On the other hand your position seems completely untenable to me. You seem to be claiming that two systems with exactly the same response will still sound different. How is that possible?
Simply put, I don't buy it and it certainly is not consistent with my experience, the experiences of dozens of others and the physics involved. If this is your claim, and the basis of this thread, then indeed I have nothing to add as the concept is simply not physically possible IMO.
I don't know, as I have not tried what you are saying. I am just saying what I have done for a sub that works well for me. If you want to build horn subs, then go for it. I have worked for a night club that had two C-V "Earthquake" horns. They had 18's inside them and were driven with 360w/ch. They needed EQ and to face into a corner to really work good. Away from the walls with no EQ, they preformed like any other 18" sub, but corner loaded they would flap you pant legs.
At 4'x4'x2' I have no room for them in my house, so I have a sealed box with 12db/oct EQ. I think some 18's have even greater output that 96db 1w/1m.
Generally, if using a half decent driver, in an adequate enclosure, id tend to agree with him..... At least at low power (well before the onset of distortion). The only difference will be output capabilities for a given voltage (post eq) and number of drivers. And as soon as you approach said output capabilities, all bets are off (port compression an the cone flapping in the wind below tuning for reflex boxes, etc)
Sorry, just to confirm here, are you saying that you think EQ can make up for all the differences between a horn and ported enclosure?
path length delay?
group delay?
room coupling?
directionality at a given frequency range?
Ok I understand by using multiples in the situation you can get a better result. Doesn't surprise me... but we're talking about compromises here aren't we? It's not the folded horn design cab type at fault, but the room it's placed in.
Solution... spend 3x more on drivers and x amount more on amps (depending on impedances) to fix it.
If anything the lengths we have to go to in terms of extra drivers and amplification should indicate that the horn is a far superior device to a reflex one. It's just not always practical.
I'd also like to point out that room EQ will do nothing to fix modal ringing in a compromised environment. In those sorts of environment I'd get in some rpg modex panels and still use a horn.
Forgive me if I'm confused here, I pretty much refuse to believe any two different systems have the 'same response', with EQ or without. Maybe we get similar %thd, -3db points in available spectrum, unified phase response, peak and rms max spl for a given %thd, dispersive qualities at a given frequency, group delay across frequency... but the variables are so great and distortions so apparent!!
And so, assuming you mean that you've also been able to use EQ to match phases between a horn and a reflex device, and group delay of course, and relative output across their response range, and forgetting increased thd at a given xmax in the ported box resulting in upper harmonics making the thing sound brighter, and forgetting changes in dispersion in the upper part of the units range, I still find it quite likely that these boxes would sound entirely different in their integration in a real system in a real environment.
path length delay?
group delay?
room coupling?
directionality at a given frequency range?
Ok I understand by using multiples in the situation you can get a better result. Doesn't surprise me... but we're talking about compromises here aren't we? It's not the folded horn design cab type at fault, but the room it's placed in.
Solution... spend 3x more on drivers and x amount more on amps (depending on impedances) to fix it.
If anything the lengths we have to go to in terms of extra drivers and amplification should indicate that the horn is a far superior device to a reflex one. It's just not always practical.
I'd also like to point out that room EQ will do nothing to fix modal ringing in a compromised environment. In those sorts of environment I'd get in some rpg modex panels and still use a horn.
Forgive me if I'm confused here, I pretty much refuse to believe any two different systems have the 'same response', with EQ or without. Maybe we get similar %thd, -3db points in available spectrum, unified phase response, peak and rms max spl for a given %thd, dispersive qualities at a given frequency, group delay across frequency... but the variables are so great and distortions so apparent!!
And so, assuming you mean that you've also been able to use EQ to match phases between a horn and a reflex device, and group delay of course, and relative output across their response range, and forgetting increased thd at a given xmax in the ported box resulting in upper harmonics making the thing sound brighter, and forgetting changes in dispersion in the upper part of the units range, I still find it quite likely that these boxes would sound entirely different in their integration in a real system in a real environment.
