Digital crossover instead of conventional crossover

[TEK]

Hi

First: By conventional crossover I mean for example a crossover as the one Jon is building for the Wavecor Ardent. By digital crossover, I for example mean a crossover like minindsp (http://www.minidsp.com/applications/...ossover-basics)

Have anyone experienced with digital crossover? If you have, but is not using it anymore - why?
If you areusing it and would never go back, why?
I'm looking for proos and cons with the two methods?

[cjd]

I have not heard the miniDSP yet, but intend to snag one to use for initial testing and setup when I get around to the next steps on my dipole project. If I find it lacking I may implement an analog active crossover. I hear mostly good things with some reserved comments here and there.

I've heard two digital crossover implementations - one was very disappointing, with tons of low level detail lost. However, I'm not sure that was the fault of digital. The other was simply different, but was A/B'd with a passive. Again, I believe low level detail loss was the biggest notable difference, though it was tough to be sure. That was a significantly more expensive setup, but was still some time ago.

If you get into the real power of miniDSP with biquads and whatnot, they're incredibly flexible. Unquestionably useful for experimentation. As with any active setup, it means more amplifiers.

If you want an analog source in, you'll have to go ADA, but if you're looking at only digital source, run that straight to the miniDSP.

In the end, given what I've heard here and there, I'm a little unsure if I'll find the results satisfactory. Too many people rave about designs that I can't even listen to for a 5 minute demo.

I remember seeing Jon posting as if he were getting one, but I don't know if he ever USED it.

[Saurav]

I found some things easier to deal with using a digital XO. The two big ones for me were: not having to worry about the driver impedance interacting with the XO components, and not having to play tricks with slope and phase to account for acoustic center offsets (with a digital XO, you can add delay without affecting phase). I "time-aligned" my drivers by inverting the signal to one and adjusting delay until I got the deepest reverse null notch.

You'll need more amps, of course. You should run the digital stuff wide open, and use a multi-channel analog volume control downstream to minimize low-level signal loss, but as Chris said, it's probably never going to be as transparent as a good analog crossover. Good ones are hard to find, and expensive. You'll also have to worry about digital clipping, especially with a dipole or anything else that needs a lot of EQ.

[CraigJ]

TEK,

If it's good enough for Dr. Linkwitz, then it's good enough for me; http://www.linkwitzlab.com/LX521/DSP_challenge.htm IMHO, Davey's dsp design shown in the prior link is simple and complete. Plug this system into a 6 channel amp or 7.1 receiver, and presto! O.k., you need and mic and measuring software.......IMHO, the forum at Linkwitz.com has excellent info on digital crossovers.

edit, not to leave out; http://www.musicanddesign.com/NaOIIDigital.html

[TEK]

CraigJ: I did not understand Your comment about 7.1 receiver.

According to my understanding, if you want to use Davy's design for stereo 3ways you need to have 6 amplifier channels, 3 pr. speaker.
I guess a advantage of this type of design would be that you probably could have a very low watt amp for the tweeter, a bit more power for the mid-tone and most power for the bass units.
But for a, for example 7.1 setup, with 5 2-ways and 2 3-ways, you would need 16 amplifier channels. That's a lot :-)

Also, for a home theater setup, I also see that you would have to think about how you are going to get the digital channels pr. amp from the source and into the crossover. Not sure if there is chips or something that you could by to get that to work.
The easiest would probably be to let the digital crossover support analog input, but then you will get unnecessarily analog to digital to analog conversion, with probably would cause signal loss and not being a wanted solution.

[CraigJ]

Tek,

You can use one 7.1 receiver for a 3 way stereo. A lot of 7.1 receivers have the ability to use 6 of the internal amplifiers separately (6 inputs with matching 6 outputs to speakers). These 6 amps of the receiver will power a 3 way stereo and you will have volume control.

Home theater setups can be implemented with a blu-ray player with 7.1 audio out into the digital crossover, then amps, or receiver(s). On my HT system, my front speakers have digital crossovers and the surrounds and rears are analog. It may sound complicated, but the front channels come out the blu-ray player and into the input of the digital crossover, then into a receiver; done (I don't use a center). The second receiver powers the side and rears, the subs have their own amps. I'm sure there are different and better ways of implementing this.....but it works well for me.

And since I feel like I'm coming out of the closet ops:, my main system is a digital 4 way, using 10 internal amps from 2 receivers. Each speaker has 2 woofers, each with its own amp.

Hope I've simplified it some.

[JonMarsh]

What Craig has described is a cost effective way to put that kind of system together.

I do have one of the more capable miniDSP systems myself, have played with it a bit, but for me it would not be the thing to do to take the output from my 24/191 NOS R2R ladder DAC (TotalDAC-D1 Dual, full balanced DAC architecture, not just output stage).

The trick with any DSP approach is to have the right kind of filters available to implement the necessary transfer function, which is far from being text book slopes and pass bands with real world drivers and cabinet designs- best to investigate carefully, or pick a solution that gives you full control over where all the poles and zeros go. That, or develop a speaker design amenable to crossover approaches that are more straight forward.

If it's good enough for Dr. Linkwitz, then it's good enough for me; http://www.linkwitzlab.com/LX521/DSP_challenge.htm IMHO, Davey's dsp design shown in the prior link is simple and complete. Plug this system into a 6 channel amp or 7.1 receiver, and presto! O.k., you need and mic and measuring software.......IMHO, the forum at Linkwitz.com has excellent info on digital crossovers.

