This is not exactly a DIY question, but since this forum is frequented by people who have knowledge and experience in speaker design and measurements I would like to ask for some opinions on the following speaker designs.
I will certainly make effort to hear speakers mentioned in my questions.
I am interested to hear about experience with and opinions about 3-way speakers employing digital (DSP) active cross-over (examples being Kii Three and D&D8c) vs. 3-way speakers employing analog active cross-over (examples being PSI Audio A23-M, ATCSCM50ASL).
If you look at those examples, PSI Audio A23-M is using propitiatory tweeter and midrange drives developed in house, with propitiatory amplification using unique technology such as Adaptive Output Impedance error correction feedback system which according to them allows for superb control of excursion. They also do phase compensation what is evident in impulse response and transients reproduction. All of this done in analog domain, what according to them is the proper way, not employing DSP. This certainly sounds like proper approach.
On the other hand, Kii Three does active cross-over in digital domain employing DSP. They also do phase compensation as well as cardioid dispersion pattern in low-mids and bass which is certainly beneficial, and their DSP processing seems to be very well done, what is all great, but on the other hand they use off-the-shelf drivers and amplification.
My thinking is that the two examples of DSP driven speakers would certainly be the ultimate examples of what could be achieved in quality sound reproduction if they employed higher quality electronics, for example not little switching supplies that feed everything, from DSP processor, to DA converters and amplifiers. Does this make sense, do you think that the design and objectives were exemplary but the execution is not that great due to many compromises to keep the cost down?
And it is probably not only the examples of DSP driven speakers, PSI Audio example probably has bunch or relatively noisy op-amps in its active analog cross-over powered by switching PSU that introduce lots of issues and they had to employ many to align phase.
So I am interested to hear how these compare based on actual experience and/or in lack of the actual experience opinions based on some sound engineering knowledge, and which approach (DSP or analog active) would degrade the signal less.
I will certainly make effort to hear speakers mentioned in my questions.
I am interested to hear about experience with and opinions about 3-way speakers employing digital (DSP) active cross-over (examples being Kii Three and D&D8c) vs. 3-way speakers employing analog active cross-over (examples being PSI Audio A23-M, ATCSCM50ASL).
If you look at those examples, PSI Audio A23-M is using propitiatory tweeter and midrange drives developed in house, with propitiatory amplification using unique technology such as Adaptive Output Impedance error correction feedback system which according to them allows for superb control of excursion. They also do phase compensation what is evident in impulse response and transients reproduction. All of this done in analog domain, what according to them is the proper way, not employing DSP. This certainly sounds like proper approach.
On the other hand, Kii Three does active cross-over in digital domain employing DSP. They also do phase compensation as well as cardioid dispersion pattern in low-mids and bass which is certainly beneficial, and their DSP processing seems to be very well done, what is all great, but on the other hand they use off-the-shelf drivers and amplification.
My thinking is that the two examples of DSP driven speakers would certainly be the ultimate examples of what could be achieved in quality sound reproduction if they employed higher quality electronics, for example not little switching supplies that feed everything, from DSP processor, to DA converters and amplifiers. Does this make sense, do you think that the design and objectives were exemplary but the execution is not that great due to many compromises to keep the cost down?
And it is probably not only the examples of DSP driven speakers, PSI Audio example probably has bunch or relatively noisy op-amps in its active analog cross-over powered by switching PSU that introduce lots of issues and they had to employ many to align phase.
So I am interested to hear how these compare based on actual experience and/or in lack of the actual experience opinions based on some sound engineering knowledge, and which approach (DSP or analog active) would degrade the signal less.
Roughly.... active systems can be more precise in the finetuning, you get direct control - no passive components in between drivers and amplifiers, amplifiers can be choosen for their specific working range and putting in numbers in a DSP is way more easy, then fiddleling with tons of different passive components, to achieve the right value.
But implementation is key. No matter what you do, you need to makes compromises. So plan your build and find the compromises you are most willing to make
But implementation is key. No matter what you do, you need to makes compromises. So plan your build and find the compromises you are most willing to make
I can't speak for PS Audio, but I know that the Kii does not use off the shelf drivers - they specify their own drivers, and also specify their own amps and power supplies.
I think the it is incorrect to say that 'they use tiny SMPS etc' therefor not optimum. You need to look at the final result. If changing to a big heavy supply wouldn't make an audible difference, why waste money and space?
You really think that the designers of such a fine product would 'forget' to check if the psu can be easily upgraded for even better sound?
