Sonce, the test comparo was blind, as i knew not which was which. I was also not the only person that felt that way, adamantly. Think what you want, but passives can do things actives cant. Shunting or parallel/conjugate networks are not applicable to DSP xovers, so all they can do is attenuate and delay.
Think what you want, but actives (with DSP) can do things which passives can't, and on top of that, actives are much better than passives. Not only my opinion, but of several listeners.
Why would I need a "shunting or parallel/conjugate network" with active DSP loudspeaker?
Are you referring to a Zobel network? It is not a passive crossover.
Why would I need a "shunting or parallel/conjugate network" with active DSP loudspeaker?
Are you referring to a Zobel network? It is not a passive crossover.
yes there is nothing preventing anybody from using passive components with DSP crossover.
personally i see no reason to but if somebody wanted to flatten the impedance of a driver they could do it running off a DSP crossover.
personally i want to use 2nd order passive high pass protection 1 octave below DSP crossover, but without any impedance flattening measures.
i think to just use a capacitor is not ideal because it presents an open circuit to the driver at low frequencies leaving it unable to damp its resonance. with an inductor shorting the driver now driver can damp itself.
you may ask why would it need to damp itself if the resonance is out of the passband ? well because a compression driver will be driven like a microphone by the bass entering the horn ... so for example i may run DSP crossover at 1 khz and passive 2nd order high pass at 500 hz but the horn may have resonance at 300 hz which will be energized by woofer energy entering the horn and getting focused by the horn into the driver ...
i made all of this up you may want to check my logic ...
Sonce-
If you mean FIR and IIR, I agree; elsewise, DSP or passive both can do pretty much everything else- except active using conjugate type/style/effective filters; impedance correction, notches for stout breakup (series LC) or Fs comp, etc. (Yes- passives can use a lot of parts and get it done.) The reason this is a problem is that breakups sometimes cannot just be attenuated via series massive resistance like a DSP can provide. Energy storage/breakup sometimes have to be shunted out to be eliminated completely and the DSP attenuation cannot do this adequately. I have seen and done this myself empirically and found this to be true.
I would not use conjugates with an active loudspeaker as my example implies, because they can't do it justice and achieve the goal I would want. Simply implying to add passive parts to an active design was not what I was inferring. I was saying you can't get the results I want in this criteria with an active design. The setup is not capable. To suppress steeply via a attenuation slider will just induce ringing which is an unavoidable artifact once the attenuation is enough. I don't want (pre)ringing of the response.
axi- you can add conjugates to an active xover system as you entertain, however, just adding it across the driver will yield impedance comp only. The usage of these circuits involves a voltage divider upstream for them to accomplish anything else; ie a coil or cap. Having an additional 12dB highpass I'm guessing could be beneficial to steepen the out of band response slope.
If you mean FIR and IIR, I agree; elsewise, DSP or passive both can do pretty much everything else- except active using conjugate type/style/effective filters; impedance correction, notches for stout breakup (series LC) or Fs comp, etc. (Yes- passives can use a lot of parts and get it done.) The reason this is a problem is that breakups sometimes cannot just be attenuated via series massive resistance like a DSP can provide. Energy storage/breakup sometimes have to be shunted out to be eliminated completely and the DSP attenuation cannot do this adequately. I have seen and done this myself empirically and found this to be true.
I would not use conjugates with an active loudspeaker as my example implies, because they can't do it justice and achieve the goal I would want. Simply implying to add passive parts to an active design was not what I was inferring. I was saying you can't get the results I want in this criteria with an active design. The setup is not capable. To suppress steeply via a attenuation slider will just induce ringing which is an unavoidable artifact once the attenuation is enough. I don't want (pre)ringing of the response.
axi- you can add conjugates to an active xover system as you entertain, however, just adding it across the driver will yield impedance comp only. The usage of these circuits involves a voltage divider upstream for them to accomplish anything else; ie a coil or cap. Having an additional 12dB highpass I'm guessing could be beneficial to steepen the out of band response slope.
