Crossover question

[Wayne A. Pflughaupt]

Anyone know what kind of crossover alignment has a 24-dB slope and is 3 dB down at the crossover frequency? I know a Linkwitz-Riley alignment would be 6-dB down at the crossover frequency, so what is that one, a Butterworth alignment?

Regards,
Wayne

[F1 fan]

Wayne A. Pflughaupt wrote:

a Butterworth alignment?

Correct.

[Wayne A. Pflughaupt]

Thanks Fred. Do the standard alignments apply to digital filters as well? It seems like in the digital domain they could do things irrespective of or to compensate for the artifacts you get with analog filters...

Regards,
Wayne

[F1 fan]

Wayne A. Pflughaupt wrote:

Thanks Fred.

Your welcome

Do the standard alignments apply to digital filters as well?

Yes,it is the Q of the filter that determines whether it is a Butterworth,Linkwitz-Riley ,Bessel etc.So as long as the Q is.707 then you will get a butterworth response regardless of whether it is done in the analog or digital domain.Change the Q and it is no longer a butterworth alignment.

[jeffreybehr]

Uhh, guys, the crossover frequency (or Filter Point) is defined as the frequency that is indeed -3dB, so there's no such thing as an 'alignment (that) would be 6-dB down at the crossover frequency'. ALL alignments by definition are down 3dB at the filter point. Of course there are different slopes and different Qs, but the FP is the FP.

[Wayne A. Pflughaupt]

Okay guys, who wants this one?

Regards,
Wayne

[jeffreybehr]

Hmm...just determined* that a 4th-order L-R filter is composed of 2 2nd-order filters in series and that the FP of such could be defined as the FP of ONE of the 2nd-order filters. Wow. I wonder what definition such people as Vance Dickason writing in The Loudspeaker Design Cookbook uses--the -3dB point or the the -6dB point? Seems rather confusing and unnecessarily so IMO.

While I'm pleading ignorance, here's a question for the educated among us--does a filter implemented digitally such as all of them in my dbx-brand DriveRackPA speaker-management system still have phase errors?

BTW I thank you all for not jumping on me for my only-semiinformed statement.



* I'm a retired accountant and an active audio hobbyist and have NO training in electronics, so I have a good excuse to be uninformed.

[F1 fan]

As you have found out the L-R network is -6db at the xover point.This means that the combined response of both the hi pass and a low pass filters will sum to flat a response( assuming the drivers are in phase).The resulting summed response when using butterworth networks will result in a 3db peak at the xover point ,again assuming drivers are in phase.

jeffreybehr wrote:

-does a filter implemented digitally such as all of them in my dbx-brand DriveRackPA speaker-management system still have phase errors?

By phase errors do you mean the relative phase differences between the drivers due to physical offset ie. tweeter acoustically closer to listener than the mid/woofer?If so then the answer is yes,unless the unit also has the capability of adding the appropriate amount of delay to the tweeter.Delay can also be added in the analog domain via an allpass network.

[jeffreybehr]

"By phase errors do you mean the relative phase differences between the drivers due to physical offset ie. tweeter acoustically closer to listener than the mid/woofer?" No, don't mean the timing errors inherent in different driver spacing.

I understand that frequencies within the slope of a filter implemented 'analogly' are not only down in level but also are delayed in time. It's this latter delay...or phase error...that I'm asking about in a digital filter.

[F1 fan]

Oh ,you are talking about the phase shift that happens with the slope of the filter.The least phase shift occuring with the shallow 1st order network and more shift occurs as you progressively increase the slope.It has been debated for years whether a speaker using a minimum phase type crossover(ie.Vandersteen,Theil) actually has a sonic advantage.The only truely phase accurate xover is the 1st order butterworth but it has other issues that rear their ugly head.However I believe that there are some digital xovers that can improve the phase response of the system.I think the DEQX digital xover used in the little NHT active system does this.

To answer the question ,whether done digitally or analog a crossover with the same frequency and slope will have the same amount phase change in the drivers roll off region.