In products like this, everything is optimized within the design and cost constraints. You can't simply improve the design by a single design point, but would have to improve a lot of interrelated factors and design choices. And thus end up with a design in a much different price bracket.
Jan
I think the it is incorrect to say that 'they use tiny SMPS etc' therefor not optimum. You need to look at the final result. If changing to a big heavy supply wouldn't make an audible difference, why waste money and space?
You really think that the designers of such a fine product would 'forget' to check if the psu can be easily upgraded for even better sound?
In products like this, everything is optimized within the design and cost constraints. You can't simply improve the design by a single design point, but would have to improve a lot of interrelated factors and design choices. And thus end up with a design in a much different price bracket.
Jan
In my opinion, there should be no difference in sound quality between a well made analog crossover and a well made DSP crossover. Of course there are shortcuts and pitfalls that may afflict those poorly made versions of either type.
I suspect that with current technology, the DSP route is a less expensive way to get a high quality crossover. I suspect that the analog crossover is going to have more components and a higher manufacturing cost.
As an example, the MiniDSP HD 2x4 is 2 input, 4 output DSP device with very good sound quality. In addition to a highly configurable crossover, it can introduce time delay and perform parametric EQ. It retails for ~ $210. It would be hard to make and sell an equivalent analog device for $210.
I suspect that with current technology, the DSP route is a less expensive way to get a high quality crossover. I suspect that the analog crossover is going to have more components and a higher manufacturing cost.
As an example, the MiniDSP HD 2x4 is 2 input, 4 output DSP device with very good sound quality. In addition to a highly configurable crossover, it can introduce time delay and perform parametric EQ. It retails for ~ $210. It would be hard to make and sell an equivalent analog device for $210.
Last edited:
I use an Ashly analog crossover at 350hz with excellent results.
I'd like to find a preamp that took the digital output from the CD player and processed it in the digital domain which includes a two or three way crossover. DACs would be at the outputs of the crossover and full range. A little bit of gain would be nice too, maybe 8-10db.
Anyone know of such a device?
I'd like to find a preamp that took the digital output from the CD player and processed it in the digital domain which includes a two or three way crossover. DACs would be at the outputs of the crossover and full range. A little bit of gain would be nice too, maybe 8-10db.
Anyone know of such a device?
Don't ask me - but they are not off the shelf.
The other things is that yes, both analog and DSP can provide good audio quality. If you mean freq response, filter characteristics, distortion, that kind of thing.
But not as far as acoustics is concerned. For instance, it is a no-brainer to separately adjust freq response and phase response with a DSP filter, something that is simply not possible with an analog filter.
Jan
Last edited:
According to PSI Audio and third party measurements their analog active cross over does what you say is impossible:
it’s phase compensated from 200Hz upwards so that the various time‑domain errors introduced by the drivers and filters are equalized.
So... they modify it.... propriotary units.... that just look very much like the standard one
It can be done in analog, but it has many constrains. In digital, it can be done in all situations.
It is not the same as regulating the total phase... and obtaining a perfect impulse response.
Could you please let me know what the constrains are?
What is their significance on the end result?
My humble opinion: PSI and many others are just old school. For over a decade now developments in directivity control have set a new benchmark in sound quality in acoustically not perfect spaces. Like our living rooms. We have to thank a lot of pioneers like Geddes, Linkwitz and Toole (and many others) for figuring out the relevance of it. Something like addressing the weakest chain. And thx to digital filtering (build an analog delay anyone?) practical solutions became possible. Listen and enjoy.
Sorry Sonce, but you are factually incorrect on this. It is possible to implement almost any filter in analog. Likewise you can make phase adjustment and delay via single or cascades of all pass filters (or in extreme cases, bucket brigade circuits). You can even obtain excellent step response. While all this is possible, it's a pain in the neck. Moving to the digital domain allows you to make any filter, delay, phase adjustment, etc. without having to alter any hardware.
It is possible that an all analog filter may be cheaper if used in a system that is known and will never change. But DSP hardware has become pretty cheap (particularly compared to the cost of drivers) so this is no longer a consideration in anything but very high volume designs (think low cost active monitors like a JBL LSR305).
As mentioned before, you can make a good or bad implementation of both analog or digital systems. Nowadays I think it's better to just go digital as you can justify spending lots of R&D money to di it right the first time as most future improvements happen in software. With analog you are constantly redesigning.