Phase is like delay only for continuous, single sine wave. But we don't listen to sine waves (lololol).
To put long story short, here is an illustration. Phase can turned forward. If you turn it 36000° at 100 Hz (1 second, 100 full cycles), the waveform doesn't change at all. The timing is the same, the signal excites the speaker exactly at the same time as with no phase shift. Phase can also be turned back, negative degrees. If you turned it backwards, -36000° at 100 Hz (-1 second), music starts playing one second before you click play button. This is false.
Dsp delay is always >= 0. Also dsp can delay the whole signal without any phase shift at any frequency. Actually delay is not frequency dependant at all.
Yeas, I mean FIR and IIR.
For passive attenuation of cone breakups, I disagree. Passive LC(R) shunt notch filter has frequency response which can be exactly cloned with DSP, only active DSP will damp cone breakups/energy storage much better because the driver is directly connected/shunted to an amp (practically - zero resistance). In passive networks (LC shunt) there is always some inductor resistance, and shunt notch network can not work alone - it must have filter before it, so there is some serious (frequency depended) resistance between amp and driver, which prevents good damping.
My empirical data contradicts yours.
If you have frequency response with deep notch done with passive crossover, it will (pre)ring exactly as much (or worse), as active DSP with cloned frequency response. Theoretically and practically, they are the same mathematical functions and they will behave the same.
Looks like many here forget, that we are discussing loudspeaker drivers, which have intrinsic inductance and capacitance, as well as series resistance. So, eg. a serial capacitor won't create open circuit for the amp.
One of the best features of multiway-dsp is to be able to adjust delay. For example a typical subwoofer-main setup asks for delay to the main speaker!
I have tried, but I can't hear difference between capacitors or resistors (of same value) in xo circuits. But I can hear spl and timing differences, which are easy to spoil or cure with dsp-xo.
Like axiperiodic, I'm too poor, lazy and dumb to invest on piles of analog/passive xo components and to learn xo mathematics or simulation software. Instead I rather easily learned to take measurements with REW and to operate dsp-xo. It took only about half a year with help from forum members!
One of the best features of multiway-dsp is to be able to adjust delay. For example a typical subwoofer-main setup asks for delay to the main speaker!
I have tried, but I can't hear difference between capacitors or resistors (of same value) in xo circuits. But I can hear spl and timing differences, which are easy to spoil or cure with dsp-xo.
Like axiperiodic, I'm too poor, lazy and dumb to invest on piles of analog/passive xo components and to learn xo mathematics or simulation software. Instead I rather easily learned to take measurements with REW and to operate dsp-xo. It took only about half a year with help from forum members!
Believe all the gobily gook you want. "I believe the increase in detail is a function of faster reactions that create time spaces and separate between the different instruments/sounds". How do instruments that are all combined into one continuous varying voltage have time spaces between them? When did people start believing in there own fantasies over science and technology. GR research (the moron tests speaker cable as an antenae with a receiver to pick the best cable) makes money selling you parts you don't need. If he new how to make speakers he'd be selling those not dubious parts for real speakers.
Last edited:
Phase is not the same as delay or you wouldn't need both to describe the wave function y(x,t)=Asin(kx−ωt+ϕ).
Show me a passive delay that delays all freqs. 100ms.
What is the right metaphor for home audiophile systems? Hard to specify "value" or "excellence" or "optimal" or "DIY-ability" parameters.
As always, we're clueless about some of OP's interests and home configuration. Maybe folks with major systems questions should complete a questionnaire before tasking the forum for answers.
For a bunch or reasons, important to keep low freq out of small drivers (and their small amps) used for mid-range bands. But for sure, hopeless to attempt XO around 100 Hz with old-fashioned passive XO - simply not feasible to have slopes sharp enough to be useful.
