Don't forget: with a current drive amp the active speaker gets a plus of lower distortion.
If voltage amp is used any resistor like 0.47 to 4 ohms with the drivers in line reduces distortion converting the voltage a little bit to current driven amp.
The combination of raw passive crossover with overall fir correction would marry both techniques if 12db slopes are enough.
If voltage amp is used any resistor like 0.47 to 4 ohms with the drivers in line reduces distortion converting the voltage a little bit to current driven amp.
The combination of raw passive crossover with overall fir correction would marry both techniques if 12db slopes are enough.
You could go hybrid and run the mid and tweeter from the tube amp with passive XO, and active XO between mid and woofer, and use a more powerful solid state amp to drive the woofer (better damping factor and tighter bass).
Very cool! Have you got a write-up of this somewhere?
I've been toying with this idea, using coaxials, each with ~+/- 60 degree beamwidth from 200Hz to 20KHz, and 3 coaxes per enclosure.
Given you love your SET amplifier the main advantages of DSP active crossover which revolve around dry technical performance makes a pretty unnatural combination. A passive crossover with boutique components (I know little about the subject) would seem a more sympathetic combination.
If you do opt for an active crossover using DSP it would be very strange to lower the technical performance of the overall speaker by using SET amplifiers. Commercial speakers with active crossovers normally promote the speakers overall technical performance without drawing much if any attention to the amplifier modules which tend to be whatever is cheap, reliable and efficient with inaudible distortion in use. It is a bit different in the DIY world with people getting enthusiastic about expensive amplifier/signal processing hardware or perhaps the cheapest depending on the project. Few though are going to attribute the quality of the overall sound with the amplifier modules to any significant degree. Quite different to SET amplifier enthusiasts.
Given you drivers and the SET amp, I would do a hybrid approach.
Use 2 outs of the DSP to the SET amp to drive just the Morel domes….they only dip to 6.5ohms and should be an easy load on the amp.
Use the remaining two outs for each side of your mid and woofer. The DSP can be set for the low pass. Design and build a passive network for the mid/ woofer section. You can still use the DSP parametric EQs to smooth the overall response and set your subsonic high pass. Now you’ll only need 1 more stereo amplifier……a nice 500w ICE amp build.
Use 2 outs of the DSP to the SET amp to drive just the Morel domes….they only dip to 6.5ohms and should be an easy load on the amp.
Use the remaining two outs for each side of your mid and woofer. The DSP can be set for the low pass. Design and build a passive network for the mid/ woofer section. You can still use the DSP parametric EQs to smooth the overall response and set your subsonic high pass. Now you’ll only need 1 more stereo amplifier……a nice 500w ICE amp build.
Crossing them over at 250Hz in 28 litres box, seems unnecessary large.
Being fond of tube amps goes hand in hand with passive crossovers.
The trouble with passive filters is substantial. Not everyone is prepared
to manufacture a fair deal of inductors for trying out and the work involved
with connecting parts is tiresome. I understand if folks would rather simplify
by resorting to dsp.
Last edited:
Sheeple, when you say "pre FIR units" does that imply that the DSP chipsets have changed?
The hybrid option mentioned by several folks sounds really interesting. I also like the Hypex plate amps, mentioned by olsond3. And at least now I have a better list of options than I did previously – though probably more questions, too! Thank you!