As far as "wimpy" SMPS (from another post on this thread) - have a look at the charger used in an electric vehicle. Smaller than a shoe-box and capable of controlling more than 100kWh!!!! A properly designed SMPS is not noisier than a linear supply (with some very esoteric exceptions - you can get super quiet linear designs for very low power, for example). In both cases performance depends on proper engineering.
It is possible that an all analog filter may be cheaper if used in a system that is known and will never change. But DSP hardware has become pretty cheap (particularly compared to the cost of drivers) so this is no longer a consideration in anything but very high volume designs (think low cost active monitors like a JBL LSR305).
As mentioned before, you can make a good or bad implementation of both analog or digital systems. Nowadays I think it's better to just go digital as you can justify spending lots of R&D money to di it right the first time as most future improvements happen in software. With analog you are constantly redesigning.
As far as "wimpy" SMPS (from another post on this thread) - have a look at the charger used in an electric vehicle. Smaller than a shoe-box and capable of controlling more than 100kWh!!!! A properly designed SMPS is not noisier than a linear supply (with some very esoteric exceptions - you can get super quiet linear designs for very low power, for example). In both cases performance depends on proper engineering.
Yes, yes and yes.
No.
Excellent? No, "excellent" is not the same as the perfect triangle of step response ("transient perfect") and perfectly flat frequency response found in the symmetrical first order crossover. And that is - you are stuck with the first order Butterworth only. All pass filter can help when tweeter and woofer are on the same vertical baffle.
The Harsch crossover (and some others) has "excellent", but not perfect step response. Also, Harsch has not perfectly flat frequency response. And you are stuck with some specific filter types and slopes.
So, no - there is not possible to have perfectly flat frequency response AND perfect step response in analog crossover, except in the first order symmetrical crossover.
Only in digital domain you are free to chose filter types and slopes AND have a perfect step response.
It seems the majority considers the new DSP driven cross-overs much better approach than active analog counterparts.
When you look at directivity and impulse response on Kii Three it is certainly incredible.
But here is my dilemma
Active large 3-way ATC has arguably exceptional drivers, their A/B amplification is also sounding and measuring well. Let’s say on one end you have such ATC system fronted by a high quality DAC.
On the other end you have Kii with drivers that are certainly not the same quality, DACs that are not top of the line, amplification that is arguably not up to the level found in ATC.
ATC will certainly not have so nicely controlled directivity as Kii, and ATC’s impulse response will be terrible in comparison.
But will not ATC ultimately sound better in terms of resolution, dynamics when placed in a decent room considering its higher quality drivers and amplification, and better external DAC?
Will not the subpar electronics in Kii make it sound very precise but ultimately kill transparency?
When you look at directivity and impulse response on Kii Three it is certainly incredible.
But here is my dilemma
Active large 3-way ATC has arguably exceptional drivers, their A/B amplification is also sounding and measuring well. Let’s say on one end you have such ATC system fronted by a high quality DAC.
On the other end you have Kii with drivers that are certainly not the same quality, DACs that are not top of the line, amplification that is arguably not up to the level found in ATC.
ATC will certainly not have so nicely controlled directivity as Kii, and ATC’s impulse response will be terrible in comparison.
But will not ATC ultimately sound better in terms of resolution, dynamics when placed in a decent room considering its higher quality drivers and amplification, and better external DAC?
Will not the subpar electronics in Kii make it sound very precise but ultimately kill transparency?
Again.... implementation is so much more important than some specifications in certain components.
Yes you will hear maybe small changes between good and bad electronics. But mainly it's how it's used. Ofcourse with maybe high noise in the background and less good resolution.... then in direct comparison, you will hear difference.
But I tried to listen between a Behringer EP4000 and huge mono blocks from Dan D'agostino - on a set of Magico Q7. There was a little difference, even though we used same extremely expensive DAC, cables, speakers, DSP controlled subwoofer and everything. The Behringer still played loud and clean.
So it could very well be, that the Kii 3 will play much better than alot of systems, simply because they did it better with what they have.
And I agree... it's aparently cheap:
Kii Three — ImgBB
SBS-160F35AL01-04 Peerless by Tymphany | Audio Products | DigiKey
FSL-0512R01-08M Peerless by Tymphany | Audio Products | DigiKey
SEAS Prestige 27TBCD/GB-DXT (H1499) Tweeter
But hey... if it works.... then they just made the best out of it all... by simply using the right components that seems to do the job just fine
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.