ProSoundWeb Community

Sound Reinforcement - Forums for Live Sound Professionals - Your Displayed Name Must Be Your Real Full Name To Post In The Live Sound Forums => LAB Lounge => Topic started by: luis Markson on April 09, 2011, 12:13:15 am

Title: Crossover question
Post by: luis Markson on April 09, 2011, 12:13:15 am
On the Mackie S410s the internal crossover in set at 180Hz 12dB/Octave. The recommendation for an external crossover is 150 Hz 24dB/ocatve.

Why is the internal unit different from the recommended settings?
Title: Re: Crossover question
Post by: Chris Carpenter on April 09, 2011, 02:36:10 am
On the Mackie S410s the internal crossover in set at 180Hz 12dB/Octave. The recommendation for an external crossover is 150 Hz 24dB/ocatve.

Why is the internal unit difference from the recommended settings?

So that you have the option to cross it over higher. I imagine they set the crossover to the highest usable range of the driver so you can cross it over wherever you like and not have the built in crossover get in the way. Unless you can bypass the internal crossover; then I'm just flat wrong.
Title: Re: Crossover question
Post by: Mike McNany on April 09, 2011, 06:12:40 pm
On the Mackie S410s the internal crossover in set at 180Hz 12dB/Octave. The recommendation for an external crossover is 150 Hz 24dB/ocatve.

Why is the internal unit different from the recommended settings?

Because an active crossover can be had with the adjustable filtering AND the proper components for a 24dB per octave filter. Using a lesser filter, the 12dB per octave, means using fewer AND cheaper components for the filter. I.e., cheaper to build, more profit. Behringer (I know, poor example) uses a 6dB/octave low pass and a 9dB/octave high pass PASSIVE filter at 170Hz for it's B1800X sub. Semi-usable for seperating subs from tops at best, with too shallow of filters and far too high a crossover frequency. But definately cheaper   :-X

Mike McNany
Title: Re: Crossover question
Post by: Ivan Beaver on April 09, 2011, 06:52:53 pm
Because an active crossover can be had with the adjustable filtering AND the proper components for a 24dB per octave filter. Using a lesser filter, the 12dB per octave, means using fewer AND cheaper components for the filter. I.e., cheaper to build, more profit. Behringer (I know, poor example) uses a 6dB/octave low pass and a 9dB/octave high pass PASSIVE filter at 170Hz for it's B1800X sub. Semi-usable for seperating subs from tops at best, with too shallow of filters and far too high a crossover frequency. But definately cheaper   :-X

Mike McNany

Where did you get that 9dB/0ctave spec from?  That would be a pretty complicated filter to construct-as the usual slopes are multiples of 6dB.

Title: Re: Crossover question
Post by: luis Markson on April 10, 2011, 07:19:19 am
Here's another question:

This is an extract from the x-over section of www.lenardaudio.com :

The crossover decreases power at a given rate from either or both sides of a given bandwidth. The slopes rotate phase in opposite directions.

Could someone please explain what is meant by, "The slopes rotate phase in opposite directions".

Is this phase "rotation" an intentional quality or is it an inherent problem with filtering. Is this related to what I read about other EQ "phase problems".

By "rotate" are we talking about a gradual shift along the slope or an inversion at the -3dB point?
Title: Re: Crossover question
Post by: luis Markson on April 11, 2011, 06:35:31 pm
Here's another question:

This is an extract from the x-over section of www.lenardaudio.com :

The crossover decreases power at a given rate from either or both sides of a given bandwidth. The slopes rotate phase in opposite directions.

Could someone please explain what is meant by, "The slopes rotate phase in opposite directions".

Is this phase "rotation" an intentional quality or is it an inherent problem with filtering. Is this related to what I read about other EQ "phase problems".

By "rotate" are we talking about a gradual shift along the slope or an inversion at the -3dB point?

I'm not sure how this forum feels about thread "bumbs", but here it is.

