Building an echo/reverb chamber

[DeadPoet]

Hello all,

I've searched around on the site but there is still very little info on building your own reverb chamber.

I have a spare basement of about 150m³ (+/-5300cubic feet). L-shaped, brick (painted) walls, a brick floor with a faint V-shape, plastered brick ceiling which is divided in several arcs.


I'm planning on putting a powered speaker in there with (for now) 2 Behringer ECM8000's but what I have no idea about is how much SPL should I be generating to get a useable reverb out of that room?


(there is a semi-used living room on top of that basement and it is just brick ceiling, wood beams and wood floor on top - that won't stop a lot of dBs)


Herwig

[spm_gl]

We often use our piano room as a reverb chamber, simply put one of the Behringers (2031?? not sure) from the cutting room on a stand. You don't really need much SPL for a nice, smooth reverb. Just make sure you don't have the TV from the livingroom in your reverb tail.

[Ethan Winer]

Originally Posted by DeadPoet

I have no idea about is how much SPL should I be generating to get a useable reverb out of that room?

Room acoustics is linear, so whatever happens at low levels happens the same at high levels. So the main issue is getting enough SPL volume to be above the ambient room noise.

--Ethan

[DanDan]

The lack of sound isolation will be a problem. Reverbrant rooms can be effectively 5-6 dB more 'sensitive' to incoming noise than deadened treated rooms. Also if you use high SPL's the will surely be audible upstairs in your Control Room. Do you use Convolution Reverb, e.g. Altiverb? If so, I suggest you sample the space at a quiet time and use that.
I have an irrational turn against anything Behringer. If you are sampling the room or experimenting with it I suggest trying some different mics. Clue- the classic chambers at Cello Studios used SM57's. I would go for an Omni Dynamic, a talkback mic.
DD

[spm_gl]

I'd wouldn't use a dedicated reverb mic, rather see the room as an intrument, and experiment with what you have and where you can place it. You'll need an assistant though.

[DeadPoet]

Thanks for the input.


Actually, i quite like the behringer omni's, they share the capsule of an Earthworks QTC1 and in a listening test they match DPA 4061 easily. I use them for drum room sometimes, mostly with them pointing up at the ceiling. Was thinking of trying that technique as well. The chamber will be mono in - stereo return.

But I'll try the 57-omni thing as well off course.
General idea seems to be omni, right?


The basement is under the house while the studio is at the other side of the house, so control/live room bleed will be no issue. The living room above the basement is not used all the time (we have the luxury to have a small living room with tv/couch/family stuff and a larger one with large dining table, piano,...) so *some* leaking won't be a real problem.



I hope to be able to realise this project somewhere in August, i'll report back then. But if anyone has more info, I'll be happy to know.



Herwig

[Northward]

Originally Posted by Ethan Winer

Room acoustics is linear

Not quite... Air absorption capacity vs SPL will play a major role in how HF are behaving in the reverb room for example. And this will have a major influence on the quality/color of the reverb. The SPL vs "color of reverb" (so response) ratio there is for sure not 1:1, in particular in HF. Also, if you have more energy in the room... It takes more time for it to transform in heat etc... So it implies perception will be different.

It's like all treatment. Semi-rigid Rockwool for ex isn't behaving in a linear way at all. But no model can predict this today. What is clear is that only a certain amount of energy can be dealt with by this type of materials - that we know. So it's important to keep this in mind whenever dealing with room acoustics.

Problem is that you can build all the equations that you want, you won't be able to put your finger on it. I discussed of this yesterday with a colleague, and he was stressing that in our field of physics, there is a real need to understand that we are dealing with partial info, based on "flawed" measurements (in the sense that you never get twice the same exact measurement even if 'conditions' are the same) and that a lot of what we do has to be dealt with carefully.

To get a good reverb in there you will need rather high SPL. And very low backgound noise if you want a good tail. Good pres, good mic with low noise.

[DeadPoet]

Originally Posted by Northward

To get a good reverb in there you will need rather high SPL. And very low backgound noise if you want a good tail. Good pres, good mic with low noise.

mr T.

Thanks for the advice. I'll experiment a bit with it. Let you have a look at it next time you get here


H.

