Measuring the effect of acoustic treatments in the room

[AndyU]

Just can’t understand why you don’t use a membrane bass trap for the low frequency issues. A bit of fibre isn’t going to impede 41Hz, any more than pebbles on a beach will stop a big wave coming in.

 

[ToTo Man]

Porous absorbers need high air particle speed. The slowest speeds will typically be seen next to a boundary at its associated bass mode frequency. (In contrast that's the ideal place for a membrane absorber, which is a pressure absorber).

A porous absorber is better at absorbing non-mode frequencies at boundary (which you observe) and bass mode frequency away from its boundary.

(As you know ToTo Man, membrane is recommended for bass, this is another reason why.)

Just can’t understand why you don’t use a membrane bass trap for the low frequency issues. A bit of fibre isn’t going to impede 41Hz, any more than pebbles on a beach will stop a big wave coming in.

I have four membrane traps in the room tuned to 41Hz, but TBH I find them a nuisance. I bought them naively thinking they'd be sonically invisible outside of their bandpass but this is not the case, especially if you have a run of multiple units. They are, after all, solid boxes, so you are going to get frequencies reflecting and diffracting off them. Putting two units in each front corner, for example, generates a big, broad peak in the response at 115Hz. They are more benign on the rear wall, which is where they are at the moment.

Porous treatment makes more sense to me in smaller rooms because it works in all axial, tangential and oblique directions. Most membrane treatment is only designed to work in a single direction. Also, porous treatment also works on the x2, harmonics, membrane treatment doesn't. Thus, if I replace the porous corner treatment with 41Hz tuned traps, my 53Hz, 86Hz and 110Hz modes will become more pronounced.

I'm not expecting miracles, I'm very aware that it's much easier to absorb axial modes with porous absorption that's stood well away from the boundaries (I did a test a couple of years ago that confirmed that a 15cm thick panel absorbed more 40Hz the further I moved it out from the wall). My point is that I should at least observe a small incremental improvement in 40Hz absorption as I stack more boxes of 40cm thick rockwool in the corners, but this does not appear to be the case this time.

 

[darrenyeats]

As mentioned, when placed at boundary, porous can indeed absorb bass but especially less at the mode frequency itself (unfortunately). In contrast to membrane or other pressure traps like HH resonator.

 

[darrenyeats]

I have four membrane traps in the room tuned to 41Hz, but TBH I find them a nuisance. I bought them naively thinking they'd be sonically invisible outside of their bandpass but this is not the case, especially if you have a run of multiple units. They are, after all, solid boxes, so you are going to get frequencies reflecting and diffracting off them. Putting two units in each front corner, for example, generates a big, broad peak in the response at 115Hz. They are more benign on the rear wall, which is where I've put them.

Porous treatment makes more sense to me in smaller rooms because it works in all axial, tangential and oblique directions. Most membrane treatment is only designed to work in a single direction. Also, porous treatment also works on the x2, harmonics, membrane treatment doesn't. Thus, if I replace the porous corner treatment with 41Hz tuned traps, my 53Hz, 86Hz and 110Hz modes will become more pronounced.

I'm not expecting miracles, I'm very aware that it's much easier to absorb axial modes with porous absorption that's stood well away from the boundaries (a did a test a couple of years ago that confirmed that a 15cm thick panel absorbed more 40Hz the further I moved it out from the wall. My point is that I should at least observe a small incremental improvement in 40Hz absorption as I stack more boxes of rockwool in the corners, but this does not appear to the case this time.

I have foam broadband panels standing in front of my membrane traps, it sounds better to me.

 

[ToTo Man]

I took the RS45 rockwool out of the cardboard boxes to see if the cardboard impedes absorption at 41Hz. It does, but only by +0.18dB, so I'm not sure if you could class that as statistically significant?!

I also measured the bagged rockwool in a couple of different places in the room, away from the axial length mode boundary. Absorption of the 41Hz axial length mode should, in theory, improve the further away from the wall I place it. I tried putting it in front of the front corner traps (90cm from the front wall) and then centred in the rear of the room (125cm from the rear wall). Putting it in front of the front corner traps reduced the 41Hz peak by -0.28dB, while putting it behind the listening position increased the peak by +0.82dB.

