Originally Posted by corradizo
Quickly took some measurements tonight. ...
The idea behind x.1 for bass is threefold:
- send the most power hungry program to a special driver(s) designed to handle the load
- separate the modal (bass) frequency generation from the specular frequency generation
- allow the program developer (who does the final x.1 mix) added capability for LFE - low freqeuncy effects.
That second one is what you're playing with here.
The audible spectrum within a room is divided into three frequency regions which are clearly evident on your unsmoothed sweeps.
- the modal region, where room boundaries define resonances, and drivers set up a sound field of frequencies with wavelengths comparable to room dimensions. Sound quality in this region is dependent on driver placement relative to room boundaries, not relative to the listener.
- the specular region, where room boundaries can be seen as mirrors, reflecting sonic energy in a manner consistent with Snell's law (angle of incidence = angle of reflection). Speaker placement is determined relative to the listener, with specific speaker locations (LCR and surrounds for 5.1) required for proper playback
- a transition region, inbetween.
Modal - looking at measurement #1
, note that below ~100Hz, your unsmoothed curves are nonetheless comparatively smooth. This smooth trend continues through the data; you're not seeing the strong peaks/dips characteristic of bass mode issues. There are resonances - tones that persist - but not cancellations, like those sharp dips seen higher in frequency. It looks like you've chanced upon a pretty good place to put those subs!
Specular - this is why we smooth FR data. The hash you see above ~400Hz is real, but ubiquitous; it's what we're used to hearing. That hash is due to wave interference; every sharp dip is a frequecy cancellation due to a second wave at the same frequency that's out of phase with the first. The key is that these cancellation modes get closer together as frequency increases, until they become inaudible. High modal density is characteristic of the specular region, as well as the ability to make predictions based on straight line propagation. Dolby speaker placement guidelines assume this behaviour. Note that speaker location relative to room boundaries can have effect, but you're in-wall so much of that goes away.
Transition - modal density is increasing, but modes are still audible. Speaker output interacts strongly with room boundaries, not to set up standing waves, but rather peaks and dips. Allison effect, floor bounce, rear wall dip all occur in this range. Here's a link
to DIY speakers, but I want you to look at the excerpts from Toole's book at the top.
- first pair show a freestanding speaker placed as designed, freestanding, and mounted in-wall, like yours. What's they don't tell you is that the big hump in the bass range is a design variable (baffle step compensation) that can be changed to make in-wall placement as flat as freestanding.
- the second pair show the effect of surface mounting vs. flush mounting a speaker on the wall. Note the dip at 200Hz due to an out-of-phase wall reflection, is in the transition region
- the third pair show the effect of bookcase mounting, with one possible remedy. Note the dips are again in the transition region.
So, let's see what you get with proper 5.1 bass management! Set everything to small and try again. Overall, this isn't that bad; the slight downward tilt at higher frequencies is a very common room curve target. The only thing I'd want to change so far is the 400-450Hz bump from L and R speakers, but not center. This could be boundary reinforcement - not all waves are out of phase - and it's consistent, appearing in app measurements with L or R.
I also did a spectrogram of #1
, and you will want to treat this room once it's fully furnished. I spoke earlier about strong resonances below 30Hz, but I also see a rise in decay time exiting the transition region. By ~400Hz, it's taking 0.6 sec for 40dB decay. You'd like half that. The trick is you're running close to 0.3 sec. in the transition region, 100-200Hz. You'd want bass-dead treatment to hit the 400Hz+ range, and bass-specific treatments below 100Hz, to avoid messing with the one frequency range that's already close to a reasonable target decay time.
Thankfully, there are ways to do that. The thickness of acoustic panels has a strong effect on LF absorption, as does the presence of air space between the panel and wall. Once the final requirements are known, you pick the thickenss and spacing that meets your goals. We can then discuss placement, to perhaps to address the 400Hz peaking if it's a boundary interaction.
Bass issues are not as easily addressed. Porous absorbers like those used in panels work better in areas of high air velocity (back to wave motion, pressure vs. velocity and their respective nodes). Bass modes have velocity nodes at the walls; the wave is all pressure, so wall-mounted bass traps need to be pressure sensitive, not velocity. I suspect you'lll find commercial product capable of absorbing below 50Hz, and there are several shown in Everest. you may not DIY, but you can always get a contractor to do the work, based on your plans.