[DanDan]

DP seems like you should just go for it. Find a loud and robust powered speaker. I suggest a PA speaker rather than a studio monitor. You don't need extended frequency response. Mackie have some very nice ones. SRM 450 if I recall correctly.
Omni will probably have the more open sound, and again you don't need extended frequency response. Lexicons start their roll offs at 2K! So an omni dynamic would be
ideal I reckon. I would say it matters little, the Cello chambers with the SM57's are really good, long, clean and bright, with not nasty colours. They appear to be a small damp stained unused basements. Try your lookalike microphone, made by one legged blind orphans, working 27 hours a day, if you must, but you will suffer in the next life.....
I am sure the microphone noise will be a problem though.
Try handclaps in there, if you find any obvious twanging flutters, kill them by hanging blankets or whatever, and avoid putting the source or mics in those areas. You will have fun, and please do get someone with Altiverb to sample it for the worldl to enjoy.
DD

[spm_gl]

The 2k rolloff is a great idea. I have a few old oval full-range speakers lying in the workshop. Perhaps we should try those for our reverb room. Maybe not having a crossover makes an even evener reverb response.

[Ethan Winer]

Originally Posted by Northward

Not quite... Air absorption capacity vs SPL will play a major role in how HF are behaving in the reverb room for example.

Okay, room acoustics is linear at SPL levels that can be tolerated by humans, but possibly become slightly nonlinear six feet behind a jet aircraft engine.

Is that better?

Semi-rigid Rockwool for ex isn't behaving in a linear way at all.

On what do you base that? If this were true, adding rock wool into a room would add audible distortion products. And we know that's not the case.

--Ethan

[Northward]

Originally Posted by Ethan Winer

On what do you base that? If this were true, adding rock wool into a room would add audible distortion products. And we know that's not the case.

--Ethan

Laws of physics. I think maybe you misread me.
Non-linearity means it's behaviour will vary with SPL variations - such that efficiency will not be linear. And more than one would think, especially in LF. There is only so much energy that can be realistically 'managed' by a defined amount of material. Then again, it's impossible to come with a realistic model of such behaviours (AFAIK), as they are too complex and particular.

Have a nice day

[spm_gl]

F.P. Mechel, Formulas of Acoustics, Berlin 2002
Chapter S, Nonlinear Acoustics by O.V. Rudenko:
"The strength of an acoustic field can be characterised by an acoustic Mach number:
M= v/c0 = d1/d0 = p1/p0 [we really need a formula editor, d is actually ro]
Parameters of the equlibirum state of the medium are: c0 = adiabatic sound speed, d0 = density and p0 = c0²ro0 = internal static pressure. The acoustic wave leads to variations of density d1 = d - d0 and and pressure p1 = p - p0, as well as to the appearance of nonzero particle velocity v. The limit M -> 0 corresponds to linear acoustics. Nonlinear acoustics deals with small, but finite values of M."

So Ethan is right, non-linear acoustics happen at M >> 0. Not the situation we deal with in normal rooms.

[spm_gl]

Originally Posted by Northward

... There is only so much energy that can be realistically 'managed' by a defined amount of material....

True, it is possible to ignite cotton with ultrasonic beams (If I remember correclty at 150dB SPL). But that's not the amount of energy we normaly deal with.

[andrebrito]

I will go even further than Thomas and state that acoustics is non linear even at the same SPL.

I did several measurements in all kinds of buildings: usually normal typical restaurants, bars and bedrooms due to building acoustics but also in some concert halls and theaters always using a pro sound level meter and an omni sound source.

When measuring RT and transmission loss which is what I usually measure at I always get different results particularly at low frequencies.

Above all acoustics is a statistical science and based in empirical measurements. Formulas are used for us to understand the real world or the perception of it and not the other way around.

[andrebrito]

Room acoustics is linear, so whatever happens at low levels happens the same at high levels. So the main issue is getting enough SPL volume to be above the ambient room noise.

--Ethan

Not only room acoustics is not linear but more important, the human perception of the acoustics varies greatly with the SPL.

Even if the room acoustics were the same at different level (which is not) we would perceive it as different

[spm_gl]

We need to define linear here, because there seems to be a misconception. If the room acoustics change linearly with SPL, it's still linear acoustics.

Above all acoustics is a statistical science

Then you need to repeat the same measurement very often to get a valid sample.

[Ethan Winer]

Originally Posted by andrebrito

When measuring RT and transmission loss which is what I usually measure at I always get different results particularly at low frequencies.

That's not nonlinearity, that's just due to not having identical test conditions. If you move the measuring microphone even an inch a lot can change, as I'm sure you know. Nonlinearity usually means distortion is added to the sound.

the human perception of the acoustics varies greatly with the SPL.

