3 Empirical and numerical estimation of room acoustic properties
This room is mainly used for games, birthday parties and everyday conversations. For this type of multipurpose social room, the acoustic requirements are relatively low. The main goal is to avoid excessive reverberation so that the room does not sound boomy or chaotic. More advanced criteria such as STI or detailed speech clarity indices are less critical here, because people communicate at short distances and can easily adapt their position and speaking level.
I chose this room for the measurement because it has an interesting geometric shape: the ground plan is slightly trapezoid. The ceiling is sloped on two sides and contains many grooves of about 2 cm depth, which act as small diffusing element for very high frequencies. Because of this design, there are only two fully parallel surfaces in the whole room, namely the window and the back wall.
Sound playback in this room:
Reverbration Time
For the hand-clap recording I estimated the reverberation time from the decay of the sound level in dB. (measured in the middle area of the room) In the impulse response, the level drops from –30 dB to –60 dB over a time interval of Δt₃₀ ≈ 0.25 s, as shown in the figure.
Assuming a roughly linear decay in this semi-logarithmic representation, this corresponds to a full 60 dB decay time of RT₆₀ ≈ 2 · Δt₃₀ ≈ 0.50 s.
The smooth, almost straight decay without strong fluctuations indicates that the reverberation is diffuse and well behaved, with no dominant late reflections imposed on the RT curve.
Listener locations in this room
When having a conversation in the middle of the room, I would expect very good clarity: the first reflection from the dome-like ceiling almost certainly reaches the listener’s ears very early, reinforcing the direct sound instead of masking it. The delay of this first reflection would only be a few millisecons and therefore perciefed as one sound source.
Closer to the side walls I would also expect good clarity, since reflections from the nearby wall dominate and the contribution from the ceiling becomes less pronounced, again avoiding discrete echoes while maintaining a pleasant, diffuse reverberant field.
Because most of the walls are not parallel, no strong back-and-forth reflections build up, so you don’t hear distinct echoes, but rather a short, warm reverberation coming from the mostly untreated wall surfaces.
This impression fits well with the measured reverberation time of about RT₆₀ ≈ 0.50 s, which is short enough to keep speech intelligible while still giving the room some acoustic “life.”
RT simulation at 500Hz and 1000Hz
I tried to use Sengpiel’s online calculator with the room dimensions and surface materials and did press the “calculate” button, but for some reason the page never returned a meaningful RT value – the output field simply stayed at “0 s”.
I do not know exactly why it did not work; most likely it is a technical issue (browser/JavaScript compatibility or some hidden input problem) rather than an error in the RT formula itself.
On top of that, the calculator assumes a simple rectangular room with idealised, frequency-independent absorption coefficients, which does not really match the trapezoid shape and diffusing, dome-like ceiling of my room.
For these reasons I decided not to rely on the web tool and instead determined the reverberation time directly from the measured impulse response, which automatically includes all the real geometric and surface effects of the room.
RT60 at 500Hz:
RT60 at 1000Hz:
Unfortunately, from the practical measurements at 500 Hz and 1000 Hz I could not determine a clear RT₆₀.
In the dB-versus-time plot the decay is not a straight line. I suspect that the playback signal did not stop abruptly but faded out, so the decay of the loudspeaker is mixed with the room response and the RT estimate becomes unreliable.
Final Assignment: HIL CLUB
Michael Lory 04.12.2025
1. Acoustic Characteristics of this Room as it is:
The room D24.1 (HIL) is a big rectangular meeting- and workspace of about
13.5 × 9.7 × 4 m. When you walk in, the first impression is dominated by the high ceiling and the large glass facade. The walls feel big, flat and sterile. Beside the Humans, there is almost nothing in the room that could absorb or scatter sound.
On closer inspection, I noticed two features that are acoustically beneficial and worth mentioning:
Industrial features on ceiling
The open grid with ducts and beams breaks up the large flat ceiling, so reflections in the mid and high frequencies (most effectivive for frequencies >= 500Hz) get strongly scattered instead of coming back as one mirror reflection. For mid-frequencies, it also acts as an absorber since soundwaves get partially trapped inside.