Andy, I imagine my preoccupation with the SET amp seems counter-intuitive. And it probably is. But I love the precise imaging it offers. Of the (admittedly, not that many) amplifiers I tested on my system, my Cary SLI-80 provided the most precise imaging, by far. I've had people jump out of the listening position, unnerved on some tracks because they heard sounds in very precise and immediate proximity. None of the solid states (Krell, Adcom, Creek, Marantz, Onkyo, Amp Camp Amp, early Tri-Path, and a couple of NADs) that I’ve tried could quite match the imaging. So that’s my attachment to the SET, not specifically its acoustic signature or tube sound. Though to be fair, I doubt I can replicate that level of imagining in my current listening environment. 🙁
Lojzek, thanks for pointing that out! You’re right! I went back to check my maths and it turns out I entered some incorrect values somewhere, so I don’t need 28 liters for that mid. I still intend to give it ample space to breathe for the low Qtc transient response, but not having to stick to the larger volume gives me a bit more flexibility in the cabinet design. 🙂
Andy, I imagine my preoccupation with the SET amp seems counter-intuitive. And it probably is. But I love the precise imaging it offers. Of the (admittedly, not that many) amplifiers I tested on my system, my Cary SLI-80 provided the most precise imaging, by far. I've had people jump out of the listening position, unnerved on some tracks because they heard sounds in very precise and immediate proximity. None of the solid states (Krell, Adcom, Creek, Marantz, Onkyo, Amp Camp Amp, early Tri-Path, and a couple of NADs) that I’ve tried could quite match the imaging. So that’s my attachment to the SET, not specifically its acoustic signature or tube sound. Though to be fair, I doubt I can replicate that level of imagining in my current listening environment. 🙁
Lojzek, thanks for pointing that out! You’re right! I went back to check my maths and it turns out I entered some incorrect values somewhere, so I don’t need 28 liters for that mid. I still intend to give it ample space to breathe for the low Qtc transient response, but not having to stick to the larger volume gives me a bit more flexibility in the cabinet design. 🙂
The imaging effect shouldn't be too hard to reproduce using DSP and an amplifier that doesn't audibly distort the signal. Our brains use several mechanisms to locate sound one of which is our Head Related Transfer Function (HRTF) which is effectively a distorted frequency response which varies with direction and elevation. The frequency response of a typical SET amplifier will vary with speaker load particularly through the crossover region (they have other forms of audible distortion that likely don't distort the imaging). The bumps and dips in the frequency response have similarities with those of an HRTF making it likely the brain interpreting part of them as spatial information. Something similar seems to happen with the edges of small speakers which are often claimed to image well despite (because of?) the distortion introduced by significant diffraction off the close edges. One can experiment with this by placing books on or around small speakers and listening for changes in the imaging. In the case of a SET amplifier one could measure the distortion of the frequency response with a particular speaker, reproduce that distortion with DSP and then listen via an audibly neutral amplifier to see if the change in imaging is reproduced.
Now most SET amplifier enthusiasts are not interested in the scientific method and what is going on in a scientific sense and instead prefer to believe something like that although their amplifiers distort (most accept this) they also extract more real/musical/whatever information from the signal in some way they can't quite specify and that is likely unknown to science. There is nothing much wrong with believing stuff like this if adds to the fun of a hobby but putting aside scientific knowledge tends to go hand in hand with putting aside things like signal processing and modifying transfer functions. Not saying it can't be done just that marrying two outlooks that are to some extent in conflict is odd and unusual.
Have you "rolled" the small signal tubes 12AU7/12AX7/6DJ8/etc, through the range of usual suspects Telefunken, Siemens, French Mazda, Amperex, Mullard, etc.? They are more responsible than power tubes for high frequency response and detail resolution, including localization/imaging. When I played with this and other tube questions (maybe 14 years ago) the Siemens silver-plate 12AU7 had the most (incredible) pin-point imaging, but Telefunken sounded more natural. Yes this suggested to me FR had a lot to do with it. Another discovery was that it took six hours for the small tubes to fully warm-up and settle down soundwise. (Equipment used: Unico hybrid amp, Fostex F120A plus electrostatic supertweeter, silicone glove.)
Tried and true, even before DSP was a thing. The crossover itself doesn't require the same level of subjective input or have the same degree of uncertainty as voicing, and is simpler than managing the room EQ.. so it's fairly natural to set the crossover, and work on the other two with the help of EQ.
