Is this a rip-off?

[saemola]

A friend of a friend of a friend of a... Is a small theatre production director and they have recently gauges access to a rehearsal space for the actors. The place is a cave, 145 sqm and 13 m tall, with insane reverberationtimes.
They asked a company to take care of the acoustic treatment of the place, enough to make dialogue intelligible and allow for shows rehearsals.
The company got paid €1500 to present an acoustic treatment project and execution plan (which invoices 32 polyester fibre panels with a surface of 2 square meters each, all of them suspended from the ceiling as the room is a historical relevant space and they don't want to touch the walls. Also, they planned for 4 corner floor to ceiling bass traps for a grand total of €13000 + VAT.
According to them, this should lower the reverberation times from about 9.2" at 250Hz to about 3".

The price seems high to me, but I wouldn't question it if it wasn't for a small detail: the acoustic firm never took acoustic measurements of the room.

They came up with this project but looking at construction for plans of the room and looking at pictures.
So my question for you is: is this standard procedure?
It seems completely insane to me that they would get paid €1500 for a plan and an estimate of the reverberation times of the room without even surveying the place.

By the way, reverberation times according to them are:

9.2"@125Hz
9.2"@250Hz
8.8"@500Hz
7.0"@1kHz
7.3"@2kHz
7.1"@4kHz

Which after treatment should go down to:

5.0"@125Hz
2.8"@250Hz
2.0"@500Hz
2.5"@1kHz
2.5"@2kHz
2.6"@4kHz

I'm skeptical. The theatre guys are afraid of turning the proposal down as they have already paid €1500 for the plan.
What do you think?

[JCBigler]

It's nearly impossible to tell without knowing what the exact dimensions of the room are and what the materials are that it's built with. Bass traps in the corner and hanging absorbers from the ceiling will soak up some of the reverb time.

A 3 second reverb time seems...ok-ish to me. But what is the reverb time at all of the other frequencies? Maybe there's some high frequency echo that needs to be dealt with, maybe there's low end build up.

It's just too difficult to know without knowing more about how the room is built.

Theatres and live performance spaces often have longer reverb times than your standard recording studio.

[saemola]

Hey Justice,

I reported the current (according to their calculations with no measurements) reverb times and estimate reverb times after treatment at the bottom of the post.

But my question was more about the way they can't up with the project: is it considered standard practice to do so without taking measurements on site?
It sounds insane to me, but then again I am no expert.

[Gregwor]

There is advanced software out there such as EASE or Odeon that can quite accurately plot acoustic responses of spaces.

They would have to use fire retardant fabric for the panels, because law states that in a commercial facility like that, any sound treatment device that is fixed to the building must be fire retardant. This would really bump up the price of materials. However, as all of us on the forum know, making 32 broadband panels does not cost €13000 + VAT.

Greg

[Soundman2020]

enough to make dialogue intelligible and allow for shows rehearsals.

OK, but what were the actual numeric goals? This should have been stated in terms of speech transmission index, or speech clarity, or some such objective system. For example: "STI greater than 0.75", or maybe "D50 > 0.85" or something. Measuring decay times at specific frequencies is part of it, but much more important for speech clarity and music clairty, is the ratio of early sound to late sound. That's what is measured by objective methods, such as C50, C80, and D50.

According to them, this should lower the reverberation times from about 9.2" at 250Hz to about 3".

Large room acoustics is not the same as small room acoustics. The principles are the same, the goals are similar, but the treatment methods and materials are not the same. For example, small rooms needs major bass trapping in the corners to get acceptable low-end response, mainly due to insufficient modal support, but large rooms DO have modal support, so they don't need as much bass trapping, and it doesn't need to go in the corners. Large rooms generally need diffusion more than absorption (although they still do need some absorption), and the mid range is where the issues are likely to be, not so much the extremes.

32 polyester fibre panels with a surface of 2 square meters each, all of them suspended from the ceiling as the room is a historical relevant space and they don't want to touch the walls. Also, they planned for 4 corner floor to ceiling bass traps for a grand total of €13000 + VAT.

So 13k Euros will buy 64 m2 of absorption and 4 large bass traps? It does sound a little high. That works out to €200 per square meter of absorber? Sounds higher than what I would have expected. But I'm not there, I haven't' seen the room, I have no idea of the complexity of the installation, so it might be justifiable too. That doesn't consider bass traps, but even if the bass traps are 1/3 of the total, that still means €135 per square meter for the absorbers. Does that include installation? If so, it's more reasonable. What type of finish would be on the absorbers? If it is high-end Guildford of Maine fabric on both sides, that could be maybe €20-30 per square meter. 4" OC-703 (more expensive than polyester!) would run about the same ( €20-30 per square meter ), plus frames, so maybe €50 - 70 / m2 in materials, plus labor, plus installation. Maybe €80 to €90 / M2. But that's for high-end materials 703 core with GoM fabric. They could be built a lot cheaper than that with polyester and lower-grade fabric...