Of course, Fletcher-Munson at work. But 1) I did't think we were talking about human perception, and 2) that's not nonlinearity either.

--Ethan

[andrebrito]

Ethan

I was talking about making the measurements on the same and exact location with the same sound source and sound level meter.

You get different results. Differences are quite significant when the room is reverberant.

[Northward]

Originally Posted by Ethan Winer

If this were true, adding rock wool into a room would add audible distortion products. And we know that's not the case.

Originally Posted by Ethan Winer

Nonlinearity usually means distortion is added to the sound.

Originally Posted by Ethan Winer

Room acoustics is linear, so whatever happens at low levels happens the same at high levels.

I'm really not sure you're with Andre and I here as to what is being discussed. What we are trying to say is that your statement "whatever happens at low levels happens the same at high levels" is not realistic, due to many factors, some mentioned in previous posts.


On 'distortion': If you take an empty room with a given response. The response at point A being the sum of modal response, ER and so on. You can call the effect of the room on the signal emitted by the speakers as distortion. Adding rockwool, even if it's behaviour in transfering acoustic energy into something else (heat) is non linear vs SPL, you are still reducing the amount of energy in the room by a certain amount. So you are reducing "distortion" from the room. Only in some cases will you get added distorsion such as re-emission (as parasite vibrations induced by a membrane releasing energy for ex).


So it's a lot of shades of grey here. Blanket statements don't apply.


Non-linearity doesn't mean at all that you're adding distortion. Just that the response at point A at 50dB isn't the same as the one at 100dB. The opposite would be strange.


Now, off for a nice cold beer!

[DanDan]

Andre, I have been banging on about 'flaky' measurements elsewhere recently. In my case I am not happy that I cannot get reliable decay measurements below say 100Hz in small rooms. I tried a little experiment last night. I got some third octave pink noise from the SoundCheck CD, 31-200 Hz. I trimmed them to stop suddenly without clicking.
I then played this back, recording the sound in the room. Old school eh? A complete waste of time! The decays are tiny and hard to see. I may look at it again with some gating but I digress.
The lower third octave waveforms are visibly very very random. Huge level variations, presumably at different frequencies, if one can say that. So stopping such a burst randomly will give a very different stimulus to the room each time. Maybe this is a clue to the variations in Building Acoustics measurements that you are seeing? I note that Standards are moving downwards these days, 63 or even 50Hz is involved. This will of course make this more difficult. We need 'quality' indicators for each sweep or measurement, and indeed some instruments and software provide that now.
DD

[Rod Gervais]

Originally Posted by andrebrito

Ethan

I was talking about making the measurements on the same and exact location with the same sound source and sound level meter.

You get different results. Differences are quite significant when the room is reverberant.

Folks,

ASTM C423 notes the following in their publication for testing in reverberation chambers:

6. Interferences

6.1 Changes in temperature and relative humidity during the course of a measurement may have a large effect on the decay rate, especially at high frequencies and at low relative humidities. The effects are described quantitatively in ANSI S1.26. These effects due to temperature and relative humidity changes
can be accounted for by the procedure in 6.2.

I've seen this - with some huge swings taking place just due to humidity when everything else is exactly the same as a previous test.

Rod

[andrebrito]

Rod,

The ISO in Europe are similar to the ASTM and they say pretty muich the same thing.

Temperature and humidity values don't change that much over 1 minute and using that interval I can make two or three acoustical measurements and they won't match particularly if we are dealing with large rooms such as a church or a theater.

[spm_gl]

I guess uncertainty in low frequency measturement has a lot to do with the nature of low frequencies, i.e. if you measure for 1s, you have 1000 "samples" at 1kHz, but only 50 at 50Hz. So what? To get the same statistical precision, you'd have to measure for 20s at 50Hz. Should be obvious, but somehow we tend to forget that frequency is actualy a temporal measure.

[Ethan Winer]

Originally Posted by andrebrito

Differences are quite significant when the room is reverberant.

I'm not trying to be combative, but I don't understand why successive RT60 measurements would change even a little if the test conditions are identical. Now, maybe that could happen if you use pink noise as a sound source rather than a swept sine wave. In that case the change is from the random nature of pink noise.

When I did the tests for my new video I measured at a few nearby locations. I didn't note the exact distances, but the two shown below were maybe six inches apart. As you can see they're pretty similar, and very different from the "no diffusors" and "no treatment at all" RT60 graphs.

