Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Irvin Pribadi]

As you go lower in freq-there is more phase shift/time in the filter.
As the filter gets steeper-there is more phase shift/time in the filter.
Put the the two together and it adds up pretty quick
I typically don't use more than a 24dB/oct slope.

Hey Ivan,
I've been passively following this thread and got a question: Does this affect both digital and analog filters alike? Thanks.

[Art Welter]

That subharmonic has to punch you fast, it has to be a "bum" not a "wooofff..." the way i see it is, once that subharmonic hits the amp, the driver needs to accelerate from standing still to that 40hz hit as fast as possible and then stop as fast as possible once it stops getting signal. These emminence drivers in the combination of the boxes they are in are simply to dull for that.
It will be to slow in accelerating to reaching that 40hz vibration, which will mean it will basicaly be late to follow the upper ~150hz kick, and then when the transient ends the cone will keep flopping around for too long insted of stopping fast. No matter what i do this is the way these boxes behave.

1)The more load you give them the less precise they become. If it's only a fast 40-50hz hit it might be somewhat decent, if it comes on top of a lasting subline, it will drown even more. 


2)TIM: Can you explain the delay factor with high pass filters? I didn't understand what you are trying to say.

1) When you push a driver (any driver) beyond Xmax, the magnet has little control over the cone, resulting in harmonic distortion, the harmonics of 50 Hz, that is 100, 150, 200, etc. may actually be louder than the fundamental.
Doubling Xmax (and power) results in 6 dB more output, the 18" drivers you are using may have only 6mm Xmax, while there are 18" with as much as 24mm Xmax, +12 dB more output potential. One of those drivers could equal the LF output of four of the lesser Xmax drivers.
2) Each "pole"(or "order") of a crossover introduces 90 degree phase shift. 6 dB per octave introduces 90, 12 dB 180 degree, 24 dB 360, etc.
360 degrees of phase rotation is one cycle. One cycle at 20 Hz lasts 1000ms/20=50ms. A 48 dB crossover introduces 720 degrees of phase shift, 100ms delay at 20 Hz. At 100 Hz, a 48 dB crossover would introduce 20ms delay.

The low end of a bass reflex (phase inversion) speaker at Fb lags the top end by 360 degrees before any HP has been introduced, the cumulative delay can be much longer when a high order HP is added.

Below is the frequency and phase response of a BR cabinet with a 38 Fb (port tuning) measured "raw", no HP or LP filters.

Art

[Eugen Jeličić]

Art, i understand well what is happening, you explained it nicely. But i don't understand why is it happening...

Basically you are saying that if i take sub, then put a highpass filter 24dB/octave at 40hz then all the frequencies in the area affected by the filter will be getting out of the sub 360 degrees late.
I'm really confused here and have no clue why a high pass filter would cause this.

[Eugen Jeličić]

Wait, i've been thinking a bit while looking at the graphs you gave me. I don't really understand phase response in depth.
But something came across my mind. A driver at 50hz, will be moving faster then it will at 40hz. This means that physics won't allow 2 identical drivers in 2 identical boxes be in phase if they are not running at the same frequency. It's perfectly logical of course. You just imagine the sinewaves of both of them. That of course explains how the phase response changes in a linear way as the frequency rises, just like on the graph above. Now i'm thinking about the high pass filter once again... it will start decreasing the drivers movement under a certain frequency... but the phase should stay the same... it should just decrease it's amplitude, i still don't get it

[Ivan Beaver]

Art, i understand well what is happening, you explained it nicely. But i don't understand why is it happening...

Basically you are saying that if i take sub, then put a highpass filter 24dB/octave at 40hz then all the frequencies in the area affected by the filter will be getting out of the sub 360 degrees late.
I'm really confused here and have no clue why a high pass filter would cause this.

No.

Phase is freq specific.

And when you say "all of the freq in the area of the filter", than can mean a lot of different things-depending on what are you calling "in the area".

The best way is to use a DSP and look at the phase response as you put different filters in-both freq and slope.

[Eugen Jeličić]

No.

Phase is freq specific.