Here’s my scenario,
I’ve had passive systems my whole life. First store bought, then diy. I’ve built amps, a dac, several different speakers, power supplies and many interconnects along the way. Generally I’ve been very satisfied with the sound of these components. Recently I purchased a pair of active speakers to hear what if any difference I could hear with my old ears. This allowed me to do a side by side comparison (passive to active) and for me the difference is not subtle. The active system has a clarity and resolution that has been missing with my passive system. I used to listen at low levels, rarely getting loud as this brought forth issues with room interaction. Dsp has eliminated that issue completely and allowed me to experiment with subtle changes to shape the sound more to my liking. I for one will not return to a passive system as sound quality has improved dramatically. Please consider that In the end the cost difference may not be so great. This is just my opinion, offered humbly.
Jcris
I’ve had passive systems my whole life. First store bought, then diy. I’ve built amps, a dac, several different speakers, power supplies and many interconnects along the way. Generally I’ve been very satisfied with the sound of these components. Recently I purchased a pair of active speakers to hear what if any difference I could hear with my old ears. This allowed me to do a side by side comparison (passive to active) and for me the difference is not subtle. The active system has a clarity and resolution that has been missing with my passive system. I used to listen at low levels, rarely getting loud as this brought forth issues with room interaction. Dsp has eliminated that issue completely and allowed me to experiment with subtle changes to shape the sound more to my liking. I for one will not return to a passive system as sound quality has improved dramatically. Please consider that In the end the cost difference may not be so great. This is just my opinion, offered humbly.
Jcris
Maybe a bit late, but I am also in for a hybrid aproach. Mid/high with a passive crossover and actively to lows. I am running a setup like this now - 300B SE for the upper two bands and Denon PM1500AE for the lows. And EQ correction where needed (plus time alignment).
Full Disclosure: We manufacture DSP crossovers including the dspNexus 2/8. I encourage everyone to check it out. I want to address neurotopia's query in broader terms.
There are many good reasons to use a decent DSP solution over a passive solution. On the Danville website, I did a "Blackboard Session" on this topic
https://danvillesignal.com/tbs-1-why-dsp-crossovers
Most of my arguments are in the blog, but I do want to elaborate on one: amplifier consideration. neurotopia has a good pair of tube amplifiers. I have no quarrel with good tube amplifiers but they don't tend to be the best for low frequency drivers. The reason is that you want the back emf of the drivers to see a short circuit. This is why damping is important. Vacuum tube amplifiers are generally poor in this regard because of output transformers. In an active system, the amplifier is essentially driving the speaker directly. This means the damping will be very high.
Regardless, when you are using a passive system, you place a huge demand on the amplifiers. In an active system, all the amplifiers perform much better since they are asked to do a much easier job. Modern Class D amplifiers from a variety of sources are surprisingly good today and they are not expensive..
At the recent CAF show, we did a demo with an old pair of Maggies (MG10QR) and a GR subwoofer. The DSP crossover allowed us to time align the system and optimize the original crossover. In our demo, we bypassed the original passive crossover. The crossover was much more complex than the original crossover and kept each section of the panel operating in a linear region. We crossed the Maggies at 160 Hz, which reduced the excursion of the panel and really helped the overall system. I think we surprised a more than a few people with the results.
I will claim that no matter how good the components are in a passive loudspeaker or the overall design of the crossover, a well designed DSP active system will always be superior. It's one of the areas where you can make significant improvements in existing systems.
Regards,
Al Clark
Last edited:
My 2 cents. Built one passive and one active pair with the same scanspeak illuminator drivers. Same box. Passive crossover from Troels.
Sound wise, I am still chasing that passive crossover sound with actives, but the joy along the way is immense. I would consider building passive crossover only for conservation/sell purpose 🙂 My next goal is to design digital filters that mimic passive crossovers, and thus save a ton of time with soldering iron.
Sound wise, I am still chasing that passive crossover sound with actives, but the joy along the way is immense. I would consider building passive crossover only for conservation/sell purpose 🙂 My next goal is to design digital filters that mimic passive crossovers, and thus save a ton of time with soldering iron.
