John Sayers' Recording Studio Design Forum Archive

Mr.Chino told me o register.
Maybe we can solve together many questions.

Hi all!
One, within a control room, will never use a diffuser in the far field (where it should be used), we will always be in the near field because of the dimensions of the actual control rooms.
In the very near field youll find a (high) pressure gradient sound field, let's say at distances less than 20cm from the scattering surface. At this zone you can convert a good diffuser into a good absorber by placing a thin resistive layer (let's say "fabric") at that distance.
By the way, it is better to use a diffuser in the near field than using a random surface (like a bookshelfs), because the spatial scattering and the energy time spreading are still valid.

Greetings.

Alejandro.

mmmhhhh. I like this forum. Sooo much information here!

Ethan Winer wrote:The measured modes were 20 percent higher than predicted!

If I'm not mistaken, this is a bit of an exaggeration. (The full threads are here and here.) The two (and only two) modes you looked at were indeed measured roughly 20% higher than predicted. There are various things - maybe some you missed in all the discussion of the above threads - that would account for this:

1. Since the room in question, presumably, was far from "rigid" at low frequencies, some shift is not unexpected.

2. The measured peaks can often lie within the modal bandwidth of the actual "center" frequency. Since modal bandwidth is dependent on decay time of the mode, you should probably go back and calculate the bandwidths to see if the predicted frequencies were actually "wrong." Have you done that?

3. The resolution of the analyzer can also have an effect on where the measured peak is in relation to the predicted mode. That doesn't appear to have come into play here, but it is worth noting.

But no mode calculator I know of can account for losses in the walls or other factors that can shift the frequencies.

The common equation used to predict modal behavior makes many assumptions. Among them, that the room is rigid and there is 0 damping across the frequency range of interest. This is usually not the case. AFAIK, there are software packages that will allow input of a nominal damping factor. There are also ways to incorporate it into a spreadsheet-type mode calculator. If you use the full derivation of the (spatially dependent) mode pressure distribution equation from Morse and Ingard's Theoretical Acoustics, you'll get much better correlation between predictions and measurements. If that's what's actually desired. (Usually, just finding out which frequencies might pose some challenges is the sole purpose of things such as mode "calculators.")

Ethan Winer wrote:> I´m working on a space that is 3,50 m L x 2,68 m W x 2,90 m H ... I´m trying to get as close as possible to the correct frequency, attack it <

Tuned absorption is not appropriate in a room that size. You'll get much better results from broadband absorption - rigid fiberglass at least four inches thick mounted straddling all of the corners including the wall-ceiling corners. Wood panel traps make more sense in large rooms.

There is no minimum room size that should be considered before using resonant absorbers. Whether or not this type or that type of treatment is appropriate comes down to the specifics of a situation. Not some broad generalization based on room size.

Chino,

Are you still trying to address the 100 Hz problem? My calcs show a front/back axial mode @ 98.4 Hz that is a likely culprit. That the first-order oblique mode is 100.4 Hz is probably not helping. While corner treatments will help, the most effective solution is going to be thick and/or resonant absorption front and/or back, centered at ear height. A thick (broadband) absorber spaced from the wall should help. I would suggest at least a 1-3 m² area on front and/or back, if possible. If you wish to build resonant devices, they can be effective. But you'll likely need some design help from one of the many sources given above in this thread. In my experience, the "Helmholtz" type devices, while usually not as absorbent as panel/membrane devices, tend to be more forgiving when it comes to exact construction details.

As for diffusors, I would only add an opinion to Steve's thoughts on "working distance." IMO, diffusors on the ceiling - which is where they're shown in your renderings - can be useful. Since we tend to localize sound in the lateral plane, diffusive energy from the ceiling tends to affect imaging and things far less than diffusive energy from nearby side walls. Add to that the nice height you have to work with - 290 cm or about 9.5' - and you should get some good results.

Finally, I concur 100% with the comments of Alejandro. And I would add that some research I've read about recently on the effect of certain diffusors in the near field bodes well for these types of ceiling applications.

Hi Jeff,

Nice to see you back in action!

> The two (and only two) modes you looked at were indeed measured roughly 20% higher than predicted. <

Yep.

> Since the room in question, presumably, was far from "rigid" at low frequencies, some shift is not unexpected. <

Sure, and as I stated on page 1 of this thread the room is standard sheet rock construction with a cement floor. Whether this is "far from rigid" I can't say, but it's absolutely typical.

> Since modal bandwidth is dependent on decay time of the mode, you should probably go back and calculate the bandwidths to see if the predicted frequencies were actually "wrong." Have you done that? <

No, I never went back. But you can see from the ETF waterfall graphs that all of those modes are very high Q - far narrower than 20 percent.

