the product in the numerator will not be zero for a single mode (the modes are similar to complex cosines which don't have zeros). The summation over many modes CAN be zero or near zero when the modes intereact. But this is nit picking. My point was that the zeros that we see in the frequency responses for "real" rooms are more likely the result of modal interaction when there is a decent modal density than a spatial nodal line, although this later aspect cannot be rulled out at the very lowest frequencies. It is highly unlikely when there is a decent modal density.
Far too many concepts in acoustics are taken from the equation that John shows which is virtually never solved with absorption present and then usually when the room is ideally rectangular with no doors or windows or furniture. Its a good thing to use equation like this, but we always have to keep in mind that they all have assumptions.
In my PhD thesis and after I graduated I did a lot of work on calculating room modes with absorption where the eigen modes and values are complex. In this case one can plot out the energy flow in the room, which cannot be done in the "real" case because the energy flow is zero everywhere. This can make a big difference, but mostly at points between the modes. At a strong mode the high modal excitation dominates the problem and the static potential energy is much larger than the kinetic energy flow. Hence, the solutions for the room are most accurate at low damping and near the modal peaks and least accurate at higher damping and away from a modal peak.
First, the equation that you show is that for a 2nd order bandpass (you forgot about the "k" out in front which is in the numerator), not a Highpass.
Second, in general, this equation is complex with complex eigenvalues and the eigenmodes. Only in the idealized case where the eigenvalues are purely real (no absorption) and hence the eigenmodes purely real will the product of the two eigenmodes in the numerator have zeros (real cosines). In an actual case where the room has damping the eigenmodes and eigen frequencies are complex and the product in the numerator will not be zero for a single mode (the modes are similar to complex cosines which don't have zeros). The summation over many modes CAN be zero or near zero when the modes intereact. But this is nit picking. My point was that the zeros that we see in the frequency responses for "real" rooms are more likely the result of modal interaction when there is a decent modal density than a spatial nodal line, although this later aspect cannot be rulled out at the very lowest frequencies. It is highly unlikely when there is a decent modal density.
Far too many concepts in acoustics are taken from the equation that John shows which is virtually never solved with absorption present and then usually when the room is ideally rectangular with no doors or windows or furniture. Its a good thing to use equation like this, but we always have to keep in mind that they all have assumptions.
In my PhD thesis and after I graduated I did a lot of work on calculating room modes with absorption where the eigen modes and values are complex. In this case one can plot out the energy flow in the room, which cannot be done in the "real" case because the energy flow is zero everywhere. This can make a big difference, but mostly at points between the modes. At a strong mode the high modal excitation dominates the problem and the static potential energy is much larger than the kinetic energy flow. Hence, the solutions for the room are most accurate at low damping and near the modal peaks and least accurate at higher damping and away from a modal peak.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.