I also took an additional couple of measurements with the rockwool back in the boxes just to confirm their influence on 41Hz absorption:

The difference between the 'best' and 'worst' 41Hz measurement is 1.6dB, so the RS45 rockwool is absorbing, just not as much as I'd expected.

 

[ToTo Man]

In my haste to get the frequency response measurements posted I didn't check the RT60s, but I was listening to the system tonight and swore it sounded less bright, - now I know the reason!... The RT60 confirms my earlier claim that plastic wrap absorbs mid and high frequencies while cardboard (at least partially) reflects.

It actually sounds better now. However, when I finally come to re-fitting the speaker grilles the output above 5kHz will be attenuated by approx. -2dB, which will increase the importance of preserving HF reverberation times in the room. The rockwool will therefore be going back into the boxes!

 

[ToTo Man]

A Gearspace member recently likened my room treatment approach to "re-arranging deck chairs on the Titanic". A fair point, perhaps, given that I still fall way short of their advice to treat every inch of every wall and the ceiling with 40cm absorption (i.e. sacrifice over a third of my room's volume!) to effectively deal with my modal and SBIR issues.

Feeling somewhat dejected, I thumbed back through my measurements to re-evaluate the progress I've made. The following graph compares my room in a completely untreated state vs as it is now (i.e. 40cm absorption in all four vertical corners plus a small a handful of GIK tuned membrane traps). The corner treatment consumes around 4% of my room's volume, the tuned treatment around 0.7%. I'd say the payback for treating less than 5% of my room isn't too bad?!

Waterfall for R speaker before and after treatment:

After EQ:

I'd obviously like to see the 41Hz and 53Hz modes come down a bit more, the 41Hz mode especially as this would allow me to do office work close to the rear wall without feeling like my ear drums are going to implode. However, for critical listening in the 'captain's chair', I'd much rather see improvements in the 45Hz and 78Hz nulls and the hole between 125Hz-175Hz. I'm hoping that adding treatment around the perimeter of the ceiling will help with this because it's the space in the room I'm most comfortable with sacrificing, but I won't know until I try, and doing a temporary experiment with cardboard boxes is impossible without fitting some form of bracketing to the walls so I'm going to have to take a leap of faith and get a carpenter to construct and hang Soffit-style frames that I can stuff with rockwool and cover with plywood.

 

[ToTo Man]

My first attempt at integrating my two subs in stereo configuration, using only the controls on the back of the subs. (Technically it's my second attempt, but my first attempt yesterday was a bit of a WTF experience so let's not get into that!). It's an incredibly tedious process because changing one setting (e.g. output level) more often than not also requires the other settings to be tweaked as well, so you've effectively got three variables contributing to a moving target! It's also incredibly frustrating to see one area of the FR improve at the expense of another area, so a lot of compromise is required.

This is the most benign setting I've managed to dial in. Phase at 180-degrees (knob fully clockwise), as 0 degrees phase delivered way too much summation at the XO frequency and a huge dip above this.

Before EQ:

After EQ:

Good points:
- Improvement in dips at 62Hz and 78Hz.

Bad points:
- No improvement in 45Hz dip.
- Wide hole at 35Hz (this is present regardless of how I have the subs' phase set and is there even when I measure the subs without the main speakers).

 

[EIffel]

Would it make more sense to use the two subs in mono and try to optimise their respective locations (first working with each individual sub to find locations where one or more of the peaks are attenuated, and then playing with level / timing so they integrate well together) so that together they get rid of the nulls (this may mean having both subs at the middle point on opposite walls, or at the 1/4 and 3/4 position against the front wall... or something completely different), the way people in home cinema usually do.

 

[ToTo Man]

Would it make more sense to use the two subs in mono and try to optimise their respective locations (first working with each individual sub to find locations where one or more of the peaks are attenuated, and then playing with level / timing so they integrate well together) so that together they get rid of the nulls (this may mean having both subs at the middle point on opposite walls, or at the 1/4 and 3/4 position against the front wall... or something completely different), the way people in home cinema usually do.

That's an interesting point. I'm not really sure what the best approach is. E.g. I could spend ages optimising the two subs together, then when I turn my main speakers on the combined response could look very different. Alternatively, I could spend ages optimising the left sub with the left main, and the right sub with the right main, then when I play both subs together the combined response could look very different. The approach I've taken at the moment is a mix of both, i.e. basically fiddle with the knobs until I see a response that looks good on the RTA not only with Lspk + Lsub and Rspk + Rsub but also 2ch spks + 2ch subs, and then listen to some music to see if my ears concur.