I don't have a position. The post you quoted for the response you made was simply telling a new participant in this thread what your position was. I credit your knowledge and I fail to resolve why they do sound different. Demonstrating that they sound different is easy. Play them and every single time, I can tell them apart. At some point, the consistent ability to tell them apart moves past statistical likelihood of guesses being correct. What I have now is two drivers that sound different in that I can pick it out without fail. If its some EQ magic process that will resolve the situation, I'm all ears.
When you speak of physics, what I don't buy and wont until its demonstrated is that a cone that is 3x as heavy can ignore Newton's laws of motion and stop and start as fast as the lighter one if the lighter cone has a better power to weight ratio motor driving it. Inertia is physics.
Last edited:
Interesting enough, this link here is to Data-bass' test of my exact TC LMSR 12" in 3 different type enclosures and after EQ there were measurable differences. EQ could not adjust for some things. What it didn't address is if you could tell which was playing if the listening was limited to say 92db.
Data-Bass
In these overlay-ed graphs you see the 3 alignments with the same driver (MINE) EQ'ed. The first is sealed vs passive rad and the second is sealed vs horn. Refer to the link to see the EQ process and its limitations.
Data-bass noted the ringing from the horn design in the 55hz range as an issue their EQ imitations couldn't fully compensate for. They do look good as far as matched curves. Note that the graphs are freq vs SPL vs time. While EQ did a fair job of matching freq vs amplitude, it didnt keep the 55hz resonance from continuing out 100ms longer after the sealed system stopped making that freq. The sealed system had ringing up near 90hz but in most sub apps, the cutoff would be lower.
Geddes paper addressed the room issue with staggered placement and as far as I'm concerned its 100% right that the room can be negated with staggered placement eliminating it as a variable.
Last edited:
I used the TC is 3 ported alignments and two sealed boxes and posted pics to prove it. All ported ones are complete mud. I think that driver's Q is too high to expect anything good from a ported alignment. What blows me away is how a sealed TC LMSR 12" sound really poor in comparison to a ported SEAS L26ROY after EQ. In my house thats a DCX2496 and DEQ2496.
Anyone thinks I can't pick them apart can bring $1000 bucks here, EQ the system and if I can't tell them apart, you leave with $2000.
In these overlay-ed graphs you see the 3 alignments with the same driver (MINE) EQ'ed. The first is sealed vs passive rad and the second is sealed vs horn. Refer to the link to see the EQ process and its limitations.
Data-bass noted the ringing from the horn design in the 55hz range as an issue their EQ imitations couldn't fully compensate for. They do look good as far as matched curves. Note that the graphs are freq vs SPL vs time. While EQ did a fair job of matching freq vs amplitude, it didnt keep the 55hz resonance from continuing out 100ms longer after the sealed system stopped making that freq. The sealed system had ringing up near 90hz but in most sub apps, the cutoff would be lower.
.
Such resonances are typically due to over shortened horns and small mouth areas. A horn of correct design would have a very low resonance across its usable range and limited group delay.
90hz... well it depends on the slope of your filter. And don't forget the acoustic crossover point will often be higher than the electronic crossover point, making this frequency range very important!
Last edited:
Nobody is ignoring physics. Some things you need to keep in mind:
1) Sound Pressure Level is proportional to cone acceleration (at least in a direct radiator) not velocity or displacement. F=ma and a=F/m, so a "better power to weight ratio" (actually a better force/mass ratio) simply means that for the same input acceleration is greater. Greater acceleration means more SPL output, NOT better high frequency or transient response as you imply. This is a common fallacy.
Within the piston region of the driver the only difference in frequency response will be the change in fundamental resonance at the low end where the stronger motor and/or lighter cone will lower the Q.
Now, in practice a driver with a thinner and lighter cone will typically have a more extended high frequency response above the piston range of the driver but this is NOT due to a simplistic force versus mass argument as you present it - you are now dealing with the highly complex cone breakup region which can't be characterised by something as simple as F=ma.
