DiyMobileAudio.com Car Stereo Forum banner
1 - 20 of 29 Posts

·
Registered
Joined
·
65 Posts
Discussion Starter · #1 ·
Why is it that setting the LPF and HPF on your head unit doesn't create as nice a "blend" as setting the LPF to 80hz and the HPF to 100hz?

I listened to my stereo where the LPF and HPF were both set at 80hz, and it didn't sound as good as the 80hz LPF / 100hz HPF setup.

-John
 

·
Registered
Joined
·
309 Posts
In an ideal world, speakers would have a perfectly flat response, with the same sensitivity across their frequency range, woofers would have identical sensitivity to tweeters, etc. A crossover point is the frequency where the output rolls off by 50%. In ideal world, both high pass and low pass frequencies would be at their 50% point at the same place and you'd have perfectly flat output. We do not live in an ideal world, however. Speakers will have different sensitivities across their frequency ranges, different amounts of power going to them, etc. Therefore there will usually be a sensitivity difference at the crossover point. Overlapping, underlapping, and different crossover slopes help us compensate for these sensitivity differences.

Or in simpler terms, your woofers are probably slightly more efficient near the crossover point than the rest of their frequecy range.
 

·
Registered
Joined
·
891 Posts
In an ideal world, speakers would have a perfectly flat response, with the same sensitivity across their frequency range, woofers would have identical sensitivity to tweeters, etc. A crossover point is the frequency where the output rolls off by 50%. In ideal world, both high pass and low pass frequencies would be at their 50% point at the same place and you'd have perfectly flat output. We do not live in an ideal world, however. Speakers will have different sensitivities across their frequency ranges, different amounts of power going to them, etc. Therefore there will usually be a sensitivity difference at the crossover point. Overlapping, underlapping, and different crossover slopes help us compensate for these sensitivity differences.

Or in simpler terms, your woofers are probably slightly more efficient near the crossover point than the rest of their frequecy range.
I'm used to seeing x-over points like -3 db or -6 db, depending on whether Butterworth, Linkwitz-Riley, etc. are used, but I've not seen a x-over point defined as a 50% reduction in output before. Am I missing something here...? :confused:
 

·
Registered
Joined
·
6,848 Posts
In an ideal world, speakers would have a perfectly flat response, with the same sensitivity across their frequency range, woofers would have identical sensitivity to tweeters, etc. A crossover point is the frequency where the output rolls off by 50%. In ideal world, both high pass and low pass frequencies would be at their 50% point at the same place and you'd have perfectly flat output. We do not live in an ideal world, however. Speakers will have different sensitivities across their frequency ranges, different amounts of power going to them, etc. Therefore there will usually be a sensitivity difference at the crossover point. Overlapping, underlapping, and different crossover slopes help us compensate for these sensitivity differences.

Or in simpler terms, your woofers are probably slightly more efficient near the crossover point than the rest of their frequecy range.
That depends more on the alignment of the xover. Butterworth alignment will yield "50%" (so to speak) for each, summing to 100%, but Linkwitz-Riley for example requires underlapping.
 

·
Registered
Joined
·
309 Posts
I'm used to seeing x-over points like -3 db or -6 db, depending on whether Butterworth, Linkwitz-Riley, etc. are used, but I've not seen a x-over point defined as a 50% reduction in output before. Am I missing something here...? :confused:
You're thinking of slopes, or db per octave. The crossover point is the frequency where it hits 50% reduction (-3db). For the slope, a 6db/octave filter at 2khz will be -3db at 2khz, -9db at 1khz, -15db at 500hz, and so on. a [email protected] filter will be [email protected], [email protected], [email protected], and so on.
 

·
Registered
Joined
·
309 Posts
That depends more on the alignment of the xover. Butterworth alignment will yield "50%" (so to speak) for each, summing to 100%, but Linkwitz-Riley for example requires underlapping.
Cool, I didn't know that. Ya learn something new every day :)
 

·
Registered
Joined
·
891 Posts
You're thinking of slopes, or db per octave. The crossover point is the frequency where it hits 50% reduction (-3db). For the slope, a 6db/octave filter at 2khz will be -3db at 2khz, -9db at 1khz, -15db at 500hz, and so on. a [email protected] filter will be [email protected], [email protected], [email protected], and so on.
No, I'm not talking about slopes. I was questioning you about your reference to a 50% output reduction at the x-over point, which as you say, is at -3db. This is not always true, as the x-over point for a fourth order Linkwitz-Riley alignment is at -6 db. So, if I were going by your math, and the -3db point yields a 50% cut in output, what, then, does the -6 db point realize in the way of reduced output?

http://rane.com/note160.html
 

·
Registered
Joined
·
309 Posts
My mistake. -6db would be about 75% reduction in power.
 

