Conventional wisdom says sealed boxes have better transient response than ported. However ported gains efficiency and reduces excursion. Some personal experience in home and car has led me to feel that undertuned ported systems with a long slow "droopy" rolloff can work really well. What I always wondered was just how much inferior the transient response really was.
Once upon a time I wanted to simulate this with LEAP and LMS...unfortunately I never had time to do it. 🙁. It's SO annoying when actual work gets in the way of doing interesting stuff! 😡
Does anyone know a system that will let you simulate tone bursts? (Because I don't feel group delay curves or impulse response curves are very relatable; at least tone burst would give some kind of visual indication of differences, from which we could then argue about the significance 😛 )
Once upon a time I wanted to simulate this with LEAP and LMS...unfortunately I never had time to do it. 🙁. It's SO annoying when actual work gets in the way of doing interesting stuff! 😡
Does anyone know a system that will let you simulate tone bursts? (Because I don't feel group delay curves or impulse response curves are very relatable; at least tone burst would give some kind of visual indication of differences, from which we could then argue about the significance 😛 )
Why tone bursts? A tone burst is not a sinusoidal signal and correctly interpreting results from a band filtered system like a woofer in an enclosure is hard.
To me, this would require standard modern impulse measurement in apps like Arta, REW and Holmimpulse. Plus adequate measurement setup.
I do not think you are going to find new results though. The impulse response of a relative simple system can also be calculated. But in home (and car) audio the in room response is largely defined by the badly damped listening room, not the almost ideally damped loudspeaker.
To me, this would require standard modern impulse measurement in apps like Arta, REW and Holmimpulse. Plus adequate measurement setup.
I do not think you are going to find new results though. The impulse response of a relative simple system can also be calculated. But in home (and car) audio the in room response is largely defined by the badly damped listening room, not the almost ideally damped loudspeaker.
Last edited:
Sealed and ported subwoofers are both minimum-phase systems. If you EQ one to match the other's frequency response, the phase response will also be identical.
As a result, I don't think it can be said that one has better time-domain performance than the other.
Chris
As a result, I don't think it can be said that one has better time-domain performance than the other.
Chris
The tone bursts are very visual. So is impulse response, but I don't think in a meaningful way really. I forget where I saw tone bursts, maybe Dick Small's thesis? And some other few places perhaps. You can see how quickly the woofer "gets up to speed" and how much it rings. Now, how meaningful are small differences? Good question. But I wanted to see if there were gross differences at various frequencies or not, to see if an undertuned port would approach the transient performance of the sealed box. On top of all this though it's a good point about the room dominating.
Hmm, well yes. Though you can't really EQ the ported box to match the sealed unless you accept infinite gain down to DC ha ha. And of course the micro-granular response is probably not exactly minimum phase and varies with signal and all that fun stuff. My main wondering was regarding the old arguments and conventional wisdom about sealed versus ported, and thinking an undertuned port might give you the best of both worlds.
If you have or can borrow an Omnimic setup, you can get a graph of shaped toneburst envelopes (energy storage) over any range of bands in a waterfall format, calculated from impulse response.
edit: that's for measurement, not a simulation. Might not be what you wanted.
edit: that's for measurement, not a simulation. Might not be what you wanted.
Attachments
Last edited:
Can you draw examples of what you would expect from the response of a sealed and a vented box? Otherwise, Chris is right and the impulse response serves you perfectly. As I wrote before, the band pass function will give results that are difficult to read.
As for ‘micro-granular’ behavior, I’d say nonlinear behavior will mainly occur when the cone breaks up. Personally this kind of non-science doesn’t do it for me.
Well...I haven't used REW yet, can't you input a frequency response file somewhere from which it could calculate that?
IIRC, LMS would back-calculate the impulse response from the frequency response* and then you could simulate tone bursts somehow.
****-uming linear phase in order to validly apply a Hilbert (??) transform. This was one thing which I never liked, I prefer to really measure the phase.
The concept of linear phase has little to do with that. Loudspeakers in boxes behave as pretty linear devices as in their phase behavior resembles that in calculations on networks.
