4-way active via dsp,
FIR based driver-by-driver correction,
and complementary linear phase xovers
I've tried FIR correction on speakers already tuned either passively, analog active, or digital active...without any real success. Not worth the trouble ime/imo.
However, the driver-by-driver approach along with linear phase xovers has definitely been worth the effort. (I think the driver-by-driver approach is more important than the linear phase aspect.)
FIR both greatly facilitates driver-by-driver tuning, and also makes xovers vastly simpler by avoiding phase rotations and group delays.
I've used 8 channels of minDSP openDRC, PC convolution via JRiver,
and currently open architecture drag-and-drop Q-Sys, on a Core110f.
Interconnects are ethernet cables
I dont understand.... you mean the relative phase between two drivers - tweeter and midrange?
Cause we agree totally that relative phase between drivers need to be right - or else there will be boost and suckout - even though each driver indivudally plays nicely.
I recently replaced my passive system with an active system.
The passive system was a 1992 B&K ST-140 100 W stereo amp (class AB) driving a 3 way design which I built in 1993: A big 85 liter box with 12 inch woofer and Vifa drivers. It sounded great for 25 years... however the whole system was getting long in the tooth.
I was looking at new amps, and the ones I liked from NAD, Parasound, Rotel and others were all $1500 - $3000 USD... I also was very interested in getting back into speaker building (I had not built or designed a speaker since 2002). Since I needed both amps and speakers, I began to look at active speakers as a possibility.
I went with the Hypex fusion amps. If they had not been available, I probably would have used MiniDSP-HD and ICEpower amps, but the all-in-one nature of the Hypex product fit my needs. The Hypex nCore technology also had a great reputation for sound quality.
My finished system is described here
New active 3-Way, Hypex and SB
I want to point out that achieving great results with an active system is not necessarily easier than a passive system. A DSP-active system is not a magic wand. It is faster to make changes to the crossover, yes. But the designer must still know how to properly measure the loudspeaker system, and there are many ways to get measurements wrong. The designer must also understand enough acoustics physics and electrical filter theory to make good decisions. If I had not had the experience of designing and building multiple passive crossovers and measuring the results, I might not have been successful in designing my active filters. Or perhaps I would have gotten success eventually, but it would have taken a lot longer.
I also want to point out that DSP can do many things, but there are some important things it can not do.
DSP can easily adjust the time delay between drivers, and this makes filter design easier. It allows separate bass cabinets to be physically dispersed from the upper frequency unit. It makes horn drivers easier to deal with.
DSP can not correct for diffraction or reflection. There is no substitute for a low diffraction baffle, or a wave guide to control directivity.
DSP can easily correct for baffle step response, and it can correct most driver response abnormalities in the pass band and stop band. But if those abnormalities are the result of cone breakup resonance, the DSP can not remove the resonance. DSP can not lower the distortion of the drivers.
DSP can modify the bass response, such as in my case with a Linkwitz Transform. DSP can correct for room effects. But the designer must be able to make good measurements, and low frequency measurements are tricky. The result is only as good as the measurements that went into it.
So Yes, I am a fan of active systems. But we need to be aware of the limitations. We need to be aware that it still takes knowledge, experience, and good measurements to get a good result.
The passive system was a 1992 B&K ST-140 100 W stereo amp (class AB) driving a 3 way design which I built in 1993: A big 85 liter box with 12 inch woofer and Vifa drivers. It sounded great for 25 years... however the whole system was getting long in the tooth.
I was looking at new amps, and the ones I liked from NAD, Parasound, Rotel and others were all $1500 - $3000 USD... I also was very interested in getting back into speaker building (I had not built or designed a speaker since 2002). Since I needed both amps and speakers, I began to look at active speakers as a possibility.
I went with the Hypex fusion amps. If they had not been available, I probably would have used MiniDSP-HD and ICEpower amps, but the all-in-one nature of the Hypex product fit my needs. The Hypex nCore technology also had a great reputation for sound quality.
My finished system is described here
New active 3-Way, Hypex and SB
I want to point out that achieving great results with an active system is not necessarily easier than a passive system. A DSP-active system is not a magic wand. It is faster to make changes to the crossover, yes. But the designer must still know how to properly measure the loudspeaker system, and there are many ways to get measurements wrong. The designer must also understand enough acoustics physics and electrical filter theory to make good decisions. If I had not had the experience of designing and building multiple passive crossovers and measuring the results, I might not have been successful in designing my active filters. Or perhaps I would have gotten success eventually, but it would have taken a lot longer.
I also want to point out that DSP can do many things, but there are some important things it can not do.
DSP can easily adjust the time delay between drivers, and this makes filter design easier. It allows separate bass cabinets to be physically dispersed from the upper frequency unit. It makes horn drivers easier to deal with.
DSP can not correct for diffraction or reflection. There is no substitute for a low diffraction baffle, or a wave guide to control directivity.
