Hi all, updates are here!
REW measurements after rear wall hangers/superchunks:
Low-end Frequency Response:
After hangers 26Hz - 340Hz SPL.png
I'm extremely happy with frequency response below 150Hz. It's currently +/- 3.2dB with no smoothing! (Until the speaker response starts to roll off below about 33Hz), this reduces to 2dB with 1/3 smoothing.
Waterfall: (Over 500ms to compare with the last measurements)
After Hangers 10Hz - 2050Hz Waterfall 500ms.png
Now again but reduced to 300ms to see more clearly specific issues still existing:
After Hangers 10Hz - 2050Hz Waterfall 300ms.png
I good to see the peaks and troughs smoothing out.
Low frequency resonant issues have now been reduced almost entirely to the axial length mode at 30-33Hz (theoretically 32.85Hz) in the room, and less so it's first harmonic at 62-63Hz (theoretically 65.71Hz).
I can now turn my attention to the mid frequencies and observe the resonant issues here.
The resonant issues around 190-200Hz appear to be a combination of the 3 close together axial modes: 4th width, 6th length and 3rd height harmonics.
My theory as regards to why the first natural modes for width and height don't appear on the waterfall is the deep slat helmholtz resonators.
I designed them using John's newer method, with construction timber varying from 150mm deep to 200mm deep and with an air gap ranging from 50mm to 400mm and an air gap of 5mm. This was to gain absorption targeted at 40Hz to 150Hz.
The decay rate for the first width and height modes was significantly reduced even before the rear wall treatments were added. Then the rear wall treatments cleaned it up even more.
Well that's the theory I'm working with at the moment!
The lower level resonances at around 137-150Hz line up with the height mode's second harmonic (136.65Hz) and width mode's third harmonic (144.12Hz) and the resonance at 165Hz matches up with the 5th length mode (164.27Hz).
If we look up the spectrum we see the repeating pattern, as multiples of these figures:
284Hz is double our 140Hz cluster.
380Hz is double our 190Hz cluster.
The 475Hz one is double the 7th axial length mode.
567Hz is double the 284Hz one.
and so on...
Plan of action:
- For testing treating width and height modes I intend to place a stack of insulation slabs, in the center of the long walls on the floor. This should reduce the energy of the higher order modes significantly.
- This should highlight the strongest modes causing issues.
You may notice the
huge null at around 650Hz. This is highlighted even more drastically in the frequency response graph:
After hangers 26Hz - 1000Hz SPL.png
This is almost certainly caused by SBIR. Annoyingly (as you will see later) it is
not caused by a ceiling or floor bounce!
Which leads me to believe the front wall can be the only culprit. I was hoping my baffles, not extending completely to the front wall would give sufficient baffle extension to avoid the null, but apparently it is a fearsome beast!
A 650Hz null would be cause by the speaker being 135mm or so form the front wall, which is a similar distance that my baffles stop from the front wall.
I will test this by placing a slab of insulation across the front wall and seeing if the null is reduced.
If it is (as I theorise), then I will have to decide between keeping the slab there, or adding a reflective surface between the two baffles.
Spectogram:
After hangers LR 20Hz - 5000Hz Spectogram -500ms to 500ms new.png
The spectogram confirms the 650Hz null, as well as (I believe) modal reinforcement to the problem modes we saw in the waterfall.
Decay rate:
After hangers LR 50Hz - 10kHz Decay Rate T30 0ms to 900ms.png
The decay rate shows the mid and high frequencies coming down to meet the low-frequencies which is great. 250Hz has gone too low though. My rear wall diffusers etc, should increase the decay rate by reflecting some of this energy back into the room though.
I'm not going to focus on 250Hz, but revisit it after the rear wall diffuser.
Now I've completed the initial RFZ treatments Let's look at the ETC:
After Hangers ETC.png
We can see some reflections here above -20dBFS before 20ms. I have confirmed 2 and identified 1 more.
Confirmed:
3.78ms: Floor reflection
- Confirmed by laying insualtion batts on the floor between mic and speakers
2.73ms: Ceiling reflection
- This was surprising, as I had assumed the ceiling reflectors would be enough to prevent this. I isolated the problem area to the joist between bay 2 and 3 from the front wall, just before the first reflector row. Direct reflected sound on the face of the joist (only 70mm!) and probably some diffracted sound wrapping around the joist seems to be to blame.
- I confirmed this by a series of measurements hanging insulation batts below this location.
Unconfirmed:
0.98ms: I strongly believe this to be the cuplrit with my 650Hz front wall SBIR problem.
The other few reflections near to 6ms are quite low level, and possibly some residual energy from my angled baffle extension "wings" being a bit too shallow an angle. I identified 40-50 degree off-axis waves could be directed towards the mix position by the baffle.
Depending on the outcome of the measurements I was going to decide whether to add a "fill-in" angle over where the "wing" meets the speaker baffle extension. I'll decide on this later on.
As you can see from the graph there are no high-level returning reflections after they reach the rear-wall hangers. This is great for removing destructive reflections, but I believe I can achieve an improved listening experience by adding a definite ITDG termination at around 20ms, where additional diffuse energy is allowed to return to the mix position to be fused into the direct signal. My plan for a rear wall diffuser, should help to address this.
Now that was a marathon to type, so I'll add additional photos another day. I have a few showing the various locations of the ceiling treatments I tried in isolating the ceiling reflection, and REW measurements to back up each picture.
Dan
Stay up at night reading books on acoustics and studio design, learn Sketchup, bang your head against a wall, redesign your studio 15 times, curse the gods of HVAC silencers and door seals .... or hire a studio designer.
But it's not something I can add to my resonators now as they're already fitted. I'm thinking of either sanding and oiling them, or hiding them behind a cloth finish (after testing the acoustic effects). My speaker baffles themselves could be veneered, but I haven't done it before (so I'm scared!), or the other option is to paint them satin/matte black, which will look cool with the speakers in them (and a lot cheaper).
(If anyone has any ideas on this, please let me know! 
) 
) seems to be in order, to target the ceiling>floor mode! I'll revisit it after the I've added the rear wall treatments. 