Porous / textured wall
This micro-bubble wall surface acts a bit like a shallow porous absorber, so it softens reflections in high frequencies a bit. (about 1–8 kHz).
Reverbration Time
When talking, voices sounded a bit echoey and harsh. As soon as we were allowed to do a quick hand-clap reverberation-time measurement, this impression was confirmed. The clap left a whirring, metallic echo hanging in the room, so instead of one clean decay you could clearly hear a bright tail of reflections bouncing back and forth after the initial impact. This effect was strongest at 2500Hz (probeably beacuse my clap excited the Room at 2500Hz the most) The large glass facade, the big parallel walls and the smooth concrete floor all contribute to this effect. Together they act like hard mirrors for sound, so the energy keeps bouncing back and forth and builds up this long, ringing echo.
The reverberation time that I measured with the hand clap actually feels too short. It is clearly shorter than what I heard with my own ears, and also shorter than the values in the official acoustic measurements. A T30 of 0.68s would be a good value for an office like this! Sadly I think this has a lot to do with the situation in which I measured: there were people sitting around, there were coats, office chairs and stored material. Especially in the lower part of the room, below about 1 m, where I placed the phone microphone, there was much more acoustic absorption than at head height, where speech and most sound energy actually matter. So my measurement mainly captured a decay that is already damped by furniture and legs and jackets, while the upper part of the room is still much more reflective.
The simulated data for this room shows a much longer decay: T30 ≈ 1.6 s in almost all frequency bands. In my opinion these simulations match the perceived reality quite well, even though they were almost certainly done for a rather empty room and with a lot of simplifications.
SIA Standarts:
For a room of about 500–600 m³ used for meetings and speech, the SIA guidelines aim for a mid-frequency reverberation time of roughly 0.6–0.8 s.
In our case, both the simulations and the provided measurements show T30 ≈ 1.5–1.7 so the room is clearly outside the recommended range and therefore does not meet the SIA standard for its current use.
Room analysis conclusion:
So the reality of the acoustic properties of this room is likely a bit better than what the empty-room T30 suggests, but the core problem remains the same: even with occupants the room is still too reverberant for clean speech, and the combination of glass façade, parallel hard walls and concrete floor produces the characteristic whirring echo that I observed in both listening and measurement.
2. Alternative use: HIL CLUB
I propose to transform this room into a club. Not because it is particularly well suited for it, but because dance-music clubs usually grow exactly out of spaces like this. Considering today’s “club and cultural space death”, During the day this Room can also be used for exhibitions, creative workshops and a place to meet new people and exchange ideas. I think creating cultural spaces would be a good thing and a very plausible scenario.
To be able to play dynamic music in this room, we need acoustic treatment in every frequency band. Clubs where artists can show their musical skills place a relative high demand on a room acoustically. When taking into consideration that clubs on top of the already problematic high frequencies, we now add strong bass frequencies below 100 Hz, which we could not produce with a humans voice.
Schroeder frequency
Below the Schroeder frequency fs, the room is dominated by discrete room modes: sound behaves like standing waves, with strong peaks and nulls that depend a lot on position, especially in the bass. Above fs, the sound field becomes more diffuse and statistical, so individual modes overlap and we can describe the room with global parameters like reverberation time and treat it effectively with broadband absorbers and diffusers.
Acoustic treatment for frequencies higher than 110 Hz:
Acoustic curtain: High frequencies treatment
For the high frequencies I decided to cover two of the walls with acoustic curtains, especially in front of the glass façade and the wall on the opposite side of the DJ booth. These surfaces currently act like strong reflectors, so the curtains turn them into porous absorbers that mainly work in the mid and high range (roughly from 500 Hz upwards). A nice side effect is that curtains are flexible: they can be opened or closed depending on whether the room is used for daylight activities like an art exhibition or workshop, or loud club events. In this way they help to reduce the harsh, bright echo without permanently loosing the Rooms usecase flexibility.
This is an Image of various acoustic curtains and their absorption coefficients. We see that choosing the right one installing it couteract the room’s echo problem in all high frequency bands a lot.