DSP can easily correct for baffle step response, and it can correct most driver response abnormalities in the pass band and stop band. But if those abnormalities are the result of cone breakup resonance, the DSP can not remove the resonance. DSP can not lower the distortion of the drivers.
DSP can modify the bass response, such as in my case with a Linkwitz Transform. DSP can correct for room effects. But the designer must be able to make good measurements, and low frequency measurements are tricky. The result is only as good as the measurements that went into it.
So Yes, I am a fan of active systems. But we need to be aware of the limitations. We need to be aware that it still takes knowledge, experience, and good measurements to get a good result.
Exactly.... no DSP can correct for a 3D distortion. But liniear distortion - 1D - can be corrected. As soon as the sound has left the driver - the baffle and the room is in charge.
You might be able to smooth it out and make some measurements that makes one think otherwise - but it is impossible.
Not sure what you're saying/asking....
I'll take a stab .... by linear phase, I mean taking phase flat to zero.
First, doing so for each driver individually, through its passband and as far as possible into the critical summation regions with adjacent drivers.
This is mostly accomplished via minimum phase EQ's in and out of band...you know, fix mag and you fix phase too.
Then, with linear phase xovers and no more than timing and level matching, the entire summed response just falls together with beautifully flat mag and phase.
There is no trying to fix phase of a previously tuned box, like when input or 'global' FIR correction is used across the entire speaker.
Fix the pieces first, and there's nothing to fix afterward.
But like you say, relative phase totally matters....and nothing says phase has to be flat to have matching relative phase.
(That's why I said driver-by-driver tuning is more important than achieving linear phase.)
There are many excellent speakers with sloping phase, huh
The big thing with multi-way FIR and linear phase in my mind, is you get guaranteed results in the simplest most straightforward fashion I've yet to find.
I am re-reading my post above. I don't want to give the impression that I have any regrets about my decision to go active. I have none. My system has exceeded every expectation. Sometimes it is so real it is spooky. I am quite confident that I could not have achieved this level of performance in a passive system, with these drivers in these cabinets. It takes a great deal of experience to design and build a large, excellent 3 way speaker... experience which I do not have. But with an active DSP system, my knowledge and experience was sufficient to achieve an excellent result.
Sorry for the late reply.
I don't have an 18" woofer.
My 3 way system is a 0.75" tweeter a 4" mid and a 10" sub woofer.
The 4 way system gets an other 2" mid speaker.
E9000 I would like the measurements of your bass bottom....I want to do something similar3 way active. FIR filters correcting phase and response for each driver designed in rephase. Reaper as a vst host fed by foobar. Using a Behringer 8 channel usb sound card as output.
DIY Pass Aleph 5 on Seos 12 horns, Crown PSA-2 on 10" mid and 18" woofer......now.....
and dont want cabinet mass....and your bass cab. seem shallow....I like...if you would not
mine posting the size of the bottom cabinet and are you happy with the 18"....Was think
of 15" for punch.....whats your feeling on the 18".....
thanks gray
^^ @grayxxx, a pro 18"driver will excel for midbass. I have used JBL2242 in the past and I have a friend with a 5 way active system that uses 18"Maelstrom subs, and PHL 18/10/6 drivers with a Morel dome tweet, but I forget which. It's all active with the M18s in separate enclosures. Many, many people have heard this system and commented on how excellent it is at all SPLs, and believe me, it will play clean to very high SPLs. Dynamics and imaging are superb.
My current system uses a pair of AE TD15 drivers/ch in the L/R and 1 TD15/ch for each of the 4 surrounds. I designed it as a 2ch system and added the surrounds later. It's all active with a custom PC doing the xover with an external 8ch pro soundcard and the surrounds using modified DCX2496s.
I would never go back to small, low efficiency passive systems again.
My current system uses a pair of AE TD15 drivers/ch in the L/R and 1 TD15/ch for each of the 4 surrounds. I designed it as a 2ch system and added the surrounds later. It's all active with a custom PC doing the xover with an external 8ch pro soundcard and the surrounds using modified DCX2496s.
I would never go back to small, low efficiency passive systems again.
In the L/R a Unity, so they're a 3 way, and in the surrounds a Faital 8HX200 coax.
thanks...Brett for the note............
I suppose its where this all falls...If the below is reference and Im not sure I would use those numbers..Im going to do separate cabinets...Low-Mid-High-and poss. a ESS
ESS Heil Air Motion Transformer AMT Tweeter with Horn driver also. Anyway just thinking
of punch and the 15". I could mount 18" in a wall I had thought of that...Im sure the sys. sounds good...What frequency was the 18" range....???
thanks ..........
Sub-bass20 to 60 HzBass60 to 250 HzLow midrange250 to 500 HzMidrange500 Hz to 2 kHz
This site act. gives the sample pitch...
