Analytical Analysis - Room Gain

[René - Acculution.com]

So-so? We see the flattening out of the SPL down to 5 Hz, with a perfect match to the theorectical response, taking into account the two rooms total volume, and they must be quite air-tight to have this response. It is not possible to show all kinds of leakage and absorption effects, as all rooms are different. The point was to show why the effect happens, and the measurements were just an unexected bonus.

 

[clemenules]

So-so? We see the flattening out of the SPL down to 5 Hz, with a perfect match to the theorectical response, taking into account the two rooms total volume, and they must be quite air-tight to have this response. It is not possible to show all kinds of leakage and absorption effects, as all rooms are different. The point was to show why the effect happens, and the measurements were just an unexected bonus.

It appears I might have reacted rather hastily! Perhaps I should not react on matters outside of my expertise at 2 AM in the morning!

What I was expecting: 9-12 db per octave boost below the first node (for a sealed sub)

What I see in the graph (with more scrutiny):

• A rise in gain from the first node down;
• A constant ~8 db rise gain from the first node down to approx 15 hz
• From then on a rather flat response to 0 Hz at approx -20db (compared to "peak")

Ported sub: no room gain

So yes there is substantial room gain, but it is not what I was expecting, which is probably due to me not understanding matters correctly.

The gain is more constant, instead of rising with lower frequency, correct?

 

[Scgorg]

The room gain does yield an increase of 12dB/octave below the first mode, but you have to remember that at some point the sub itself starts rolling off. The low-pass filter of the room and the high-pass filter of the sub cancel each other out, yielding a flat line.

 

[clemenules]

The room gain does yield an increase of 12dB/octave below the first mode, but you have to remember that at some point the sub itself starts rolling off. The low-pass filter of the room and the high-pass filter of the sub cancel each other out, yielding a flat line.

OK... Do I understand you correctly if I conclude that in this case, the high pass filter of the sub is steeper than the low pass of the room? Otherwise the measured in room response would be "flatter" closer to the first node than it is now, no?

 

[Scgorg]

OK... Do I understand you correctly if I conclude that in this case, the high pass filter of the sub is steeper than the low pass of the room? Otherwise the measured in room response would be "flatter" closer to the first node than it is now, no?

There are confounding factors at play, such as the room not being theoretically ideal, but let's look away from that. For the response to be completely flat you'd need the sub to be high-passed right around the first mode, falling 12dB/octave below that. If the sub's high-pass starts at a lower frequency, then it will still have the room gain low-pass, but not lose any output by itself, which yields a rising response toward lower frequencies until the sub starts rolling off itself.

 

[clemenules]

@Scgorg @René - Acculution.com

So.... I did some measurements. Happy to receive constructive criticism to gain a better understanding.

#1 with sub in ported (standard) and sealed (improvised) mode;
#2 with one main (bookshelf speaker, ELAC dbr62) in ported and sealed (improvised) mode.

For all measurements, db's were not calibrated. Just focussing on the shape of slopes and delta's between the different curves. Also, no EQ in place (normally I use minidsp to EQ peaks flat in the range to ~300 hz).

Conclusions / questions:

Overall: taking below into account, should I conclude that for my current circumstances, I have nothing to gain from a sealed sub in my small room? A 12db / octave gain does not seem to materialize, so I might as well go for a ported sub?

@ #1: I'm not seeing anything resembling a 12db gain per octave gain in the shape of the in room response slope. Can my room be considered "leaky"?
@ #2: Sealed slope: 40hz-20hz = 75-64.3 = 10.7db. Compared to near field slope = 14.3 db. So room gain = 14.3 - 10.7 = 3.6 db per octave in my case? Not questioning the science, but can I conclude I have a leaky room, despite it being pretty much sealed?

#1 with sub in ported (standard) and sealed (improvised) mode;
My cheapo presonus 8 inch sub, near field, in normal ported operation and in improvised sealed mode (sock + ducktape):

As expected. Sub seems to be tuned @ approx 48 hz.