·
Registered
Joined
·
6,848 Posts
No, I'm not talking about slopes. I was questioning you about your reference to a 50% output reduction at the x-over point, which as you say, is at -3db. This is not always true, as the x-over point for a fourth order Linkwitz-Riley alignment is at -6 db. So, if I were going by your math, and the -3db point yields a 50% cut in output, what, then, does the -6 db point realize in the way of reduced output?

Linkwitz-Riley Crossovers: A Primer
I think you're both right on that point. -6dB voltage = -3dB power response. I think "50%" holds.
 

·
Registered
Joined
·
309 Posts
...and now I'm really confused
 

·
Registered
Joined
·
4,712 Posts
Why is it that setting the LPF and HPF on your head unit doesn't create as nice a "blend" as setting the LPF to 80hz and the HPF to 100hz?

I listened to my stereo where the LPF and HPF were both set at 80hz, and it didn't sound as good as the 80hz LPF / 100hz HPF setup.

-John
I'm going to offer two different aspects, I find these to be very prevalent in car audio:
1. your subwoofer level is much higher than the midbass/midrange. The rolloff of the subwoofer even if steep at a high level it will acount for a good part of the midbass region as well. This might give you a peak around 100hz.

2. Your midbass/midrange distort when crossed as low as 80hz. A higher filter eases the load on these drivers. Even if underlapping in this manner you might get a good transition to the mid if the sub level is set high like explained above.
 

·
Registered
Joined
·
65 Posts
Discussion Starter · #16 ·
I'm going to offer two different aspects, I find these to be very prevalent in car audio:
1. your subwoofer level is much higher than the midbass/midrange. The rolloff of the subwoofer even if steep at a high level it will acount for a good part of the midbass region as well. This might give you a peak around 100hz.

2. Your midbass/midrange distort when crossed as low as 80hz. A higher filter eases the load on these drivers. Even if underlapping in this manner you might get a good transition to the mid if the sub level is set high like explained above.
Interesting stuff. I don't get much (if any) noticeable distortion at the volumes I listen at, but it makes sense to me.

John
 

·
Registered
Joined
·
731 Posts
Also keep in mind that different enclosure alignments will produce different roll off slopes at the lower end of the response. For example, if your midbass start rolling off at 12db/octave at 80hz (for sealed enclosures) AND you add a 12sb slope, the total acoustic crossover slope is 24 db/octave. In addition, the total idealized phase shift is 360 degrees, acoustically. When compared to the subwoofer, the lowpass only rolls off at 12db per octave (assuming the sub has a flat response through the crossover region) and has a phase shift of 180 degrees. So in this case, the crossover point may have to be adjusted AND the phase of the woofer should be adjusted.

Again, the above is completely idealized. Actual in car response will be drastically different. This is why its useful to have a lot of crossover flexibility and slope adjustment. I usually set my highpass for the front stage FIRST. I then adjust the crossover on the subwoofer amp since its continuously variable. I then go back and adjust the slope of the fronts and/or the phase of the subwoofer. This requires a lot of iteration to get right. Don't forget about the levels either!
 

·
Registered
Joined
·
1,357 Posts
-3dB is NOT "50% of output".

You seem to be confusing the power required to increase a speaker's output by 3dB and the effect of a 3dB change on the sound amplitude.

3dB is the SMALLEST change in volume which we can reliably detect as a change in volume. We hear changes of less than 3dB as differences in tonality, not intensity.

So -3dB is used for crossovers because it's the point at which it makes a noticeable impact - just like freq response is given with +/-3dB deviations.

But a sound which is 3dB lower in intensity is not "50% lower".
 

·
Registered
Joined
·
1,357 Posts
Why is it that setting the LPF and HPF on your head unit doesn't create as nice a "blend" as setting the LPF to 80hz and the HPF to 100hz?

I listened to my stereo where the LPF and HPF were both set at 80hz, and it didn't sound as good as the 80hz LPF / 100hz HPF setup.

-John
Due to the common midbass cabin gain, Rockford's XV-1 active crossover came with a 100 Hz LP and a 150 (or was it 275?) Hz HP. Since the filters were at 12dB/octave, there was significant output in the stopband of both filters, and along with the cabin gain, RF figured that would be fine.

But most of us changed them to be the same xover anyway :)
 

·
Registered
Joined
·
4,712 Posts
doubling the amplifier power will give you 3db
human hearing:
6db increase is perceived as 50% increase in volume
10db is perceived as 100% or doubling of volume
1db is the smallest change in volume that we can detect
 
1 - 20 of 29 Posts
Top