TOPIC:

My setup so far is analog, so I guess I'll use the XLR in/out optional board unless there is a compelling reason to go some other route (Why are you going with the curry man DACs BTW? is there a matching ADC available?)

Also I should've asked where do I go to learn how to use this stuff.
I assume I'll need the Dayton omnimic v2 to measure, REW to fart around with room problems and your files & minidsp plug in to stuff all this data into the minisharc...am I missing anything?

If you strictly implement the parameters listed in the document you will be in the exact same situation as a "normal" M2 user with the crown amp, or any good loudspeaker: the speaker will have a flat anechoic response, but it will still benefit from specific EQ in the LF (say <200Hz).
You might also want to gently alter the HF response depending on your taste and situation, but this can be done by ear.

So, you will need a mic and measurement software if you want to adapt the speaker to your room, but that is the same as with any other speaker.

I'm also starting on a DIY M2 project. I already have a miniDSP 2x4. Is the 2x4 unit capable of handling the basic crossover filters? I assume the openDRC has better capabilities for implementing room correction and is a superior unit. I was mainly curious if the 2x4 was worth using as a starter or entry level system....or is it just a waste of time given the potential of the M2 components.

Thanks!

Don

@ Don,

I doubt that you'd be able to do what pos built for the JBL M2. The 2x4 uses only IIR filters while the M2 design was heavily dependent on the use of FIR thanks to rephase. Not apples for apples I'm afraid..

DevTeam

In fact the M2 filter and EQ settings as implemented in the Crown and BSS unit only use minimum-phase IIR corrections.
They did not use the FIR capabilities of these units, probably because these are limited to textbook filters which do not take driver responses into account...
No such thing with the openDRC of course, so I plan to implement a fully phase-linear version as soon as I can measure the real thing. As it is now, and based on published measurements, it looks like a simple LR36 linearisation at 800Hz would go a loooong way toward achiving linear phase acoustical reponse in the passband (ie ignoring the BR rolloff down low).

So as it is now (ie factory IIR settings) implementing these correction is just a matter of having the right EQ and shelving implementations (but some equivalent are given in the google doc linked in the first post), enough EQ points, and a good biquad implementation (ie avoiding too much quantization noises for LF EQs, and not loosing to much accuracy for UHF EQs).
I think the miniDSP 2x4 would qualify for the first two points, not sure about the last point tho.

Here is the expert the rescue!
I wasn't aware of this and thanks for the clarifications. I assumed rephase was behind it..

@ Don, appologies for the confusion.

DevTeam

rePhase is indeed behind it and you can find rephase presets in the google document linked in the first post.
But the rePhase correction is only minimum-phase for now, replicating the original M2 settings.
I plan to add phase linearisation as soon as I get a chance to precisely measure its behavior.

To me the openDRC (one per loudspeaker) is the perfect platform for applying these settings in pure FIR, be it minimum-phase or not.

Thanks for the input guys! I'll start with the 2x4 since I have it and will most likely upgrade to the openDRC package in the future. Right now I'm waiting on parts and have to build cabinets so I am many weeks away from making noise and many months from completion.

Cabinets nearing completion. Decided to skip the 2x4 experiment so a pair of openDRC-AN are in transit. I should be back in about a week...probably with questions on going from rePhase to the plug-in. Time to get get back to the cabs...

EDIT: double message...

If you are planning on building M2 clones then beware of the passive crossover published in the official M2 spec sheet.
Turned out that the C2 capacitor value was misprinted 33.3uF instead of 3.3uf : www.audioheritage.org/vbulletin/showthre...iewfull=1#post383264

Here is the correct schematic :
docs.google.com/document/d/1bzJyui75ZG_-...ading=h.267sxg2itu1h

As for the openDRC presets, I had them published on the first post, but they were subsequently removed from by dropbox.
I will try to rebuild them and publish them again.
This is pretty simple stuff: everything is handled in the FIR, even the relative delay and levels, so all you have to do is generate bin files based on the rephase presets published in the document (load from paste and press generate ), load them in a fresh openDRC settings, and adjust routing...

Would a single openDRC DA8 be up to the task of running a stereo set of M2s?
If it can do that and still have processing power left over to EQ four subs, I'm getting one.

It probably can, but you would have to mix FIR and IIR EQs to get the job done, because it won't have enough taps.
That means extracting the lowest frequency EQs from the FIR and turning them into IIR (beware of the constant Q vs proportional Q thingy...), one after the other, until the result curves matches the target one close enough in rephase for the number of taps you can afford...

About mixing IIR and FIR I just want to make a comment... I do this for a stereo 4-way system (low channels IIR and hi channels FIR, 80/500/5k, all RL48dB/8) and it make a great difference compared to the same system using all around IIR's.

I could even go lower FIR by mixing another set of channels (midbass) 500Hz FIR-LPF with 80Hz IIR-HPF , but this would mess things out because of about 10ms time difference between low cut and high cut of the same band-pass (composed) filter.

I add this observation because I have the feeling that you've just recommended such a configuration (mixing IIR and FIR in the same channel), which I think is a mistake as the actual software do not let you delay just HPF cut (IIR) to be in line with LPF cut (FIR). You can delay a channel as an all, not part of it. Sorry if it's a misunderstanding from my part

Not sure what you did exactly, but by "mix" I meant chaining IIR and FIR on the same channel, not mixing two channels together.