You didn't say what the actual dimensions are, but at a rough guess there's going to be about 900 m2 of surface area in there, assuming 4 walls plus floor plus ceiling. This plan is for 64m2 of treatment, to deal with 900m2 of surface... ? Hmmm.... In theory, you'd need something like 2200 sabins of absorption, and 64m2 is only about 700 sabins... However, if all of those panels are double-sided, then there's more like 1400 sabins... Closer to what is needed in theory, but still falling short.

So, bottom line: the plan will have some effect, and will certainly improve the acoustics of that room, but I'm not convinced that it's going to give you a room that is great for speech clarity, if that's what you are looking for. At 3 seconds RT-60, that's still rather long for speech, or most types of music, considering that the room is not huge. For good speech clarity, I'd be looking for RT-60 of maybe 1.2 seconds or less, and STI over 0.8. For music, RT and C80 could be higher or maybe C80 lower, depending on the music genre (jazz vs. Gregorian chant, for example, are rather different).

Maybe your friend could insist on getting some type of guaranteed result here, in terms of objective speech clarity numbers, and also insist on "no results, no payment". Eg, if the spec is for "C50 > 0 dB, full spectrum", but the result is only C50 = -3 dB, then there's no payment, and the company must remove all of the treatment at no cost. So your friend would only pay if the agreed spec is actually met. Of course, the spec would need to be defined carefully! "make dialogue intelligible and allow for shows rehearsals" is not a valid specification! It must be defined properly, based on what they need for their rehearsals.

The price is a little high, maybe, but without knowing what is involved in the manufacture and installation, it's hard to say if it is excessive. If that is all high quality materials, an the installation is complex (hanging everything at a height of 13m implies renting scaffolding, and safety gear, and maybe there are restrictions on how it can be mounted if the room is historic, etc.) then it might be reasonable.

But getting that "no result = no pay" guarantee, and a solid spec number to judge it by, would be they best way to do this.

- Stuart -

[saemola]

Maybe your friend could insist on getting some type of guaranteed result here, in terms of objective speech clarity numbers, and also insist on "no results, no payment". Eg, if the spec is for "C50 > 0 dB, full spectrum", but the result is only C50 = -3 dB, then there's no payment, and the company must remove all of the treatment at no cost. So your friend would only pay if the agreed spec is actually met. Of course, the spec would need to be defined carefully! "make dialogue intelligible and allow for shows rehearsals" is not a valid specification! It must be defined properly, based on what they need for their rehearsals.

That's exactly what I suggested. If the acoustic firm thinks they can hit those numbers and doesn't feel like taking measurements, then they should at least guarantee the results. Especially for that price.
I was simply shocked by the fact that they didn't inspect the site and take measurements, but that seems to be just me as you guys seem to be quite ok with it.

With less understanding and experience on the subject, my math came up similar to yours. I think it's expensive, but probably not crazy.

Being that this is not a direct friend I don't know much about the situation, and your question about target RT makes perfect sense, but I think this was more of a case of: we can't touch the walls and we have X money, what can we get out of this?
Either way, my question was more about the legitimacy of the project without having taken measurements.

A few extra notes:
- panels are indeed double sided
- will be suspended by the ceiling mounted to metal wires. should be a standard installation as it's presented on the project.
- room is 16x9m, 13m tall

Projected reverb times with the room filled with 100 people are:
2.0"@125Hz
1.5"@250Hz
1.0"@500Hz
1.0"@1kHz
1.0"@2kHz
1.0"@4kHz

However, it's unlikely the place will be filled with that many people. More likely 20-50.
I was wondering about suggesting thick curtains to place on the walls if that's an option for them, as that would allow them to not permanently damage the walls but still get that extra absorption by unrolling the curtains when required. Not sure about the cost though, so I will have to investigate.

Thanks all for chiming in!

[Soundman2020]

I was simply shocked by the fact that they didn't inspect the site and take measurements, but that seems to be just me as you guys seem to be quite ok with it.

In my opinion, its always better to take measurements! Yes, it is possible to predict performance, but that might not be totally accurate, since the predictions are based on assumptions that might not actually be correct. For example, when predicting the response of a room, you normally just consider the dimensions, but if there happens to be a wall that is resonating at some frequencies, that would change the room response, so it would be different from the prediction. etc. So it is valid to predict, yes, but it's better to measure as well, to confirm the predictions.