"My next goal is to design digital filters that mimic passive crossovers,"
Here's a tool to help with that: https://www.beis.de/Elektronik/Filter/AnaDigFilt/AnaDigFilt.html#AnaDigFilt
Here's a tool to help with that: https://www.beis.de/Elektronik/Filter/AnaDigFilt/AnaDigFilt.html#AnaDigFilt
I have heard a few variations of this
Of course, you can make IIR filters that are very similar to passive analog filters. It can be a little harder than you think given interaction with the driver impedances and the problems of non ideal components. The better choice is to improve on the filters (and probably use more). In general, you want to make the direct sound flat. You should also time align the drivers. This is trivial in a DSP implementation.
Al Clark
Danville Signal
I have heard a few variations of this idea. One was a product made by ATC. They were in a room adjacent to our demo room at Axpona. The had an "Identical" loudspeaker where one was passive and the other active. I had nothing to do with their design, but their proximity to my room was helpful. It really wasn't hard to tell which was which - the active version was clearly much better.
Of course, you can make IIR filters that are very similar to passive analog filters. It can be a little harder than you think given interaction with the driver impedances and the problems of non ideal components. The better choice is to improve on the filters (and probably use more). In general, you want to make the direct sound flat. You should also time align the drivers. This is trivial in a DSP implementation.
Al Clark
Danville Signal
Thanks! This is what makes this forum the most awesome place on the internet 🙂
My setup is somewhat unusual, I use Matlab on my PC for filtering. I was trying FIR for the last couple of years, until I did direct comparison between IIR and FIR of the same magnitude response. I played single driver at the time with a simple second order filter. I could switch filters with zero pause in between. The result was subtle but enough for me to be temporarily mad at FIR. It did something to the decay/air in the sound that made sound a tiny bit less natural. We are talking nuances here, but these drivers are unforgiving. I am now experimenting with IIR to see where how far it can get me.
Anyone with similar experience with FIR/IIR? Is it possible that we perceive linear phase of the filter as less natural?
My setup is somewhat unusual, I use Matlab on my PC for filtering. I was trying FIR for the last couple of years, until I did direct comparison between IIR and FIR of the same magnitude response. I played single driver at the time with a simple second order filter. I could switch filters with zero pause in between. The result was subtle but enough for me to be temporarily mad at FIR. It did something to the decay/air in the sound that made sound a tiny bit less natural. We are talking nuances here, but these drivers are unforgiving. I am now experimenting with IIR to see where how far it can get me.
Anyone with similar experience with FIR/IIR? Is it possible that we perceive linear phase of the filter as less natural?
If FIR filters are correctly made they must affect decay/air less than IIR filters as FIR filters had pre ringing and generally less post ringing than IIR. If decay was increased it come from increased post ringing and probably mean not very correctly made FIR filter.
@IVAudio
I used minidsp 2x4, behringer deq2496 all with iir on a diy fullrange driver and was very happy with it.
But wondered myself if fir could be even better. Never tested it as a fullrange has minimal phase behavior and EQed to the point made it phase linear, too.
FIR is known to have pre ringing in the time domain and some claim this is audible. As you made A/b comparison with fast switch I trust your hearing experience.
However a time corrected multi way speaker with FIR will profit.
My fullrange driver could withstand a simultaneous playing of original instruments against their recording being played on the boxes (no need to switch between sources).
After that I knew: loudspeakers can sound like the original if really adjusted to the point - measurement assisted.
But you have to invest some time for doing dsp programming work - quite similar to making a passive crossover.
Real professionals for making passive crossovers can make fast developments.
I used minidsp 2x4, behringer deq2496 all with iir on a diy fullrange driver and was very happy with it.
But wondered myself if fir could be even better. Never tested it as a fullrange has minimal phase behavior and EQed to the point made it phase linear, too.
FIR is known to have pre ringing in the time domain and some claim this is audible. As you made A/b comparison with fast switch I trust your hearing experience.
However a time corrected multi way speaker with FIR will profit.
My fullrange driver could withstand a simultaneous playing of original instruments against their recording being played on the boxes (no need to switch between sources).
After that I knew: loudspeakers can sound like the original if really adjusted to the point - measurement assisted.
But you have to invest some time for doing dsp programming work - quite similar to making a passive crossover.
Real professionals for making passive crossovers can make fast developments.