Audio Spectrum Explained - Teach Me Audio
I suppose its where this all falls...If the below is reference and Im not sure I would use those numbers..Im going to do separate cabinets...Low-Mid-High-and poss. a ESS
ESS Heil Air Motion Transformer AMT Tweeter with Horn driver also. Anyway just thinking
of punch and the 15". I could mount 18" in a wall I had thought of that...Im sure the sys. sounds good...What frequency was the 18" range....???
thanks ..........
Sub-bass20 to 60 HzBass60 to 250 HzLow midrange250 to 500 HzMidrange500 Hz to 2 kHz
This site act. gives the sample pitch...
Audio Spectrum Explained - Teach Me Audio
Going from memory, <30, 30-200, 200-500, 500-2k, >2k.
The new 5 way I'm building after the move, which is mostly designed and partly built, will be <20, 20-60, 60-350, 350-1k, >1k. All FLH above 60Hz.
There's another advantage of active/dsp over passive that specifically applies to horn systems with big drivers and low crossover points: sensitivity.
Complex passive filters easily swallow 3 dB (or more), which is significant, as we all know.
The noble craft of voicing a loudspeaker system - which requires many years of experience, is lost by going active. However, the benefits are manifold and undeniable.
Complex passive filters easily swallow 3 dB (or more), which is significant, as we all know.
The noble craft of voicing a loudspeaker system - which requires many years of experience, is lost by going active. However, the benefits are manifold and undeniable.
Last edited:
My experience:
I tried DSP, the flexibility when tuning your system and speed at which you could get to a good measured response is impressive, the sound quality, for me, is not.
Minidsp (4x10HD) is not particularly quite, it can produce reasonable results with digital sources, but if you have a TT just forget it, the ADC side is rubbish and your analogue source will take a BIG step backwards in sound quality.
Hypex DLCP has less noise than minidsp and the ADC is slightly better for your TT, but still a step backward. I kept the hypex for a while to help develop an analogue active circuit, then sold it on.
The manufacturers seem to concentrate on digital media and the analogue side is almost a token effort. So if you have a TT, expect to be disappointed with these, I can't comment about other makes. It seems a good ADC is either difficult or expensive to implement, so bear this in mind if you use an analogue source.
I tried DSP, the flexibility when tuning your system and speed at which you could get to a good measured response is impressive, the sound quality, for me, is not.
Minidsp (4x10HD) is not particularly quite, it can produce reasonable results with digital sources, but if you have a TT just forget it, the ADC side is rubbish and your analogue source will take a BIG step backwards in sound quality.
Hypex DLCP has less noise than minidsp and the ADC is slightly better for your TT, but still a step backward. I kept the hypex for a while to help develop an analogue active circuit, then sold it on.
The manufacturers seem to concentrate on digital media and the analogue side is almost a token effort. So if you have a TT, expect to be disappointed with these, I can't comment about other makes. It seems a good ADC is either difficult or expensive to implement, so bear this in mind if you use an analogue source.
Still active but looking towards passive.
My system has Three attenuators. on DAC (1) which i believe is analog, next is on processor input gain (2), third is after the processor (3) and these two i must believe they are digital.
these amps has minimum gain of 22db which means if i look netflix volumes are (1) -18db, (2) 0db (3) 0db
Recently cause it is attenuate analog signal which goes into processor to chop into 4-pieces i´ve attenuate after processor gain and on netfix case it goes (1) 0db (2) 0db (3) -18db.
And the most last audioautism i have on my mind cause my processor has AES/EBU in i could go less AD/DA with RME ADI-2 Pro FS R Black Edition which has attenuated AES/EBU i could do the same fully digital.
Yeah let´s see about that passive side when i start to measure things more properly, current bass needs -12db bafflestep to get right amount of bass
My system has Three attenuators. on DAC (1) which i believe is analog, next is on processor input gain (2), third is after the processor (3) and these two i must believe they are digital.
these amps has minimum gain of 22db which means if i look netflix volumes are (1) -18db, (2) 0db (3) 0db
Recently cause it is attenuate analog signal which goes into processor to chop into 4-pieces i´ve attenuate after processor gain and on netfix case it goes (1) 0db (2) 0db (3) -18db.
And the most last audioautism i have on my mind cause my processor has AES/EBU in i could go less AD/DA with RME ADI-2 Pro FS R Black Edition which has attenuated AES/EBU i could do the same fully digital.
Yeah let´s see about that passive side when i start to measure things more properly, current bass needs -12db bafflestep to get right amount of bass
The subtle sense of irony was clearly unnoticed and no, I've never developed a passive crossover, only upgraded existing ones.
However, I happen to know such an artisan and appreciate his talent and expertise to 'compose' passive filters.
Those are expensive and time consuming exercises though. At least 1 member of this forum shares the same passion.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.