So, I installed the speakers and did some tests !



DECAYS



First thing, the decays :





And the one I especially like, 63 Hz band vs full range ETC ! note that this is with a poorly optimized desk which is blocking a bit the floor bounce, hence it's absence in the ETC.



Next step was to test the door with treatment or not. Bad news is, there is some slight feedback showing every 30 ms on the 1K, 2K and 4K filtered ETC.
Good news is I tried with 40 mm and 200 mm of porous, and it is very very close. So It seems I will be good with a thin porous absorber on the door, don't have graphs for those as I did not save them apparently.

As you can see also in the ETC, there is some clutter in the first 5 ms at -20 which are desk and uncovered beams in the front part of the ceiling. Also the trimming around the speaker is missing for now so it's giving some diffraction. All that should be corrected when everything will be finished.

I have some parasites also in some measurements, as the seals of the interior doors were bad quality, so I removed them, and I removed the door closer to paint them. Long story short, they are freely vibrating in their frame with loud measurement, and there is some leaking from the back of the casemate also. Still some painting to do on the framing, I'll put new seals when it's done.



PMC COMPENSATION



Now, we enter the harder part, the compensation of the transmission line when mounted in wall. At first I thought I would just need to put a "reverse" baffle step filter, but there is more to do with these speakers. Many thanks to Thomas @ Northward to point that, because I would not have thought about that when looking at my first measurements. I measured the speaker in close miking, and at the listening position, and did that also with a big wood board compressed against the ATL terminus to effectively block it. Results ? loosing some low-end (getting even a big anemic) but getting way smoother, lower group delay, and funny thing, the 20-30 Hz longer decay you see in the spectrogram, 80% gone. At some point, in NE rooms, you begin to see the inherent decay of the speaker.

So as blocking the TL would not be practical, I thinked about the stuffing of the transmission line.
- not much stuffing : resonant, low mid getting out of the terminus, ragged response.
- Stuffed to the max : it's getting close to a leaky closed box.
- A lot of stuffing but less : you get an aperiodic alignement, which is close to a closed box in terms of group delay, with still a bit of support in the deep lows. That was what I was shooting for.

So, opening the speakers, and trying different stuffing ! I'm always amazed at the magnet size of the ATC SM75-150S, it's similar in weight to the 12" Volt woofer ! I opted for left overs of Cotton waste felt I had.



With a lot of trial and measurements, here are the results. Notice the scale : 2 dB.
- Green curve : speaker stock, with a BIG bump 6 dB bump between 50 and 100 Hz.
- Purple curve : speaker with stuffing, getting way more closed box like in the response. I tried so many things, more stuffed than than and it was getting a bit "forced", that compromise is the best I found between tightness and effortlessness. As an added bonus, the low mid range cleared up, and the group delay lowered by 30 % at 30 Hz.
- Red curve : as it's not the intended alignment of that box, I applied a small corrective linkwitz transform to flatten the low-end, as it's usually done in closed box speakers. It not that big and more importantly, it's a smooth EQ curve with around +3dB on deep low-end and -2.5 dB in the upper bass.





FR



Which get us to the measured FR at the listening position, with desk, all racks, uncovered beams and missing diffraction trim around the speaker. You can see the desk related things in the mids, but overall, I'm close to the finish line !



And more importantly, subjectively, it's sounding awesome. The low-end curve I settled with (around +5 dB in the lows) seems to work great. I'm around the projected sweet spot (actually 15 cm in front of it), and at the listening position and the sealed with plaster concrete front wall, it's surprisingly live with self noises.



FINAL THOUGHTS



I've still some work to do to finish all the details, building a better desk, and all the cosmetic stuffs. The room is workable from now, better sounding than my now previous room, and after 13 months of intense labor and 5 years of planning, thinking, learning, I'm quite happy of my 100% DIY results ! And quite frankly, I loosed a lot of pressure ! Just before listening to the first test track, my heart was beating above 200 bpm !