In room measurement at MLP, sealed and ported mode:

Sealed sub, near field compared to in room. Db levels not calibrated, this is just about the shape of the curve:

Amcoustics predicts first node at ~ 54 hz. Predicted room gain calculated next to that.

Room is "sealed": closed windows and door. Floor and one wall concrete, floor with laminate on it. Two "gypsum" block walls if that is the correct term, one slanted sealing covered with gypsum plates. Door with 2 cm opening at bottom and one corner in gypsym brick wall that has 2cm hole for network UTP cable.

I'm not seeing anything resembling a 12db gain per octave gain in the shape of the in room response slope. Can my room be considered "leaky"? I was not expecting this.

#2 with one main (bookshelf speaker, ELAC dbr62) in ported and sealed (improvised) mode.
Dbr62 near field mid-bass measurement near field, ported and sealed (sock) mode.

Sealed roll-off from 50-25 hz = 13.1 db , 40-20 hz = 14.3db , 30-15 hz = 16.2 db.

Dbr62 in room response ported and sealed (sock).

There is a node (peak) at close to 54 hz, as predicted by amcoustics. Sealed slope: 40hz-20hz = 75-64.3 = 10.7db. Compared to near field slope = 14.3 db. So room gain = 14.3 - 10.7 = 3.6 db per octave in my case?

NB in above graph, sealed and ported cross at 35 Hz. For giggles, I modelled a run of the mill 6,5 inch driver (scan speak 18W4434G00) in WinISD in ported (green), sealed (blue) in a 13L box and a 50 hz tuning for the ported variant (analogous to dbr62). I added the predicted in room slope for the sealed variant (+12db / octave from approx 50hz downwards). Result in below graph. The anechoic response curves cross at 33 Hz, as did (more or less) the in room response curves for the dbr62. In measured response above, nothing close to resembling a 12 db gain / octave can be seen (which would have resulted in a more or less flat response from ~50hz downwards, akin to the red line below).

 

[audiofooled]

Rooms can only do so much for the frequencies that cannot fit and for the loudspeakers that are not tuned that low. Your room dimensions say that you have a mode at 54 and 63. You may try WinISD Linkwitz transform where you enter 63Hz as F0 and 54 as Fp, where your room gain would start to naturally roll off and see if it would make sense.

If you would like to extend your in room low frequency response, you would need a much bigger sealed sub that can actually reproduce those frequencies and then use DSP to level the peaks around 54 Hz.

 

[neRok]

#1 with sub in ported (standard) and sealed (improvised) mode;
My cheapo presonus 8 inch sub, near field, in normal ported operation and in improvised sealed mode (sock + ducktape):

Ported subs need HPF to protect from overexcursion below the tuning frequency. It's probably built in to the plate amp and always active, thus affecting the "sealed" response.

 

[clemenules]

Rooms can only do so much for the frequencies that cannot fit and for the loudspeakers that are not tuned that low. Your room dimensions say that you have a mode at 54 and 63. You may try WinISD Linkwitz transform where you enter 63Hz as F0 and 54 as Fp, where your room gain would start to naturally roll off and see if it would make sense.

Please tolerate my stupidity, but can you imagine this is very confusing for a non-engineer? Your statement seems to be in direct contradiction with the "12 db per octave in room gain" due to pressure vessel gain. Using your suggestion gives me a ~3db boost (constant) in WinISD. Also much lower than what for example SVS states as 7-9 db in room gain in practice.

I have measurement data near field and in room for sealed and ported operation of both my sub and bookshelves. I would like to further investigate room gain in my small sealed room.

