[JohnR]

I think Jim just means that Meridian uses digital crossovers.

[Redefy Audio]

Yeah, it seems  ,

i always curious about meridian, they've been using dsp for a long time. Never heard of them, but many are impressed with their performance.

im looking for 3way active solution for my future modular speakers. i got 2x4 miniDSP, i dont have any complain about their performances. but yet, im going to use more expensive drivers than before, so i guess a step up is necessary.

MOTU 828mkII, Apogee Ensemble and Universal Audio Apollo is on top of my list. I can see that higher end interfaces will use Thunderbolt as it provide lots of flexibility.

How's ur experiment with miniDSP so far JohnR?

cheers
henry

[ccclapp]

Hi

Im looking to do the same (x/o, DRC in JRiver MC) but am looking at a DIY M-ch DAC via Buffalo-III with the exaU2I PC USB interface. 

Do any of you have any comments on that approach?

[*Scotty*]

Here is a link to some of the pitfalls one might encounter when implementing a digital crossover.
And some useful information regarding how to design a loudspeaker for a digital crossover.
http://grimmaudio.com/whitepapers/speakers.pdf
Scotty

[JohnR]

Strikes me as little more than a rant. What part specifically did you find interesting?

[*Scotty*]

What I found interesting were the well explained consequences of attempting to use DSP crossovers to perfect the performance an existing design. It appears that if you are going to build a loudspeaker that has flat response in both the vertical plane and the horizontal plane and exhibits phase coherence as well as transient perfect behavior in the time domain you have to first design the enclosure and the baffle layout with this goal in mind to the realize the potential inherent in the DSP based crossover.
 In other words the DSP based crossover is not a slam dunk easy to do right alternative to a conventional crossover. It will take as much work and measurement to get right as a conventional approach. However, when it is implemented correctly, it can deliver performance far in excess of what is possible with most conventionally designed crossovers.
Scotty
 

[JohnR]

The "good" and "bad" examples were done with the same MTM cabinet though, weren't they?

None of the examples there are transient-perfect - the paper is deceptive, only posting the impulse response of the "bad" example.

[*Scotty*]

I think the first thing to do about diffraction problems that are a result of baffle irregularities in the vicinity of the tweeter is to remember John Dunlavey's approach to solving this type of problem.
Felt and lots of it!
A similar baffle layout and use of felt would enable one to closely duplicate Dunlavey's results with a far lower parts count and probably less intermodulation distortion.
Scotty

[jimdgoulding]

I think Jim just means that Meridian uses digital crossovers.

Yah, I was referring to the subject title. The object, I believe, is to get the same results in the time domain w/o staggering the drivers ala Dunlavy, or more recently Acoustic Zen, for example.  Meridian makes active speakers as do several British and Scandanavian makers.

I use a small room friendly pair.  Formerly, I had crossoverless electrostats.  Conventional driver/crossover speakers usually manage to focus my attention on the treble range I'm so used to a time coherent, diffractionless sound.  I think that's the reason they do.

http://www.audiocircle.com/index.php?action=systems;area=browse;system=1448

[JohnR]

Hi

Im looking to do the same (x/o, DRC in JRiver MC) but am looking at a DIY M-ch DAC via Buffalo-III with the exaU2I PC USB interface. 

Do any of you have any comments on that approach?

No, but it looks awesome! (If it can be made to work of course.)

I've lucked out and been loaned a 6ch Firewire DAC (not new or anything but hopefully I can make good use of it). Well, one of those audio interfaces for recording... Since I'm mostly a Mac user I think I'll just try Pure Music's IIR filters to start with....

I'm still not clear on whether Firewire has the same issues that USB apparently suffers from. No matter, really.

[JohnR]

How's ur experiment with miniDSP so far JohnR?

Hi Henry, not much changed in the last six months or so. Been experimenting with some different driver configurations but nothing particularly interesting to report so far. As per this thread though, I'm interested in the idea of doing it on the computer - should be plenty of horsepower and resolution on tap there. In theory anyway...  Still feels like there is nothing "plug and play" at present, but worth having a go at I reckon.