Acoustic Panels on walls: Mid frequencies treatment
For the mid and low-mid frequencies I want to mount rockwool-filled acoustic panels directly on the walls. With a thickness of 30 cm, these porous absorbers work mainly in the roughly 200–1000 Hz range, where a lot of the energy of voices and instruments sits, and they help to bring the reverberation time in this band down. At very low frequencies they will not fully absorb the standing waves, but their deep surface and framing still slightly diffuse and soften the reflections compared to a flat hard wall.
The Graph in green indicates an acoustic panel filling with rockwool and a thickness of 30cm. Down to our Schroeder frequency, these pannels will work as a great absorber with absorption coefficients >= 0.8. These Panels will mainly be on the Wall behind the DJ and on the big wall parallel to the Window. Since the ceiling has a great scattering coefficient but only moderate absorption, we will also place some into the suspended ceiling, while trying not to looe its Industrial look.
Acoustic treatment for frequencies lower than 110 Hz:
Basstraps: Low Frequency resonance treatment
When treating the low frequencies, we have to focus much more on the room geometry and its resonances as a whole. Parallel walls and simple rectangular rooms have a high potential for strong standing waves, so we try to break this parallelism. One option is to add 45° angled elements in at least two corners of the room.
Standing waves simulation with cornered edges
We see that Resonances in this Room are less dominated my the dimension in which the Basstraps are facing. We see strong modes in the rooms width but only weak ones in its length. Of course this picture will also change with given impulse frequencies but its clear that there are still strong room resonances which need to be treated.
Estimating Modes and problematic low resonant frequencies
Like we saw in the Basstraps graph, we still have a lot of resonances in this room. We could say: “leave this challenge to the sound-engineer who installs this soundsystem.” But I want to propose a semi flexible damping method for specific frequencies.
First, lets have a look at which frequencies might be problematic. This can be calculated with the following Formula:
Our Room:
Lengh: Lx = 13.55m
Width: Ly = 9.73m
height: Lz = 4.0m (estimation)
Where our n-th Mode in the corresponding direction is noted as: (nx, ny, nz)
Since finding the frequencies at which multiple dimensions’ modes are active is mathematically difficult to calculate, ChatGPT helped:
Helmholz resonators: specific low frequency band treatment
These Frquency Bands can be treated wit Helmholtz-resonators. Classically you would first measure the Room an then measure your problematic frequencies and design an Helmholz-resonator afterwards. Since I want to propose a acoustically treated an finished room, we will install them, but leave some finetuning options.
This is one of the three Helmholtz resonators I want to place in the room. It is shaped like a sofa, with soft cushioning on top. The box can be opened so that additional backplates (indicated in red and green) can be inserted to reduce the internal volume and raise the resonance frequency. This allows fine-tuning after everything is built, and also means the resonators can be retuned later if the room is renovated or new elements change its characteristics.
Since we can only correct or fine-tune the frequncy upwards. I will round down the Values ChatGPT gave me.
Cluser 1: => 35Hz
Cluster 2: => 50Hz
Cluster 3: => 85Hz
Example Helmholzresonator Sofa:
Since in Electronic Dance Music, the frequencyband which transmits the most energy sits around 50-60Hz, I will only provide Example Values for a Helmholzresonator tuned to 50Hz (targetting Cluster 2).
Putting in a 10cm thick Back Insert would then for example lift te resonance frequency to 54.8 Hz.
3. Final Setup and Conclusion
Together, the acoustic curtains, the 30 cm rockwool wall panels, the additional ceiling absorption and the tuned Helmholtz resonators significantly reduce the mid- and high-frequency reverberation, smooth out the low-frequency room modes, and improve clarity. In this configuration the room should come close to the SIA target range of about 0.6–0.8 s (T30) in the important mid bands, while still keeping enough liveliness for club music. Otimally we would have now a C80 of about +3dB to +4dB which would mean the rooms reverb is quite low, and sharp precise sounds can be played and percieved.
Hopefully we can now play music that is perceived fairly linear across all frequencies. And if it is still not perfect, we are probably not doing much worse than most clubs in Zurich anyway. Low frequencies will always cause problems, but I tried to mitigate them as much as possible.