I'm working my way up to the math heavy x-over information (its going to take a while...) and would like to know whats going on within the x-over region with regards to phase.
Title: Re: Crossover question
Post by: Art Hays on April 11, 2011, 09:37:30 pm
As far as what phase is-

A continuous-time signal can be decomposed into a collection of sine-waves, each with a particular freq, amplitude and phase.  If you were to connect the mains output of a mixer to a scope or suitable laptop program and look while a band was playing you would see a rapidly changing display that wouldn't tell you much.  This is viewing in the time-domain: x-axis is time.  If you switched the scope to spectrum analysis you would see lots of vertical lines.  This is the frequency-domain, x-axis is freq.

So if you have a low-pass filter it will attenuate freq's above a cut-off.  BUT, it also can change the phase of freq's (phase can also be thought of as a delay of those freq's).  Some filters don't change phase much in the pass-band (Bessel), others do.  There are trade-offs with the different types of filters.  Bessel is great on phase but has a gradual cut-off.  These trade-offs are why speaker processors give you a choice of a few different types of filters.

So phase in the case of filters refers to how the filter changes the phase (or delay) of the individual freq's that make up the full signal.  I have no idea how phase and the different filter types affect what one hears.  That would be a nice post!
Title: Re: Crossover question
Post by: luis Markson on April 12, 2011, 06:46:44 pm
As far as what phase is-

A continuous-time signal can be decomposed into a collection of sine-waves, each with a particular freq, amplitude and phase.  If you were to connect the mains output of a mixer to a scope or suitable laptop program and look while a band was playing you would see a rapidly changing display that wouldn't tell you much.  This is viewing in the time-domain: x-axis is time.  If you switched the scope to spectrum analysis you would see lots of vertical lines.  This is the frequency-domain, x-axis is freq.

So if you have a low-pass filter it will attenuate freq's above a cut-off.  BUT, it also can change the phase of freq's (phase can also be thought of as a delay of those freq's).  Some filters don't change phase much in the pass-band (Bessel), others do.  There are trade-offs with the different types of filters.  Bessel is great on phase but has a gradual cut-off.  These trade-offs are why speaker processors give you a choice of a few different types of filters.

So phase in the case of filters refers to how the filter changes the phase (or delay) of the individual freq's that make up the full signal.  I have no idea how phase and the different filter types affect what one hears.  That would be a nice post!

Would it be true to say then that the steeper the attenuation in cut band, the greater effect on phase on the pass band?
Title: Re: Crossover question
Post by: Art Hays on April 14, 2011, 12:35:12 am
Would it be true to say then that the steeper the attenuation in cut band, the greater effect on phase on the pass band?

I'm getting out of my depth (someone please help here).  After a little reading it seems the thing you want to avoid in filters is varying group delay in the passband.  Group delay is the rate of change of the phase with respect to freq.  So the phase vs. freq can change as long as it's linear (a straight line).  If all freq's go thru the filter with the same delay (group delay is the same for all freq's in passband) then the output signal will look like the input.  Otherwise there is distortion.

I found a post on the net that said group delay increases as the order of the filter (steepness of cutoff).  http://www.microwaves101.com/encyclopedia/groupdelayfilters.cfm
Likewise wikipedia says this about Linkwitz-Riley filters:
However, crossovers of higher order than 4th may have less usability due to their increasing peak in group delay around crossover frequency and complexity.

So as best I'm able to answer this question (and I'm on thin ice here, I don't work with filters every day) if you increase the db/octave (or order) of the filter then you make group delay worse in the passband. 

I guess the question is can you hear this?  Is there a conventional wisdom about which filter to use in a crossover (e.g the DR260 provides Bessel, Butterworth, Linkwitz-Riley) and how many db/octave?
Title: Re: Crossover question
Post by: Art Hays on April 25, 2011, 12:38:29 pm
Just an FYI- this thread http://forums.prosoundweb.com/index.php/topic,2045.0.html has lots more detailed tech info on filters.