And when you say "all of the freq in the area of the filter", than can mean a lot of different things-depending on what are you calling "in the area".

The best way is to use a DSP and look at the phase response as you put different filters in-both freq and slope.

Yes, look at my upper post, i just sat to think about it and it of course logical. By "the area" i meant the frequencies that the high pass filter affects.
DSP and looking at phase response doesn't make much sense right now, i can see what is happening but i still won't understand why is it happening which is the point right now

[Ivan Beaver]

Wait, i've been thinking a bit while looking at the graphs you gave me. I don't really understand phase response in depth.
But something came across my mind. A driver at 50hz, will be moving faster then it will at 40hz. This means that physics won't allow 2 identical drivers in 2 identical boxes be in phase if they are not running at the same frequency. It's perfectly logical of course. You just imagine the sinewaves of both of them. That of course explains how the phase response changes in a linear way as the frequency rises, just like on the graph above. Now i'm thinking about the high pass filter once again... it will start decreasing the drivers movement under a certain frequency... but the phase should stay the same... it should just decrease it's amplitude, i still don't get it

Think of it in the passive way.

Ac components (coils and capacitors) have a phase "aspect" to them-all be it opposite of each other.

Since the signal is AC-there is an impedance-which contains both resistive and reactive components.

These reactive components are what cause the phase shift.

The more poles (steeper) the filter, the greater the phase shift.  They just add up.

Different freq will be "in phase" and "out of phase" at different points in their curves.  It does not matter if they are in the same speaker cabinet or not.

This is why when you mix different freq you can get some addition and some subtraction-depend on the phase relationships.

[Ivan Beaver]

Yes, look at my upper post, i just sat to think about it and it of course logical. By "the area" i meant the frequencies that the high pass filter affects.
DSP and looking at phase response doesn't make much sense right now, i can see what is happening but i still won't understand why is it happening which is the point right now

The filter "affects" freq above it and below it.

How much depends on the range of interest.

[Stephen Kirby]

It still seems that Eugen is confused about the "speed" of the drivers motion.  As was pointed out, the speed of the cone's motion is related to the frequency and displacement it's going through.  If you look at FFT theory faster "speeds" (i.e. electrical rise times of non-sinusoidal waveforms) are made up of a combination of frequencies.  So the "faster" frequencies are reproduced by the other drivers in a system.  Which brings us back to the alignment of the various bandpasses so that the "impact" of the waveform coming from the upper bandpass elements of the system "speaks" together with the lower fundamental.  And that the first harmonic which is reproduced by both the sub and low/mid speaker doesn't get muddled between the two.
Any reactive system is going to have frequency dependent phase shift.  Whether it's an electronic crossover or a mechanical reproducer.  So now it becomes apparent how difficult it is to get the various components or harmonics of a sound to line up perfectly.

[Ivan Beaver]

It still seems that Eugen is confused about the "speed" of the drivers motion.  As was pointed out, the speed of the cone's motion is related to the frequency and displacement it's going through.  If you look at FFT theory faster "speeds" (i.e. electrical rise times of non-sinusoidal waveforms) are made up of a combination of frequencies.  So the "faster" frequencies are reproduced by the other drivers in a system.  Which brings us back to the alignment of the various bandpasses so that the "impact" of the waveform coming from the upper bandpass elements of the system "speaks" together with the lower fundamental.  And that the first harmonic which is reproduced by both the sub and low/mid speaker doesn't get muddled between the two.
Any reactive system is going to have frequency dependent phase shift.  Whether it's an electronic crossover or a mechanical reproducer.  So now it becomes apparent how difficult it is to get the various components or harmonics of a sound to line up perfectly.

Even a single driver (be it a woofer-full range speaker or HF driver) is a band pass device (meaning highs and lows roll off) and the phase response will look like an electronic active or passive filter that is assembled to be a bandpass filter.

The basic shape is such that the phase response will slope down from the low freq-then flatten out, and then slope down again at the high freq.

The slopes will more or less slope off as the freq response drops.

So the electronics follow the natural response-if everything is "perfect".

But it is not-so there starts to be more involved.

[ProSoundWeb Community]