Amp upgrade for JBL M2 (would be submitted for testing)

[grizaudio]

Check out Powersoft Quatrocanali 2404 and Armonia + DSP
Armonia is a free DSP down load from Powersoft . You can also use Smaart V8 to hone frequency response in the DSP.

I actually used the Powersoft Quattrocanali 2404 DSP in my active setup for a while.

Pros:
- Dynamics
- Dante
- Bass weight
- Extensive on board IIR DSP

Cons:
- Sound quality, not on par with Hypex/Purifi sound quality
- Limited FIR TAPs for convolution DSP crossovers and phase correction
- My Purifi amplifiers sound much better, using software convolution engine

I settled on using Roon's convolution FIR engine to implement POS convolution active crossovers.
I also applied an additional 2 FIR additional EQ points, providing a 2db gentle roll off.
Low shelf 90, 1db, Q0.6
High shelf 500, -1db, Q0.6

Signal path is:
Intel NUC S4 running Roon Rock > (USB) > Raspberry Pi4b >(USB) > Motu Mk5 Ultralite 10 channel DAC

3x NAD c298 amplifiers operate in low gain, 8.5db
Amp 1, Runs in stereo feeding JBL D2430K L&R Compression drivers
Amp 2, Bridged output to first JBL 2216ND woofer
Amp 3, Bridged output to second JBL 2216ND woofer

I highly recommend this setup for the JBL M2's.

 

[grizaudio]

When I measured my Crown cdi 4/1200 drivecore (original M2 settings in the FIR section of the dsp) and compared it to the manual settings found on internet I found that the shelving in the FIR setting for the woofer was +8 dB low shelf centered at 135 hz and not -8dB high shelf as shown. The frequency response remained the same for the woofer isolated, but I got far too low output relative to the tweeter when I put in -8dB high shelf as told by the internet pictures. I installed the rephase settings into minidsp 2x4 hd and found the same issue, the settings there equaled a -8 db high shelf leading to way too low woofer output relative to the tweeter.
I can't explain this findings, but somewhere I should have the measurements to confirm if I ever went to trial for my statements.

I'm curious, how can this be the case, we can see below from a previous post that @pos settings measure the same as the OEM filters? and match the BSS settings?
+8db, provides a very different response in rephase.... and one which doesn't allow for flat correction to 30hz?

I have personally implemented the rephase setting as convolution in Roon. Sounds great to me.
I added a -1db, 0.6Q shelf from 500hz, and +2db, 0.6Q of shelving below 90hz. This sounds about right to me, in room.

View attachment 193851

JBL M2 Reference Master Monitor Review

I haven't looked at the perceived similarities but if it were a measurement rig issue then every other speaker would show the same behavior. Aside from that, this speaker was measured on its side. The HTM-12 was measured vertically. But, to top it off, I do sanity checks with a groundplane...

www.audiosciencereview.com

 

[grizaudio]

Is the dip around 475Hz on purpose?

Are you referring to the measured OEM JBL response curve? By Erin above and here…. https://www.audiosciencereview.com/...jbl-m2-reference-master-monitor-review.26051/

 

[grizaudio]

The one in your post, at least. I'm not sure of its provenance.

Read the speaker review. All is answered.

 

[DubbyMcDubs]

The M2 rephase settings as found in this document should match the measured outputs of a BSS with factory M2 settings within a fraction of a dB.

You have done a wonderful job putting that together!

In your opinion, is the MiniDSP Flex a good option for using this approach with your rephase settings? Or is there a better option? I see that the Flex only has 4096 taps, where your document recommends 6144.

 

[pos]

Thanks for the kind words @DubbyMcDubs
That amount of taps at 96kHz will not cut it unfortunately, so you would have to use the IIR/biquad section.
The Flex 8 does have a single stereo FIR section in front of the crossovers (contrary to the Flex 4), so that could be used to linearize the phase of the crossover (which could also have been done in the rephase/openDRC settings, but was not), and would also so leave plenty of channels for subs.

I do have a Flex 8 here to test, so I could definitely adapt the M2 settings for both the Flex 4 and Flex 8.

 

[DubbyMcDubs]

Thanks for the kind words @DubbyMcDubs
That amount of taps at 96kHz will not cut it unfortunately, so you would have to use the IIR/biquad section.
The Flex 8 does have a single stereo FIR section in front of the crossovers (contrary to the Flex 4), so that could be used to linearize the phase of the crossover (which could also have been done in the rephase/openDRC settings, but was not), and would also so leave plenty of channels for subs.

I do have a Flex 8 here to test, so I could definitely adapt the M2 settings for both the Flex 4 and Flex 8.

I have ceased looking into the M2 at this stage. Appreciate you offering to modify the settings!