I make the following assumptions:

• The sub's amplifier has a protection filter on the output, which matters for the sub's response, but that should not matter for actual room gain.
• For the sub, the near field response sealed equals the near field response ported from 100 Hz up;
• Thus, if I match the in room response (sealed and ported) from 100 Hz, I can judge the relative responses 100 Hz down;
• The near field response differs for sealed and ported below 100 Hz, I should compensate for this when comparing the in room responses

I thus do the following:

• Export REW measurement data to excel;
• Result 1: subtract the near field ported from the near field sealed response;
• Result 2: subtract the in room ported from the in room sealed response;
• Result 3: subtract result 1 from result 2. This gives me the in room gain for the sealed response relative to the ported response;
• Result 4: from 54 hz (my first in room node) calculate predicted in room gain of 12db per octave down;
• Result 5: compare result 3 and 4

There is a flaw in this methodology, and that is that I only measured the ported cone response near field, and not the ported port near field, spliced them together, to end up with the "true" near field ported response. So my near field ported response is too low from around the port's tuning frequency (50 hz). Which means that the delta InRoom-NearField (grey line) is probably higher then estimated below for the frequency range in which the port contributes useful output.

Result 5, graphically. We see a large difference at the port's tuning frequency, likely due to me not having taken into account the port's output during near field measurements. I'm assuming the port's output is virtually zero db at 20hz, where we see the calculated grey line getting much closer to the blue prediction.

Result 5, numerically, between 20-25 hz the average difference between grey and blue = 3 db difference between sub delta IR-NF and theoretical room of 12db per octave. I.e., leading me to believe I have a ~9db per octave room gain effect assuming that the port's contribution at these frequencies is virtually zero.

This result seems inline with SVS's predictions of actual in room gain for a sealed sub of ~7-9 db per octave from the lowest room node frequency down.

For now, I will assume a 7-9 db gain per octave for my small room as well, letting this guide my sub choice for our small viewing room.

 

[spalmgre]

Can you show your room waterfall? I think it looks like you might have a standing wave at about 53 Hz. If so EQ that down.

Don’t use Excell use Rew —-> All SPL / Action for All SPL / Trance Arithmetic. You can add measured traces like the nearfield and the port.

I also think that you should measure both speakers playing when under about 200Hz. I don’t think anyone mixing left and right at low frequencies.
I EQ my base as mono and this does work well.

 

[spalmgre]

Here are my recent measurements. I got the best base till now by measuring L, R and L+R speakers. Then I EQ the R speaker so the L+R looks nice.

The "room gain" that I get is probably a sum of three different phenomena that all seem to be related to as room gain.
Hz0 (Rene), Hz1 (Erin) and boundary gain in my case probably the floor.
See my earlier post "83 on Hz1 (Erin). Hz0 (Rene) is the initial post of this thread. Boundary gain relates to 2pi and 4pi as acoustic power gains when being concentrated into a smaller volume like a corner, close to the floor or wall.

There was also the question of my room size. The room is an open floor L-shape. The longest distance is 9,7m and the next is 5,3m height is about 2,6m.
What is special about this is that it is a new building in Helsinki that has the highest energy-saving standards. This means that windows and doors are absolutely airtight. The only leak is the ventilation that pushes air in through ducts in all rooms. The air-out ventilation is through a duct in the bathroom. So the room is very tight but has these long tubes for the airflow in and out.

 

[ayane]

Yes, but I will call that effect Boundary Gain in the upcoming article: Infinite boundaries that constrain the source to limit the solid angle it looks into, will change the acoustic power and other acoustic quantities (4pi, 2pi, ...), but that is different from having a room totally enclosing the wall with modes occuring, and different again from SBIR (boundaries(s) as well as distance from the source to the boundary(s)), so I will show these steps towards what is ultimately just the room at low frequencies acting as an acoustic compliance, with an impedance that decreases with frequency, and with displacement increasing with frequency for a flat velocity, the two counterbalance each other. So lumped or modal, the result is the same, as it has to be.

René, thank you for your excellent writeup at the beginning of the thread. Do you have a writeup on applying acoustic theory to boundary loading? Please padon the late response.

 

[René - Acculution.com]

René, thank you for your excellent writeup at the beginning of the thread. Do you have a writeup on applying acoustic theory to boundary loading? Please padon the late response.

Not really, but perhaps the Room Gain article here (https://audioxpress.com/authors/23634/rene-christensen) has some more than what is in this thread.

 

[killitmore]

Hi guys, I just measured my speakers in room and there is a peak in 1-2khz region. Is there a way to counter that peak by speakers placement/absorption instead of using EQ?