I'm definitely hearing a lot more heft down below now, and there is less level variation between bass notes. I haven't done enough A/Bs yet to identify what the drawbacks are, but it has tilted the presentation noticeably warmer. I'm not sure how much of this is due to the increased output from 60Hz-90Hz vs the increased output below 30Hz. I'd actually like to hear the subs with a subsonic filter applied below 30Hz because on certain material I get a slight headache when the subs are on and I can't quite put my finger on why this is as there is no obvious 'boom' or 'pounding'. (I also experienced this when I used the subs a few years ago, - maybe I'm allergic to full-range bass?!).

I'm limited by where I can position the subs but yes, I'm going to try them in different positions to hopefully find a spot where the nulls are minimised. Ultimately I think I'll need a DSP box for best results, though I'm still reluctant to run my main speakers through this as I like the sound of my existing DAC!...

 

[Riotvan]

I'm feeding my Yggdrasil OG with a Minidsp SHD studio for Dirac and Roon and it corrects up to 250Hz(where treatment takes over). Still keeps the sound of the Dac in tact imo you just feed it a bit of an edited signal, sure not bit perfect but i don't really care about that. Please note that for me summing the subs gives less desirable results than stereo subs with Dirac.

If you get a good result integrating the subs using the speakers full range maybe this is an option for you? A proper crossover is probably still better though and something i'm going to play with using an analog one from Xkitz. Using a preamp with active speakers though so it's easier to insert in this chain than yours.

Maybe you can get a pre and use your Yamaha as a power amp?

 

[IWC Doppel]

From my experience if you can EQ some of the excess energy from the primary room node and you can get you sub 50Hz RT60's under 700Ms, ideally 500Ms then you are pretty much there, I'd aim for 350Ms from 100 if possible. Ive heard many a system with a heap of bass trapping that measures well but sounds overly dry and anaemic. I have reduced some or my EQ and the sound, not measurement are all the better for it. Less is more with a lot of this stuff

 

[ToTo Man]

I moved the subs to the midway points on the side walls. The previous settings were no longer applicable so another hour or two of knob tweakery was required!

The left sub cancels out the peaks from the left speaker very well indeed and completely fills in the 45Hz dip, producing a virtually flat response from 45Hz-80Hz.

The right sub was more difficult to integrate in this position, I was unable to improve 45Hz and cancelling out some of the output around 53Hz also resulted in the 78Hz dip getting worse.

The combined FR is now much leaner in the 50Hz-75Hz region, I suppose that's what happens when you bust room modes, you end up with a low end that's more like the anechoic response, i.e. too lean relative to the mids and treble.

How does it sound? A bit disjointed, like there's a lot of deep bass but not enough midbass punch. On some tracks it also sounds as if the bass is coming from a location different than the main speakers. I don't know if this is because I've moved the subs further away from the speakers and my ears are localising them, or if it's the long decay times from the subs <30Hz output that's evident in the waterfall plots. (I get a similar sensation when I'm playing only my main speakers and walk towards the back wall, the increased amplitude and ringing at my 41Hz axial mode sounds like it's 'rebounding' off the walls and lagging behind the speakers).

 

[IWC Doppel]

can you check the alignment of the subs and main channels by measuring each sub and each speaker independently with a timing ref and then use the REW alignment tool to check the alignment you can also use the impulse graph for initial response, even start with the distances and delaying the subs by 1ms per foot. I am close to the expected delays when you measure well then I find sitting, listening and tweaking the delays can have benefits, even though the graph and alignment may be 'worse'. The challenge with combining two, three, four or more drivers in different locations is that they can only be aligned for one frequency at one location, so you have to compromise. I have managed this with 4 subs supporting another 7 channels. On two channel is sounds great regarding timing, plenty of slam and scale

 

[ToTo Man]

can you check the alignment of the subs and main channels by measuring each sub and each speaker independently with a timing ref and then use the REW alignment tool to check the alignment you can also use the impulse graph for initial response, even start with the distances and delaying the subs by 1ms per foot. I am close to the expected delays when you measure well then I find sitting, listening and tweaking the delays can have benefits, even though the graph and alignment may be 'worse'. The challenge with combining two, three, four or more drivers in different locations is that they can only be aligned for one frequency at one location, so you have to compromise. I have managed this with 4 subs supporting another 7 channels. On two channel is sounds great regarding timing, plenty of slam and scale

I don't yet have the ability to delay the subs, all I have is the 0-180 degree variable phase knob on each sub. This is really my first time exploring the approach of using subs to smooth out the response of the mains and I'm very much flying by the seat of my pants! I can't believe how much the additions/cancellations with the main speakers change with the tiniest change in the output level of the subs!