Projected reverb times with the room filled with 100 people are:

Hang on a sec! Something strange there...

First they said

Which after treatment should go down to:

5.0"@125Hz
2.8"@250Hz
2.0"@500Hz
2.5"@1kHz
2.5"@2kHz
2.6"@4kHz

And now they are saying:

2.0"@125Hz
1.5"@250Hz
1.0"@500Hz
1.0"@1kHz
1.0"@2kHz
1.0"@4kHz

Just from having 100 people in the room? They project a reduction from 5 SECONDS to 2 SECONDS at 125 Hz, from just 100 people in the room? And going from 2.8 to 1.0 at 250 Hz? I don't think so!

I'm calling BS on that one! Contrary to popular belief the human body is NOT a good acoustic absorber. It acts much more like a diffuser than an absorber. Acoustically, you can model it as an elongated bag of water of roughly the same dimensions, and you won't be far wrong. The naked human body has a total absorption of about 0.1 to 0.5 sabin. See if you can find a report released by AMS Acoustics a few years ago, reference number IR/57. They actually measured this in an acoustics lab (put real people in the reverb chamber, and measured the response), and came up with real numbers that were much lower than the "assumed" numbers that had been used before. They found that absorption is almost none existent for low frequencies, somewhat better in highs. At 250 Hz, they measured 0.16 sabins/m2, rising to 1.03 sabins/m2 at 4 kHz.

So I have a real problem believing that having a hundred people standing in that room will drop the decay time by more than half at 250 Hz!, and even more at 125 Hz!

Clothing makes a difference, yes, but even if all of those people are dressed up in thick parkas with thick Russian hats on their heads, as well as gloves, scarves, and woolen boots, I'm still not convinced you'd see reductions of 50% to 70% in RT-60 times. Perhaps if you also added seats with a lot of upholstery and fabric covering, you might get close...

First-hand experience: the church I attend had to temporarily move to meeting in an event hall late last year, as the building we rented before was sold by the owner. The hall we moved to is about 186 m2, and attendance runs anywhere between about 100 and 165 people. The room is a concrete cave! I did the math before we moved in, and figured we would need 161 m2 of absorption to get STI where I wanted it, and RT-60 of around 1 to 1.5 seconds. 180 upholstered chairs would provide about 50% of that, 150 people would provide another 35%, and the rest I would have to make up with panels. I put 20 panels in the room, on the walls, covering a total of 37m2, and that provided the remaining 25% (actually, 23%). So when we have 150 people in there, it does indeed sound fine. But when there are only 100, it sounds a bit off. And when there is nobody in there (just chairs and panels), it is rather poor, but still possible to talk to someone at the other end of the room. However, when there are just 150 people in there (no chairs, no panels) it is a nightmare! Ditto when there are people and panels, but no chairs. Of course, when the room is completely empty, it is impossible to have a conversation with anyone further away than about 3m, and music sounds disgusting!

I would be skeptical that they will hit those RT60 times you quoted, from occupancy of 100 people, unless there are also upholstered chairs, and the people are wearing thick clothing. I suspect those guys are using the "assumed" absorption of the human body, not the real, actual, measured absorption. I would expect that the RT-60 with only 50 or so people in there is going to be too high for good speech clarity, and also too high for musical performance (except maybe for a church pipe organ and Gregorian choir: that sounds really nice with long reverb times....). I think they are overestimating the effect of the human body, and also overestimating occupancy. If I had to guess, wildly, I'd say that your overall RT-60 time (full spectrum) is going to be over 3 seconds for the "typical" scenario that you mentioned, of maybe 50 people in there and only 32 panels hanging from the ceiling. That's too long, in my opinion. I would want to see it down near 1 second for speech, and maybe 1.5 to 2 seconds for music.

- Stuart -

[saemola]

Just from having 100 people in the room? They project a reduction from 5 SECONDS to 2 SECONDS at 125 Hz, from just 100 people in the room? And going from 2.8 to 1.0 at 250 Hz? I don't think so!
I'm calling BS on that one!

It does sound like bullshit to me as well.
Talking with the people from the theater, they told me that another guy came in, did the balloon popping test and told them their reverb time is around 3 seconds...
At this point I had to see for myself.
I was in the area today and decided to do a quick test.
I'm not an expert, and this wasn't conducted with the most appropriate gear (AKG 414 ULS, Tascam DR-60D and a balloon) but I think simply using a phone would have told me whether the room was actually closer to the estimated 8" or to the 3" of the last guy.