> There is no minimum room size that should be considered before using resonant absorbers. <

Okay, if you say so.

Thanks for clarifying.

--Ethan

Ethan Winer wrote:Nice to see you back in action!

I was never out of action.

Sure, and as I stated on page 1 of this thread the room is standard sheet rock construction with a cement floor. Whether this is "far from rigid" I can't say, but it's absolutely typical.

I didn't say it wasn't typical. And it is "far from rigid" in the acoustical sense.

> Since modal bandwidth is dependent on decay time of the mode, you should probably go back and calculate the bandwidths to see if the predicted frequencies were actually "wrong." Have you done that? <

No, I never went back. But you can see from the ETF waterfall graphs that all of those modes are very high Q - far narrower than 20 percent.

For the first axial mode in the thread we're referring to, the resonance can occur within 7-8 Hz of the predicted 34.8 Hz provided the decay is ~0.3 seconds (or lower). From the waterfall, it's tough to tell how much decay is present in that range. Maybe it's around 0.3 s, maybe it's not. Let's say it is: Since the source used was very likely not flat over the entire LF range, all it would take is a stronger source resonance at a frequency within the modal bandwidth to appear to cause a "shift," so to speak. In other words, what the measurements are showing might be exactly what would be predicted with more "beefy" methods. E.g., measure the source response in an anechoic environment and correlate it with the room measurement.

That's one possibility.

Another - probably much more likely - one is that the room's damping over the range of interest is enough to cause a shift in the resonance. This is not uncommon. And there are methods to predict it.

> There is no minimum room size that should be considered before using resonant absorbers. <

Okay, if you say so.

I don't. Physics say so.

Thanks for clarifying.

You're welcome!

Jeff,

> the room's damping over the range of interest is enough to cause a shift in the resonance. This is not uncommon. <

I'm sure, but my understanding is that adding absorption or wall losses etc will always shift the frequency down, never up. This is the part I'm having trouble understanding since the shifts I measured, and others have reported, were to a higher frequency. What do you think might account for that?

--Ethan

Ethan Winer wrote:I'm sure, but my understanding is that adding absorption or wall losses etc will always shift the frequency down, never up.

That's not what I know. Adding "traditional" absorption typically does shift the modes down. However, considering the mass impedance of something like a typical gyp-board wall, the modes usually shift up. Since the latter is typical for the environment measured in the thread cited above, I would say that it accounts for a large part of the shift you observed. This would tend to conform to many of the measurements I have taken in similar rooms.

Jeff,

> Adding "traditional" absorption typically does shift the modes down. However, considering the mass impedance of something like a typical gyp-board wall, the modes usually shift up. <

This is very interesting, please explain more. I had ASSumed any change from the walls would be simple absorption at some range of low frequencies. How is this different from "mass impedance" you mentioned?

--Ethan

Ethan,

I think there's a discussion of non-rigid wall behavior in Kuttruff's Room Acoustics. From what I can recollect (I don't have my copy here with me), a non-rigid boundary with a spring impedance lowers the eigenfrequencies (modes) and a non-rigid boundary with a mass impedance raises them. I believe it was Morse who did most of the practical investigations.

Jeff,

> a non-rigid boundary with a spring impedance lowers the eigenfrequencies (modes) and a non-rigid boundary with a mass impedance raises them. <

Okay, thanks, that makes sense. I ran it by my expert friend Bill Eppler who put it into electrical terms for me. One is like a transmission line with an out-of-phase back-reflection caused by a short circuit at the far end (mass impedance, like sound hitting a brick wall), and the other is like an in-phase reflection caused by an open circuit.

I'm still not sure how a wall could have other than a mass impedance, but I'll gladly take your word for it that this can occur. (Though if you'd like to elaborate, I'm all ears.)

--Ethan

Ethan Winer wrote:I'm still not sure how a wall could have other than a mass impedance, but I'll gladly take your word for it that this can occur. (Though if you'd like to elaborate, I'm all ears.)

Um, make it soft? Add foam/fuzz? Perhaps this is where you got the original idea of lower modes? Porous absorption in the sense of foam/fuzz acts as a spring impedance. Membrane absorption - which is effectively what a gyp-board over studs wall provides - would be mass impedance.

I should note that it's been quite a while since I've reviewed the concepts involving non-rigid walls. So, if there are any physicists lurking, I disclaim everything - I'm going from memory here and it would not be surprising if I have greatly over-simplified things!

However, since the behavior we're talking about jives with what I've seen in countless measurements, I would be inclined to think it may have something to do with it. YMMV

Jeff,

Thanks.

I'm really looking forward to seeing you on our panel at AES this year.

And I absolutely promise not to mention anything about modes versus standing waves there.

--Ethan