Dr. Linkwitz is a very smart guy, a very capable designer, yet there were things he was addressing in different aspects of the Orion 4 crossover design that surprised me in the sense that I would have thought they would have been sorted out earlier...

There's a saying about if you meet the Buddha on the road, kill him... it's a metaphor, not to kill the Buddha, but to realize that giving over your own judgement and knowledge development to another has it's risk- it's like killing yourself on a small level. I'm not keen on tons of analog opamps in my signal path, and I'm not keen on the digital I've heard to date in DSP crossovers, excepting my own experiments with using the LIO-8 DAC as a crossover with it's DSP functions- but then the LIO-8/ULN-8 has a very high reputation as a DAC and mixing control system, and is an exceptional value proposition in the $4k to $5K price range (LIO-8 is a ULN-8 without the full DSP license and without mic preamps). (my LIO-8 is upgraded to ULN-8 configuration).

I still have the Avo Picante project on hold, based on this concept; after connecting up with Mike I switched resources to the Isiris.

Basically, I'd suggest that if you want to do DSP, by all means, do DSP, and expect to experiment a bit and perhaps go through a couple of spins of equipment to find something you're really happy with, depending on expectations. Having had full active 3 and 4 way setups in the past, I also found the reliability of that kind of setup to be less than a more straight forward signal path- more parts, sometimes more problems.

Last, CJD's comments reflect my own experience with some other DSP setups I've heard. But don't take our word for it, and approach it in with a sense of adventure and train blazing- you may enjoy it quite a bit.

[cjd]

The question is whether I can get dipole correction to 16Hz passively. It actually may be possible, I'm just worried about insertion loss.

[Saurav]

All you need is a 10' wide baffle

[cjd]

All you need is a 10' wide baffle

It will be a W baffle, dual opposed-mounted Maelstrom-X 18" and I don't remember how to calculate the numbers on that. IIRC it only needs a few dB. On the other hand, I'm pretty sure I'd have to flatten the impedance peak at Fs, and that'd be some huuuuuuge components!

[JonMarsh]

Biamping a dipole speaker system makes sense, if you don't want to go full active or full passive. LF with EQ and the upper stuff passive, I think- that's the plan of record for the Arvo Picante. Hope I get back to that some day!

[Saurav]

dual opposed-mounted Maelstrom-X 18"

When will you get out of your rut of building cute little speakers and tackle something more... manly?

FWIW, even with my OB12s, I found the room filled in quite a bit of the predicted dipole rolloff. Also, since I was handing off to a sub, I didn't need the full shelving EQ only to have the high-pass filter fighting it, I was able to get the slope I wanted with much gentler EQ.

[cjd]

When will you get out of your rut of building cute little speakers and tackle something more... manly?

FWIW, even with my OB12s, I found the room filled in quite a bit of the predicted dipole rolloff. Also, since I was handing off to a sub, I didn't need the full shelving EQ only to have the high-pass filter fighting it, I was able to get the slope I wanted with much gentler EQ.

You are in luck. That's a stereo pair totaling 4 of the Maelstrom. Room gain may well negate the need for eq. I'll try it for sure! The woofer is a single 15".

[sfdoddsy]

After many years of using digital crossovers for my Linkwitz derived speakers (Driverack PA, Behringer 2496, MiniDSP) I am pleased SL has granted them his seal of approval. Especially after being told by some during those years that digital crossovers could never match what an ASP could.

Assuming (as I do) that sound quality is at least equivalent to an analog crossover, the big advantage is flexibility. If you want to change or try something it is easy. If you want to do this with an analog crossover be prepared for fun and games with caps, soldering, parts, bad connections etc etc.

If you're comparing a digital crossover to a passive crossover, I'd say the same applies. Active crossovers have theoretical advantages over passive, but it all comes down to design.

Having said all that I would prefer a passover crossover for my speakers rather than active simply for practicality. I'm sick of cables, crossover, more cables, amp, more cables blah blah blah.

[JonMarsh]

Having said all that I would prefer a passover crossover for my speakers rather than active simply for practicality. I'm sick of cables, crossover, more cables, amp, more cables blah blah blah.

This part is very familiar with me... I just did the active stuff a little earlier than some of you guys (70's and 80's). Designed and built some very high performance electronic crossovers in that time (high slew rate, low distortion and noise, etc).

[MickRS]

Just discussed the DSP issue with a friend. And although it sure has some pros, sometimes it looks like cheating a bit to me. In a sense of: no matter how bad your speakers are, with dsp you can correct it. Its like photography: with photoshop you can make a pretty good picture of a bad shot...

Now I'm a bit of a less is more kind of guy in hifi, having a se tube amp and fr speakers, so it will probably take some time to get used to this.... But this weekend I listened to a very expensive set up with dsp and had the possibility to a/b the set. And there was no denying that with dsp on there was LF certainly were a lot better.

[JonMarsh]

It can be very useful for equaling LF performance at one listening position... as long as you understand that limitation! IMO, it's best to position the speakers correctly for room response in both the bottom end and imaging (eliminating comb filtering by preventing early high level reflections with short delay), as DSP can't touch that.