2) Any arguments about the high frequency transient response of a woofer ignore the dominating effect of the low pass filter which drives it. If two woofers have different high frequency performance, so long as both comfortably exceed the target response its largely irrelevant by how much as the low pass filter is going to shape the final response and therefore the "speed" of the woofer.
For example if you had one woofer that was flat up to 600Hz and another that was flat to 1200Hz but you wanted to low pass filter the response at 200Hz then you could achieve the same target acoustic response from both woofers using slightly different electrical filters.
Once low pass filtered to the target response neither woofer is "faster" than the other, because the final low pass function of woofer + filter is dictating the performance of the woofer.
Looking at the motor strength and cone mass of the driver alone really doesn't tell you much of interest, certainly not "speed".
Simon
Thanks for taking the time to correct the distortions of the facts. I agree with everything that you say, I just don't have the patience to write it all down.
There are a couple of second order effects that can sometimes "temper" my position. These are that the larger BL drivers will "tend" to have a lower thermal compression, which can be a factor, and they "tend" to have higher efficiency so the amps don't clip as much. (People don't realize how easy it is to clip a sub amp. It happens a lot more than most would think it does.)
The other factor is minimum phase (MP) aspects of the LF sound field. If the LF field is MP then correcting the frequency response will necessitate correcting the time response, i.e. the waterfall response. For the most part most systems will be MP at LFs. BUT, and SpinMonster's data shows this, if the response is not MP then this will not necessarily occur. Take his horn data, which has a log decay at around 60 Hz. If you look carefully you can see that this decay has two different slopes. This means that this system is not MP. If the frequency response is flattened then the early steep decay will go away, but the later slower decay will not. That is because this decay comes from a secondary coupled system which has a very high Q resonance, but is not so tightly coupled that this resonance is seen in the steady state frequency response, only in the decay (its what is called a multipath signal). Hence, room EQ will not completely resolve this situation (it should be fixed at the source of the problem, the horn design itself.)
These later aspects are not primary factors in that they "make all the difference", but they can become significant factors if not dealt with directly at the source of the problem.
Thanks for taking the time to correct the distortions of the facts. I agree with everything that you say, I just don't have the patience to write it all down.
There are a couple of second order effects that can sometimes "temper" my position. These are that the larger BL drivers will "tend" to have a lower thermal compression, which can be a factor, and they "tend" to have higher efficiency so the amps don't clip as much. (People don't realize how easy it is to clip a sub amp. It happens a lot more than most would think it does.)
The other factor is minimum phase (MP) aspects of the LF sound field. If the LF field is MP then correcting the frequency response will necessitate correcting the time response, i.e. the waterfall response. For the most part most systems will be MP at LFs. BUT, and SpinMonster's data shows this, if the response is not MP then this will not necessarily occur. Take his horn data, which has a log decay at around 60 Hz. If you look carefully you can see that this decay has two different slopes. This means that this system is not MP. If the frequency response is flattened then the early steep decay will go away, but the later slower decay will not. That is because this decay comes from a secondary coupled system which has a very high Q resonance, but is not so tightly coupled that this resonance is seen in the steady state frequency response, only in the decay (its what is called a multipath signal). Hence, room EQ will not completely resolve this situation (it should be fixed at the source of the problem, the horn design itself.)
These later aspects are not primary factors in that they "make all the difference", but they can become significant factors if not dealt with directly at the source of the problem.
Last edited:
How would it manifest itself if that condition isn't met?
Again, I am not presenting any argument or view point as you suggested more than once. I simply didn't resolve that issue yet. I was trying to understand why I hear a difference between two drivers after EQ which according to so many posts, shouldn't be the case. The fact that there is a difference is demonstrated by the consistent fact that I and others easily pick it out.
Someone questioning things isn't a distortion of a fact. I stated a number of times that I fail to resolve something so I started a thread to get input on it. I never came in here stating "Geddes is wrong and this is a fact that I'm not open to debate on."
What is apparent to me is that some alignments and drivers are easier to work with.
I wonder, is it the case that most systems are minimum phase at low frequencies though ? Low frequency room response certainly can be minimum phase, but unless you follow certain design practices, I wonder if it's more often the case that it isn't ?