[JohnR]

I've lucked out and been loaned a 6ch Firewire DAC (not new or anything but hopefully I can make good use of it). Well, one of those audio interfaces for recording... Since I'm mostly a Mac user I think I'll just try Pure Music's IIR filters to start with....

Well, a quick setup with one channel of a 2-way xover - it all seems to work. That was easier than i thought it was going to be... will try a full system setup in the next few days.

[randytsuch]

I am slowly progressing.  I started checking out Allocator, to use for the digital cross over function.
Right now, I am planning to use Jriver media center as the user interface.

Interesting JRiver thread here where about their convolution capability
http://yabb.jriver.com/interact/index.php?PHPSESSID=teoj0pilj3lve71q5015dgbon1&topic=68828.0

I was reading some stuff about Allocator, and they recommend measuring the drivers first, and developing the crossover based on the anochoic driver response.  Then, if you want, apply room correction before the crossover function.  You should measure room response as a 2nd measurement.

I also read somewhere to make the driver measurements outside if possible, to eliminate room response.

Still looking for the DAC, then I will be able to start measuring the drivers.  Allocator recommended the Speaker Workshop program as freeware to use to make the measurements.

Randy

[JohnR]

Good to hear Randy.. I'm proceeding somewhat similarly on a Mac path... although I'm not worried about FIR at this point, just want to get a prototype going.

WRT measurements, my understanding is that it's a fairly pragmatic tradeoff. Gated measurements will only go down so far in a room, because of reflections. Below that point, use in-room measurements and do what you can, or move outside. Fortunately, this boundary also roughly coincides with the Shroeder frequency.

[JohnR]

Well, so far I can report that computer-based crossover are... a mixed kettle of fish. I have one working now but I did think at one point I melted a woofer. Typical (likely) issues:

* Getting a measurement signal into the player
* Getting the right audio signal to go to the right place
* Odd noises from the audio interface when restarting/reconfiguring etc (some pretty loud - I thought for a few minutes that I had melted a woofer..)
* Background noise / SNR
* User interface issues - navigating between crossovers, eq plugins, etc
* Getting your analog sources to play (haven't figured that one out yet)

It's probably a good idea to figure this all out with only a cheap pair of sacrificial drivers connected.

Still, it sounds pretty decent now for a couple of hours messing around and zero hardware cost.

[mcgsxr]

Interesting reading as I am running a Mac mini based system at the moment.

I found a utility by Googling that silences the noises on wake up etc.  Helped a lot.

I am using Play on the Mini, and notice it has some interesting DSP capabilities - but, I am not using any at the moment.  It apparently can do high pass, low pass, custom bandpass etc, but likely only one at a time, so it is far from a complex digital crossover like you folks are playing around with.

[JohnR]

Hi Mark, I think the issues I'm having with noise and various burps and squeals may be largely to do with the audio interface hardware, it's a few years old now. I'd like to try something more recent but I'm not sure how (without buying it myself that is)

Still, I think this exercise is showing up limitations that are simply the nature of things - using desktop hardware/software to do embedded work is always going to be fraught with issues....

[JohnR]

I have added Pro-Q to the mix (have a license for it now).

I'm wondering if we should shunt the Mac stuff into a separate thread.... Randy, it's your thread, so just say what you think.

[randytsuch]

I have added Pro-Q to the mix (have a license for it now).

I'm wondering if we should shunt the Mac stuff into a separate thread.... Randy, it's your thread, so just say what you think.

I don't have any problem "sharing" the thread with you Mac people.  I would like to try a Mac, but I can't pry my kids laptops out of their hands for this, and I don't want to spend the money to buy a mini right not.

BTW, I think I remember reading about trying things out with a sacrificial driver first, so you don't blow up something expensive.  I was planning to use digital volume control, but wondering if I should do analog volume control, to take care of issues like this.

Randy

[JohnR]

You'd need four or six channels of analog volume, that's the catch.

I was pointed to this today, might be of interest: http://www.bodziosoftware.com.au/