I always only build closed box systems. This is my FIRST-TIME experience with a bass-reflex design.
Usually, in the closed box design, I do prefer system with low Q; a critical damping (Q = 0.5), or a Butterworth alignment (Q = 0.707).
In ported design, I'd still prefer the lower Q. From my searching, I see the recommendation of a low-Q system on bass-reflex design is 4th-order Bessel alignment, or its abbreviation is BE4.
1. Is Q value of BE4 alignment be 0.577, same as in the closed box design, or 0.33 (1/3)?
2. Why BE4 is not popular compared to other alignments? Since I could find very little information about BE4 calculation.
3. Could anybody here provide the method of BE4 calculation, or links to it, please?
Usually, in the closed box design, I do prefer system with low Q; a critical damping (Q = 0.5), or a Butterworth alignment (Q = 0.707).
In ported design, I'd still prefer the lower Q. From my searching, I see the recommendation of a low-Q system on bass-reflex design is 4th-order Bessel alignment, or its abbreviation is BE4.
1. Is Q value of BE4 alignment be 0.577, same as in the closed box design, or 0.33 (1/3)?
2. Why BE4 is not popular compared to other alignments? Since I could find very little information about BE4 calculation.
3. Could anybody here provide the method of BE4 calculation, or links to it, please?
Last edited:
Hi,
I m more or less on the same situation as you ( except i've already done some experiments with reflex box and came to conclusion i liked bessel alignement better than the other i've tried. I may have biased conclusions thoughts as i might had made mistakes or used an inadapted driver for some alignements i've tried).
1: yes it should be. The difference with sealed is in the slope steepness: 2pole ( 12db/oct) sealed, 4pole ( 24db/octave) vented.
2: my experience with it is it doesn't sound as 'loud' in bass as some other alignements. If you are used to sealed 0,577 it should not bother you that much imho. But in the end it will depend of your room: if the room is bass 'leaky' it may seems to lack a bit ( versus other alignements). If your room is a concrete shell it may integrate way better and the 'lesser bass' will be compensated by room gain in low end.
I would say the difference are not this far as comparing sealed with Q of 0,577, 0,707 and 1: in some room one is better than the other subjectively speaking.
That said if we forget freq domain and we take a look at time domain there is no question bessel as an advantage.
Group delay plots will tell you...
Jbl Everest S9500 or 7500 ( series from 90's) used bessel alignements iirc. They targeted Japanese market where room are as an average 'smaller'. That said from reviews on eurpoean or american room nobody seemee to lack bass extension at the time ( i never heard one).
3: go there and follow the scenario, when asked for 'bass reflex box volume' follow bessel alignement and automated procedure.
Dôme acoustique : Calculs d'une enceinte bass-reflex, 1/8
I m more or less on the same situation as you ( except i've already done some experiments with reflex box and came to conclusion i liked bessel alignement better than the other i've tried. I may have biased conclusions thoughts as i might had made mistakes or used an inadapted driver for some alignements i've tried).
1: yes it should be. The difference with sealed is in the slope steepness: 2pole ( 12db/oct) sealed, 4pole ( 24db/octave) vented.
2: my experience with it is it doesn't sound as 'loud' in bass as some other alignements. If you are used to sealed 0,577 it should not bother you that much imho. But in the end it will depend of your room: if the room is bass 'leaky' it may seems to lack a bit ( versus other alignements). If your room is a concrete shell it may integrate way better and the 'lesser bass' will be compensated by room gain in low end.
I would say the difference are not this far as comparing sealed with Q of 0,577, 0,707 and 1: in some room one is better than the other subjectively speaking.
That said if we forget freq domain and we take a look at time domain there is no question bessel as an advantage.
Group delay plots will tell you...
Jbl Everest S9500 or 7500 ( series from 90's) used bessel alignements iirc. They targeted Japanese market where room are as an average 'smaller'. That said from reviews on eurpoean or american room nobody seemee to lack bass extension at the time ( i never heard one).
3: go there and follow the scenario, when asked for 'bass reflex box volume' follow bessel alignement and automated procedure.