Again, I'm no expert, but unless I'm screwing something up, the RT60 of the room is DEFINITELY closer to 3 seconds.
I'm attaching 2 files (in mp3... cause the forum won't allow wav or flac).



Walking in I was expecting a place where it would have been impossible to talk, whereas it is actually much more decent than that. Granted, standing 10 meters from another person made talking hard, but I would guess 8 seconds would make it hard standing 1 meter apart.

What do you think?
I wonder how their math could have been THAT wrong (unless mine is).
There is a large skylight window (about 4x4m) at the center of a half DOME SHAPED ceiling, and the entry door (which will remain open and leads to another small area) is about 2,5x3m. Also, the room is actually 7.5m tall (this didn't affect the acoustic firm dudgement, as they did go on site to take measurements and did have the blueprints of the room. Simply, person talking to me told me it was 13m, and he was wrong, but he never said that to the acoustic firm).

The glass of the skylight seems to be pretty thin. Maybe there is a lot of low frequency dispersion there?
Another thing that makes me call bullshit on this acoustic firm is that they said that a balloon popping test wasn't reliable enough as it doesn't have enough SPL in the speech frequencies (?!?!?!). I wasn't the one talking to them, but I would have argued that RT60 is a relative measurement: you check the balloon SPL at 125Hz and then check again 60dB later... as long as it is above noise floor, I don't see why the balloon having little low frequency energy would make any difference (and... about speech intelligibility...?!)

[Soundman2020]

Strange situation!

Talking with the people from the theater, they told me that another guy came in, did the balloon popping test and told them their reverb time is around 3 seconds...

That agrees quite well with the data you provided.

I'm not an expert, and this wasn't conducted with the most appropriate gear (AKG 414 ULS, Tascam DR-60D and a balloon) but I think simply using a phone would have told me whether the room was actually closer to the estimated 8" or to the 3" of the last guy.

Again, I'm no expert, but unless I'm screwing something up, the RT60 of the room is DEFINITELY closer to 3 seconds.
I'm attaching 2 files (in mp3... cause the forum won't allow wav or flac)

Definitely!

Please don't take this as being valid, since this was obtained by converting an MP3 (which lacks a huge amount of info) to WAV, then pretending that it is really an impulse response file, and importing it into REW like that. Assuming that this is a reasonably good resemblance of the actual IR for the point you had the mic, this is what the REAL RT-60 graph looks like:

BPOP-2-RT-50-12k.png

And here is what the actual impulse response look like:

BPOP-2-IR.png

As you can see, the real RT60 is about 3 seconds, but there is a longer "tail" around 80 Hz, going out to nearly 6 seconds, with around 4 seconds across the entire low end, as well as a peak in the mid range at 1k. So at a guess, I'd say that room sounds rather "boomy", but not entirely "dull". The high end is not a problem, though. But your estimate is right, as is the estimate of the guy who went there with his balloon. Round about 3 seconds. Getting that down to 1.5 or so across the board would be fine. 1.7 for music maybe, and 1 for speech, as long as most of it is in the early reflections, not the reverberant field (this is very different from what we try to achieve in a control room! Don't get confused: two very different scenarios).

To be more precise, REW judges the decay times at the location where you had the mic, as follows: Topt = 3.025 seconds, T20=3.025 seconds. T30=3.148 seconds. I'd go with T30, and put it at about 3.1 to 3.2 seconds.

Walking in I was expecting a place where it would have been impossible to talk, whereas it is actually much more decent than that. Granted, standing 10 meters from another person made talking hard, but I would guess 8 seconds would make it hard standing 1 meter apart.

Right, but it's not very good: I don't have any means for getting an STI or RASTI index from the data you sent, but I can get C50 and C80:

BPOP2-C50-C80.png

Red is C50, green is C80. Both show the relationship between early reflections and overall energy. C50 is for the first 50ms, and C80 is for 80ms. C50 tells you about speech clarity, C80 is about music clarity. In both cases you want to see those curves ABOVE the 0 dB line, indicating that there is more energy in the early-arriving reflections, and thus better clarity (the early stuff is not overwhelmed by the overall reverberant field). In your case, C50 is pretty bad: -5 to -10 dB, and C80 isn't much better: -5 to -1. This room is no use at all for either music recitals, rehearsal, or speech. It DOES need treatment, for sure, but it definitely isn't as bad as it sounded like from the initial analysis by the "acoustic" company!

What do you think?
I wonder how their math could have been THAT wrong (unless mine is).