When you integrated your subs with your mains to achieve a flat response, did you find that the overall output level of your system's bass was reduced? If it was, how did you fix it? Is it simply a case of using EQ to shelve the overall level of the bass up?

 

[darrenyeats]

Sub's phase setting is really a delay (IOW degrees shown are valid at a specific frequency only).

Even if the position of the subs can be adjusted for best FR, the delay should be a function of distance to listener (and any intrinsic delay in the sub and/or mains e.g. active) for correct transient response IMV. The best sub system I've heard used that approach. In many cases a delay to the mains (not the sub) will be correct which requires some front end processor/circuit upstream of both. AV processors typically include such a function. Such processing needs to be truly transparent - you decide what that means for you!

(It's easy to achieve gains with a sub, but very very hard not to lose something in the process. Even if you get it right, your system will be more complex. In my case I've been happier since I let go of the sub.)

 

[ToTo Man]

@Riotvan / @IWC Doppel , when using REW to help integrate your subs with your mains, how do you tell if a stereo pair of subs is superior or inferior to a mono pair of subs? AIUI REW outputs the same signal through L and R channels, so the subs' measured performance will (in theory) be identical whether they're running in mono or stereo. The differences will only become evident when playing music that has bass frequencies mixed in stereo, which of course means the subs' integration must be evaluated through subjective listening. Or have I misunderstood?

 

[Riotvan]

Well i mainly use Dirac but it measures for left and right separately. It gives all the info i need and then i make a final judgement by ear. I had the subs asymmetrical and summed mono for a while but it was just impractical and looked weird with one sub on a stand behind me at ear level which was the optimal position at that time.
It also annoyed the neighbors a bit since it was close to a shared wall. Then i also mistook a lack of decoupling problem for a room mode.
Once the coupling was sorted with sorbothane and concrete blocks below the subs there was no need for a sub on a stand behind me, this made a huge quality difference.

So now they are both at the same height on a sandwhich of rubber and concrete tiles with sorbothane under the subs and placed between the mains. Them being close together at 1/4 from the left and right side wall each respectively also kind of couples them to make a big sub. Summing or not summing was just a difference in output really and there are no nulls going lower than the output of the mains.
Just some peaks i chop off with dirac and it's now a gentle slope from 20-20khz with a slope of about 5-6db. Still a bit of small narrow dip at 100hz i'm playing with but nothing major. Correction only goes up to 250hz here then it's all passive.

Having them stereo subjectively does do a noticeable better job on hard panned bass instruments and Dirac likes it better as well, i think mono subs confuses it a bit. There are Dirac versions with bass management for subs but mine is the standard one.
Another point i want to make is something i mentioned to you in a pm but it's worth repeating and that is time aligning using the acoustic center and not the front baffle of the speakers and subs. Which is about equal to the radius of the driver.

Found a good read here: https://www.sounddesignlive.com/do-subwoofers-need-time-alignment/

 

[geoturbo]

I am silently following your findings @ToTo Man , it's really interesting stuff.
I am afraid there are many methods to set up subs and I am not sure which is the correct/best one; I use multi subs in mono ala Earl Geddes approach (about which a lot of material could be spot online) with a Behringer DCX2496, main speakers are direct from the preamp-amp with no signal treatment. I applied counterintuitively some Geddes suggestions and they appear to work for me, for the integration one of the tricks IMHO is to get the phase right at crossover frequency and level of subwoofer too of course (phase of main speakers and subwoofer shall be the same at subwoofer crossover frequency "doing it by ear", continuous phase variation pot in the processor or in the sub is a must imho).

 

[Tony L]

Likewise following with great interest. I suspect it explains why I’ve never liked subs!