I've only measured a couple of rooms with a mind to seeing if they measure minimum phase at low frequencies at the normal listening distance, (using an excess group delay measurement) both times it was a system with two main speakers and no separate subs, and in both cases the low frequencies were definitely not minimum phase at or anywhere near the normal listening position.
I don't recall the exact figures and I'm not sure if I still have the measurements files either, but from what I remember the bass was mostly minimum phase except for a couple of really large spikes in excess group delay - one exceeding 60ms at around 65Hz - not coincidentally quite close to a notch in the amplitude response caused by room modes/boundary cancellation.
The large spike in group delay presumably being caused by the sudden 180 degree phase reversal either side of the notch.
I'd need to measure more rooms but it seems likely that any time there is a deep cancellation due to a boundary effect that a large spike in excess group delay (and hence a non minimum phase response) will likely be present.
So not only can that notch not be EQ'ed because it's too deep and varies too much with listening position, the response at that frequency is highly non minimum phase as well.
It seems to me that a properly set-up distributed multi-sub system has the potential to make the response at the listening position actually minimum phase across the entire bass region by avoiding the deep boundary cancellations that can otherwise happen.
At the frequencies where a notch with rapid phase reversal would otherwise occur, one or more of the distributed subs will be at a suitable distance to "fill in" the hole.
Although the summed amplitude response still may not be completely flat it will be a lot better than before and perhaps more importantly it may now be completely minimum phase and thus amenable to complete correction by standard minimum phase EQ, at least at that listening point.
Could it be that one of the benefits of a distributed multi-sub system could be in achieving a minimum phase bass response ? Have you ever performed an excess group delay measurement on a multi-sub system with and without the subs enabled ?
Could the excess group delay measurement be used to help optimise the positioning and blending of subs rather than relying entirely on amplitude response ?
It's something that I will definitely try when I'm back in a position to measure and test speakers, unfortunately I don't even have a stereo system set up at the moment, let alone the facility to take measurements...
Interesting and insightful explanation - I've certainly seen CSD's with multiple slope decays before and theorised that it must be due to two independent decoupled radiating sources - with the high amplitude fast decaying resonance obscuring the low amplitude but higher Q slower decaying resonance.
It seems to be quite common in whizzer cone full range drivers - if you EQ the amplitude response, as you say the initial fast decaying resonance is eliminated but the slower decaying "buried" resonance remains. (However if its more than about 30dB down its probably inaudible)
In the case of the whizzer cone driver the decay characteristics of the two cones is very different as they don't have resonances at the same frequencies, hence one cone might decay quickly at a frequency the other cone has a high Q resonance and vica versa.
It's hard to know whether the buried resonance at 60Hz in SpinMonsters plots is from the speaker itself or is a room effect though...
How accurate is your EQ and measurement ? If it's just 1/3rd Octave resolution then the narrow band response of the drivers won't be the same at higher frequencies where the drivers are breaking up. Relatively broad EQ won't correct for highly audible high Q resonances.
Either that or you're noticing some other attribute about the speaker other than frequency response, such as distortion, or dynamic range compression. A bit hard to know unless you can describe in what way they sound different.
Fair enough. But I was also not saying that you, in particular, were distorting anything. I was just agreeing with Simon that "there is a common misconception" about many of the facts. I never said who had them and who didn't.
This is very clearly the question. I don't know either. I know that rooms can be non-MP, but maybe "good ones" never are. I don't have any data either way.
Consider this example. Let's say that I connect a very large "bottle" with a small neck to the room such that its resonance is about 60 Hz or so. Now since this is a small neck, it will only weakly couple to the room and its effect on the "overall" response will be small (a small notch would be expected). But since this is a very high Q system it will take a long time to build to its maximum and just as long to decay. So this will appear just like the example shown (and I agree we don't know if this is from the speaker or the room.) This situation is clearly not MP, but just as clearly "not a good idea".
Generally speaking, I think that any resonances in the playback are not good things. The resonances belong in the performance (music is nothing but a well controlled set of specific resonances of various instruments), not in the reproduction.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.