--Ethan

[Ethan Winer]

Originally Posted by Northward

What we are trying to say is that your statement "whatever happens at low levels happens the same at high levels" is not realistic, due to many factors, some mentioned in previous posts.

I'd sure like to see some evidence of that. What do you think would change if a test signal is played at 70 dB SPL versus 100 dB? I'm sure the frequency response and ringing would be identical, and I can't see why RT60 would be different either. So what then would be different? And more important, why?

If you take an empty room with a given response. The response at point A being the sum of modal response, ER and so on. You can call the effect of the room on the signal emitted by the speakers as distortion.

I guess it depends on how you define distortion. All nonlinearity is distortion, but not all distortion is nonlinearity. I prefer to use "distortion" only for nonlinear behavior, but I accept that others consider phase shift and frequency response changes to be a form of distortion. But we're talking about nonlinearity here, and my initial statement was, "Room acoustics is linear, so whatever happens at low levels happens the same at high levels."

Non-linearity doesn't mean at all that you're adding distortion. Just that the response at point A at 50dB isn't the same as the one at 100dB.

Can you show a pair of in-room response measurements at those two levels where the response is not identical? The only thing that could make two tests at 50 and 100 dB SPL seem to have a different responses is if the noise floor prevents the full extent of nulls from showing. If the noise floor is 30 dB and a null is 30 dB deep, at 50 dB SPL the null will appear only 20 dB deep because it's filled in by the room noise. At a higher SPL level the true null depth will be revealed. I mentioned earlier the importance of being above the noise floor.

--Ethan

[jwl]

Agreed, if the measurements change, then SOMETHING in the room must have changed, whether it's humidity, or the random nature of pink noise producing a different waveform at each instant, or humans standing in a different location during each test, or ....?

[DanDan]

In my and obviously others experiences the Measurements do change. This seems to be widely understood and dealt with. High end equipment makes a number of sweeps in each band to be sure of a result. There are 'quality indicators' provided to give an idea of how valid. Sometimes the result is rejected and one has to go again. I believe REW has this quality or confidence feature?
Two of us have suggested reasons why this happens at LF. i.e. The Stimulus signal is quite variable. If you stop a 63 Hz tone or third octave of pink suddenly, the waveform and energy content will be very different at each random stop. Also the period of the Waveform can be close to the first 10% of the slope used to derive EDT. Take a look at the waveform of some recorded third octave 63Hz and you will see what I mean.
I don't know if this old school tone burst understanding can be translated to Sine Sweep and MLS methods but I suspect similar anomalies will happen. I note REW does not do RT below 80Hz. Rather than present unreliable info John has left it out, a reasonable call IMHO.
Regarding the Non Linearity. That seems a bit deep, and Ethan can turn into a bit of a Terrier sometimes....:-) However, just out of interest, I also would like to see some numbers on the Non Linearity alluded to
i.e. what type of change, how much in level or time, at what frequency area, at what SPL?
Rufff, DD

[hendrikxix]

I am going to be working on building a reverb chamber at U. Mass Lowell in the Fall with Connor Smith. One of the rules that we created was to use no absorption only diffusion.

As soon as we start, I will send pix.

[Ethan Winer]

Originally Posted by DanDan

If you stop a 63 Hz tone or third octave of pink suddenly, the waveform and energy content will be very different at each random stop.

That could be, which is probably another reason REW uses a swept sine wave, in addition to a sine sweep being better than pink noise for frequency response too. When software sweeps a sine wave the identical signal is output every time. And a room should respond the same way every time too.

Regarding the Non Linearity. That seems a bit deep, and Ethan can turn into a bit of a Terrier sometimes....:-)

It's a fine line between being passionate and standing up for what you believe in, versus being an asshole. I try hard to stick to the former but I may not always succeed.

If anyone can show a room responding differently at different SPL levels, I'll be glad to change my position!

More on that: Earlier I said that nonlinearity adds distortion into the room, but I didn't explain why. Assuming at least one person here cares, I'm glad to explain.

Using a sine wave as an example, SPL is an average of all the points along the sine wave. A single sine wave at 100 dB SPL has a minimum (zero) and a maximum (about 1.4 times the average level an SPL meter would display). So a sine wave does not have a single level, but rather a constantly changing level. Now, if you add a material into the room that "behaves" nonlinearly (whatever that means), as was suggested earlier for rockwool, that means the low-level portions of the wave will be affected differently than the high-level peaks. And if that's true the result would be distorted music with the inevitable added harmonics and IM products! And we all know that's not true.

--Ethan