And again: how would you interpret the waveforms on the screen of your scope? No way you’ll see a clean burst, even your amp output won”t probably produce one.
I won’t take any holy grails away but you are on a quest for things that you will not find. And I mentioned it before: room interaction is of far more importance than the behavior of the speaker alone. Weakest link stuff and all.
And again: how would you interpret the waveforms on the screen of your scope? No way you’ll see a clean burst, even your amp output won”t probably produce one.
I won’t take any holy grails away but you are on a quest for things that you will not find. And I mentioned it before: room interaction is of far more importance than the behavior of the speaker alone. Weakest link stuff and all.
Could not have shown it any better. Hope that article makes the point I tried to.
BTW IIRC KEF (and possibly others) started Fourier transform measurements with real Dirac pulses in anechoic rooms. So no convolution, just the real response to an impulse. Turned out to give the same results as swept sines, periodic noise and the like, but only at far greater cost (extreme measuring conditions).
BTW IIRC KEF (and possibly others) started Fourier transform measurements with real Dirac pulses in anechoic rooms. So no convolution, just the real response to an impulse. Turned out to give the same results as swept sines, periodic noise and the like, but only at far greater cost (extreme measuring conditions).
Last edited:
Very interesting...there is no perfect lunch! I always wanted to experiment with listening tests, same woofer in a box with insertable panels or blocks to reduce the internal volume. Never had time, SO annoying when actual projects interfere with playing around at work 😉
And interesting find from Kreskovsky's article, is from the very end:
"......The further the fundamental of the input signal is above the high pass system resonance, the lower the level of phase distortion, but it can never be completely eliminated. From this point of view, it would be desirable to build a woofer system with Fs well below the lowest frequency the system is asked to reproduce. However, this is contradictory to the consideration of room gain effects unless the room is suitable large. Overall what we see here is how difficult it really is to produce a woofer system in any environment that will accurately reproduce an input with low frequency content. It's no wonder the bass is never right. "
So if you want accurate 80 to 100 hz fundamental waveform bass, implement woofer systems with system resonances in the 40 to 50 Hz range. This is illustrated by comparing Fig. 18 to the immediately preceding figures. Now the question arises as to how sensitive our hearing is to waveforms?
"......The further the fundamental of the input signal is above the high pass system resonance, the lower the level of phase distortion, but it can never be completely eliminated. From this point of view, it would be desirable to build a woofer system with Fs well below the lowest frequency the system is asked to reproduce. However, this is contradictory to the consideration of room gain effects unless the room is suitable large. Overall what we see here is how difficult it really is to produce a woofer system in any environment that will accurately reproduce an input with low frequency content. It's no wonder the bass is never right. "
So if you want accurate 80 to 100 hz fundamental waveform bass, implement woofer systems with system resonances in the 40 to 50 Hz range. This is illustrated by comparing Fig. 18 to the immediately preceding figures. Now the question arises as to how sensitive our hearing is to waveforms?
Last edited:
...which is why I prefer towers or at least blocking off ports even with a subwoofer. Hmm, does blocking off the ports have a similar help? Good question. And I'm not sure anyone really knows still how sensitive we are too bass distortions (including amplitude and time)
I think this subject has been studied long ago already. (Pre internet...) General outcome is that our ears are not that sensitive to the more gradual phase shifts. Put it the other way around. When you think of the inner ear and how sound is collected, an assumption that phase shifts matter does not come directly to mind.
Bill Waslo - I am a fan of Omni Mic. I wonder if you could elaborate on how to interpret the energy storage plot, namely this: I get quite different result depending on if I select 1, 2, 3, 4, or 5 cycle tone burst. The software default is 5 cycle. The fact that I get a difference between 2, 3 or 5 cycles, means... what ecactly?
I would think that is true. However, I would also think that delayed energy from the woofer feeding into the resonant room system would not be a good thing.
Somewhere I saw some tone bursts versus Q, some article. It was interesting showing that a low Q had "too slow" of a transient response, the rise time lagged. A higher Q was actually closer to ideal. Of course the wild card is this likely depends on excitation frequency versus speaker resonance frequency.
And in any case I still want to be able to play around and see what I find.