The room already has some "treatment" probably, in the sense that the structure itself and the materials are absorbing partially, and the shape of the room, plus the features on the walls could be adding some diffusion (pillars, paneling, columns, balcony, stage, doors, windows, etc). Any drapes, chairs, carpets, or suchlike in there? I suspect there might be... But even so, that still doesn't explain such a huge difference.... And I can't figure out how to get an RT60 of 9 seconds for that size room anyway.... My guess is they screwed up somewhere.

at the center of a half DOME SHAPED ceiling,

Oooops! That's gonna be "fun" to treat! Focusing effects there for sure....

Another thing that makes me call bullshit on this acoustic firm is that they said that a balloon popping test wasn't reliable enough as it doesn't have enough SPL in the speech frequencies (?!?!?!).

Then they don't even understand what an "impulse response is"!!! Wow. It's OK to say that a balloon pop doesn't have enough sound POWER for true RT60 measurements, or even that it doesn't have enough SPL for true RT60 measurements, but saying that it doesn't have enough in a certain frequency is garbage! The concept of a true acoustics impulse is that it is a single "IMPULSE"! It has NO frequency at all. It's just one single peak value with an infinitely short rise time and infinitely short decay time. Theoretically, of course. In practice, it's impossible to generate such a sound. But if you could, and you were to look at it on a digital time-line, all of the values at all times would be zero, except for one single spike where the value is full-scale. That's an impulse. You can't measure the "frequency" of that, because it has none. It does not repeat, therefore it has no period, therefore it has no wavelength, therefore it has no frequency. It's just a pulse. A balloon pop is pretty close to fitting that, although it isn't perfect by any means. You measure the RESPONSE of the room to that impulse, and the response DOES change over time, so there are then frequencies and wavelengths and times that you can measure. I guess you could consider that from this point of view the impulse does "contain" frequencies, but in pure mathematics it does not. The Dirac function is an infinitely short, infinitely loud pulse. Of course, the best approximations are neither infinitely loud nor infinitely short: there is a limit on both, and even if there was a such a signal, it would be impossible to play it on a loud speaker, because a loud speaker takes time for the cone to move. But theoretically, there are no actual "frequencies" present "in" an impulse. What matters is how the room RESPONDS to that impulse

So the mere fact of them saying "a balloon popping ... doesn't have enough SPL in the speech frequencies" is pure ignorance. It's like saying you can't measure the speed of a car by counting how fast the wheels turn, because the wheels aren't yellow enough... !

So, it's true that there isn't enough sound power in a balloon pop to fully create a true impulse, but from that point of view, there is also not enough sound power in a 100,000 watt speaker stack at a major rock concert ,to fully create a true impulse. Because it is impossible to create one! Neither one has enough brute power, and neither one can generate a true acoustic impulse.... And you don't need to anyway. You can simulate an impulse in any of several ways (one of which is a balloon pop), and if there isn't enough sound power to get a true 60 dB decay, then it is also valid to measure the 30 dB decay time, and double it... That's what the "T30" measurement in REW does: it extrapolates what the RT60 decay would be, by just measuring the RT30 decay. Ditto for the "T20" graph. It's not perfectly accurate, to be sure, but it's pretty darn close in most cases.

That said, it's far better to just use REW to measure the Impulse Response, because REW doesn't actually create an impulse to do that, because it is impossible! What REW does is to generate every possible frequency in turn, measure the decay over time, then use clever mathematics to track that backwards in time and create an "impulse" out of that. It simulates the impulse by going backwards from the result, and thus gets a MUCH better, much cleaner picture of the true impulse response, that is not hampered by the limitations of needing to actually generate a true impulse.

That's why the graphs above from REW are not really valid, because they are based on a real-world, low-quality impulse that was actually generated, rather than the much more accurate method that REW uses (sign sweep with FFT integration). But even considering that, it's still probably not far off from reality.

I wasn't the one talking to them, but I would have argued that RT60 is a relative measurement: you check the balloon SPL at 125Hz and then check again 60dB later...

Exactly! That's the theory. And as you correctly noted, there are no frequencies involved. Just brute sound power.

as long as it is above noise floor,

Correct! As it turns out, though, there isn't enough energy in each individual frequency band OF THE ROOM RESPONSE spread out over time, to be able to get accurate results down to -60 dB (which is why REW's method is much better), but there should be enough for T30 and T20.

I don't see why the balloon having little low frequency energy would make any difference (and... about speech intelligibility...?!)

Exactly. You clearly understand this much better than they do...

[saemola]

Thank you Stuart... I think you convinced me to convince them to ditch those guys.
Now I'm gonna give them a call and see what we can do about getting at least part of the €1500 payment back since their project is based on math that is about 200% off the mark...

Thank you!