As always, we're clueless about some of OP's interests and home configuration. Maybe folks with major systems questions should complete a questionnaire before tasking the forum for answers.
For a bunch or reasons, important to keep low freq out of small drivers (and their small amps) used for mid-range bands. But for sure, hopeless to attempt XO around 100 Hz with old-fashioned passive XO - simply not feasible to have slopes sharp enough to be useful.
i don't need additional slope. i just need a protection cap in case of accidents like DSP crossover crashing. however i feel like using a series capacitor by itself with nothing else in the circuit may create unintended consequences such as allow bass energy from woofer to enter the compression driver horn, get focused by horn and amplified by horn resonance and modulate the compression driver which will be unable to damp the movement of the diaphragm because at resonance frequency the cap will be an open circuit and drivers can only damp fundamental resonance electrically with terminals that see a short circuit
an amplifier normally presents a short circuit to the driver, which is ideal and many people just run tweeters directly into the amp ... but this leaves the tweeter vulnerable to accidents like DSP crashing etc ...
i am trying to work out a solution where i have backup passive protection for tweeters without creating problems in the process ... i think a 2nd order HP filter 1 octave below crossover point may be optimum in terms of protection, simplicity and not creating unintended issues ...
JukkaM
I've shown here a regular filter with input and output waveform, showing phase behaving as delay at one frequency.
All that is then needed is to have the right amount of relative phase shift per frequency to fit the delay characteristic.
I've shown here a regular filter with input and output waveform, showing phase behaving as delay at one frequency.
All that is then needed is to have the right amount of relative phase shift per frequency to fit the delay characteristic.
Fixed delays fall toward zero degrees at DC.
No, it only appears this way if you subtract the delay.. which then misses the point.
100ms, that's 35m. I'm sure we'd have fun with that... the assertion was that delay isn't possible passive.
I've done the same many times and have come to the opposite preference. Plus there are things I can do with a DSP that can't be done easily, or at all, with a passive xover. This sort of comparison limits ultimate system performance to what a passive can do.
Back on topic. Full respect wgh64 for honestly reporting results and good effort on keeping an open mind during your experimentation!
Don't be disappointed and I would ask you to view your results in the context of the equipment used for the test. All I mean is the kit may not be capable of taking advantage of the improved components, or that those components were not the limiting factor in the speaker at this point.
I've done similar (and continued to go a lot further) and experienced outstanding results. For me the greatest crossover improvement came from replacing the two iron core inductors for air core parts. The capacitors definitely made an improvement to clarity and separation - things previously got a bit muddy and smeared together on tracks with a lot going on.
I've also tried things that haven't worked out for the better and removed them (Zobel network on the original tweeter).
So stay with the modding and don't take any notice of people telling you that you should only do this, or the opinionated referring to 'shills', etc. They detract from the forum and most of them clearly don't know what they're talking about and certainly have no experience with decent quality kit.
Don't be disappointed and I would ask you to view your results in the context of the equipment used for the test. All I mean is the kit may not be capable of taking advantage of the improved components, or that those components were not the limiting factor in the speaker at this point.
I've done similar (and continued to go a lot further) and experienced outstanding results. For me the greatest crossover improvement came from replacing the two iron core inductors for air core parts. The capacitors definitely made an improvement to clarity and separation - things previously got a bit muddy and smeared together on tracks with a lot going on.
I've also tried things that haven't worked out for the better and removed them (Zobel network on the original tweeter).
So stay with the modding and don't take any notice of people telling you that you should only do this, or the opinionated referring to 'shills', etc. They detract from the forum and most of them clearly don't know what they're talking about and certainly have no experience with decent quality kit.
Of course that's just an opinion, and a popular one. I was once seduced by the hyper real technicolor pop of active crossovers - "like wow man!" To me there are the Jessica Rabbit of audio systems. Our fascination is understandable. We can't help it, they're built that way. 😀
Wow, way to draw a long bow and insult the many who have done extensive testing and measuring.