Dôme acoustique : Calculs d'une enceinte bass-reflex, 1/8
Last edited:
The denominator polynomial of a 2nd-order Bessel filter is
s^2 + 3.00000000000000s + 3.00000000000000 (Q = 0.57735)
The denominator polynomial of a 2nd-order Butterworth filter is
s^2 + 1.414214s + 1.000000 (Q = 0.70711)
The denominator polynomials of a 4th-order Bessel filter are
s^2 + 4.20757879435926s + 11.4878004768712 (Q = 0.80554)
s^2 + 5.79242120564074s + 9.14013089027793 (Q = 0.52193)
The denominator polynomials of a 4th-order Butterworth filter are
s^2 + 0.76536s + 1.00000 (Q = 1.30657)
s^2 + 1.84776s + 1.00000 (Q = 0.54120)
I've attached graphics showing the frequency and impulse responses of all four, with a -3dB cutoff frequency of 30 Hz.
s^2 + 3.00000000000000s + 3.00000000000000 (Q = 0.57735)
The denominator polynomial of a 2nd-order Butterworth filter is
s^2 + 1.414214s + 1.000000 (Q = 0.70711)
The denominator polynomials of a 4th-order Bessel filter are
s^2 + 4.20757879435926s + 11.4878004768712 (Q = 0.80554)
s^2 + 5.79242120564074s + 9.14013089027793 (Q = 0.52193)
The denominator polynomials of a 4th-order Butterworth filter are
s^2 + 0.76536s + 1.00000 (Q = 1.30657)
s^2 + 1.84776s + 1.00000 (Q = 0.54120)
I've attached graphics showing the frequency and impulse responses of all four, with a -3dB cutoff frequency of 30 Hz.
Attachments
Hi gberchin,
Where did you get this info? Could you help to provide me, please? Is it from research papers? I’m really interested in it.
I'm a signal processing engineer. The polynomials can be found in signal processing textbooks. The program that displays the responses I wrote myself, for my own use.
There's a good reason nobody uses it. I'd tell you, but I can't remember...
I was reading about the alignments a few weeks ago and the case against the Bessel 4 box was so compelling and complete I just thought to myself "this version of venting should be disregarded". And I did. Don't have room for garbage!
I'll look for what I was reading
I was reading about the alignments a few weeks ago and the case against the Bessel 4 box was so compelling and complete I just thought to myself "this version of venting should be disregarded". And I did. Don't have room for garbage!
I'll look for what I was reading
Hi,
Gberchin i am not familiar with polynomial expression of filters ( i'm a user not a designer of this tools).
I understood a 4pole LR is equivalent to 2 butterworth in serie and so i don't get how to use the Q values you gave us for the q of a 4pole LR ?
And while at it is it possible to stack bessel 2 poles to have a 4p bessel?
Mike7877, yes please give us theorical reason to discard a solution without even try it, i'm curious about them.
Transient integrity is of prime importance in audio reproduction: the firsts ms of a sound contains most informations about family of instrument, material it use, etc,etc,... it may appear to be unimportant for bass but if you think about it an acoustic kick drum is mainly... a transient signal ( as are most acoustic bass signals anyway - orgue being the exception).
Gberchin i am not familiar with polynomial expression of filters ( i'm a user not a designer of this tools).
I understood a 4pole LR is equivalent to 2 butterworth in serie and so i don't get how to use the Q values you gave us for the q of a 4pole LR ?
And while at it is it possible to stack bessel 2 poles to have a 4p bessel?
Mike7877, yes please give us theorical reason to discard a solution without even try it, i'm curious about them.
Transient integrity is of prime importance in audio reproduction: the firsts ms of a sound contains most informations about family of instrument, material it use, etc,etc,... it may appear to be unimportant for bass but if you think about it an acoustic kick drum is mainly... a transient signal ( as are most acoustic bass signals anyway - orgue being the exception).
Last edited:
A 4th-order Linkwitz-Riley filter and a 4th-order Butterworth filter are not the same thing. A 4th-order Linkwitz-Riley filter is the cascade of two 2nd-order Butterworth filters, but a 4th-order Butterworth filter is the cascade of the two 2nd-order filters that I showed in my original message.
You cannot just cascade 2nd-order Bessel filters and get a higher-order Bessel filter. Same applies to Butterworth filters.
You cannot just cascade 2nd-order Bessel filters and get a higher-order Bessel filter. Same applies to Butterworth filters.
Please do as it can only be bested by an EBS alignment AFAIK, which is several times larger and even then, must be tuned very low to keep its ultimately much greater delay down low enough that it has decayed enough by the time our hearing starts becoming more acute.
Keele's EBS [Thiele #9.5]: Vb = 2.6*Vas, Fb = 0.52*Fs, 0.625 Qts'
Regarding transient response, take a look at this post and the six that follow.
HT 2 12" subs vs 1 15" sub - Home Theater - The Klipsch Audio Community
Note that a significant degradation of transient response has resulted in less than half an octave of bass extension. Is it worth it?
HT 2 12" subs vs 1 15" sub - Home Theater - The Klipsch Audio Community
Note that a significant degradation of transient response has resulted in less than half an octave of bass extension. Is it worth it?