PHYS341/2024/Project9

From UBC Wiki

Building Acoustics: Reverberation

Reverberation is reflected sound waves that are perceived as separate or distinct from the original (direct) sound. It occurs when sound is reflected off of surrounding surfaces at varying rates in time. Even after the original sound has stopped, the reverberations will continue to be heard as they reflect off surrounding objects in the environment. This continued (reverberated) sound is temporary as the reflected waves decay until they are imperceivable.

The quality (perceivability) and intensity (loudness) of reverberation is dependent on the architecture / geometry, materials, air temperature, etc. of the environment in which the sound is produced. Too much reverberation (too little time in between reflection waves) interferes with speech/sound intelligibility.

Frequencies

Prior to further discussion, it is important to understand how specific frequencies ranges will affect the reverberation quality. Frequencies being high, mid, or low play into the perceived quality of reverberation: lower frequencies produce a perceptually warmer atmospheric quality; higher frequencies produce a perceptually cooler atmospheric quality.

A study done by Traer & McDermott (2016) found that larger spaces with more reflective walls (hard & smooth materials) produced greater reverberation. Traer & McDermott (2016) also noted that the longest decay rates were within the Mid Frequencies (200 - 2000 Hz). Rougher or more porous materials will absorb frequencies around the mid - high range. [1]

  • Low Frequencies: 20 - 200 Hz. (average f0 of adult speech; non sibilant)
  • Mid Frequencies: 200 - 2000 Hz. (sibilants)
  • High Frequencies: 2000 - 20,000 Hz.

*Sibilants are speech sounds like 's', 'z', 'sh', 'f'... they are characterised by more turbulent airflow and thus have higher frequency values.

Frequency Chart [2]
Summary Range Values
Low Frequencies Sub-Bass 20 - 60 Hz.
Bass 60 - 200 Hz.
Mid Frequencies Low Midrange 200 - 500 Hz.
Midrange 500 - 2000 Hz.
High Frequencies Upper Midrange 2000 - 4000 Hz.
Presence 4000 - 6000 Hz.
Brilliance 6000 - 20,000 Hz.

Reverberation in Practice

Solids are impermeable materials like marble, steel, smooth stone, cement, and some plasters. Porous materials are more absorptive; fabrics, carpet, wood.

Reverberation is often considered in building/architectural acoustics. When designing a space that will be primarily used for speech and/or music one must consider how sound will interact with the structure. Acoustic spaces can either enhance or degrade a sound's intelligibility. The waves reflect off of each surface while decaying; some energy is absorbed by a surface while most is reflected. For smooth, hard surfaces almost all of the energy is reflected.

Spaces that are made with reflective materials will have a fuller sound that gives the building life; more absorptive materials dampen excess noise and reflections, making a room sound dry and silent.

In cathedrals and churches, reverberation creates 'spiritual' ambience characterised by fullness and a lasting linger, allowing for a feeling of life in these places of worship. Thus, some music has been designed to sound best in reverberant spaces, like churches and cathedrals. Contrarily, in venues that are intended for rock or pop music, most reverberation will interfere with the listener and performer's perception due to the tempo and loudness of these music genres. Here, shorter decay times are better suited for optimal sound intelligibility: a medium reverberation tail is desired in order to prevent the sounds from blending and becoming unintelligible.

Four Acoustic Features

Simple 2D Model of Reflected Sound Energy. Sound energy from a direct source is reflected off of surrounding boundaries.

There are four acoustic features considered in a room's acoustic properties, reverberation is one of these features. Over time, the energy from the direct sound and the reflected waves decreases as it is absorbed into surrounding surfaces.

Direct Sound: the original sound waves emanating directly from the source. These are the first sound waves that reach a listener's ear, allowing localisation of the direct sound within the auditory system.

Early Reflection: sounds that arrive at the external ear after approximately 1/10th of a second after the direct sound; perceived as being within the direct sound. These early reflections allow an acoustic space to perceptually "hold" the sound, thus giving it a certain quality of life and enhancing the clarity and strength of the direct sound.

Reverberation: reflected sound waves that are perceived as separate to the direct sound. Reverberation is also referred to as late reflection and can interfere with sound intelligibility depending on the time between reflections. More or less overall reverberation is dependent on the absorption properties of the surrounding environmental surfaces and air; a larger space with more reflective walls will see higher reverberation.

Echo: discrete sound reflections that are perceived as distinct acoustic events. Echos interfere with a sound's intelligibility and are more common in larger spaces as they have a longer reflection tail.

  • Slap echo: single sound return.[3]
  • Flutter echo: rapid series of decaying echoes. [3]

Measuring Reverberation

When measuring reverberation in a given environment, a few considerations must be taken...

Frequency Balance: the strength of the frequencies involved in reverberation.

Intensity: this is the amount of energy in a sound wave. If the intensity of a reflected sound is greater than the direct sound, the brain will perceive it as unique to its source (direct sound) and it will interfere with the intelligibility of the direct sound. This is measured in decibels (dB).

Decay: how long it takes for the reflected sound (reverberation) to fade after the direct sound stops. Larger rooms produce fewer reflections per unit time, thus reverberation is slower to decay and has a longer tail.

  • Lower frequencies give a ‘warm' reverberation tail.
  • Higher frequencies give a ‘bright' reverberation tail.

D50: ratio of received early (0-50ms post arrival of direct sound) sound energy to total received energy.

  • Early reflection is considered to have positive effects for sound intelligibility (0-50ms).
  • Late (delayed) reflection is considered detrimental to sound intelligibility (> 50ms).

Reverberation Time (RT, RT60).

Used to predict the sound that any source would produce in x environment and thus describe the quality of reverberation.

RT is a measurement of how quickly a sound decays to inaudibility in a given environment.

RT60 is the rate at which amplitude (sound volume) decreases 60dB from the direct sound in a given environment.

Impulse Response (IR)

The IR is a short sound burst (impulse) that measures a sound's decay rate (RT) in a given environment. Simply put, it is a brief signal that measures how a space responds to sound; defining the acoustic characteristics of a room or building. Thus, in order to calculate RT and RT60, one must measure the Impulse Response (IR) of a room: “energy is absorbed by environmental surfaces with each reflection (as well as by air), longer paths produce lower amplitudes, and the overlapping echoes produce a “tail” in the IR that decays with time.” [1]

Not only does IR help measure the RT and RT60 of a room or building, it also gives the listener an indication of where the direct sound is coming from. This distance is determined by the ratio of direct to reverberant sound received by the listener. Smaller rooms need a shorter RT60 of 30ms for optimal listening; larger rooms require around 40 - 60 ms of RT60. These values increase when dealing with lower frequencies. [4]

Experiment

Recording setup: impulse was recorded in a mostly cement room approximately 4x4m large in the UBC Rose Garden Parkade. Laptop was positioned in the centre of the room. I stood behind the laptop and clapped once.

Methods

To demonstrate reverberation and the four acoustic features discussed earlier, I conducted my own reverberation test using recording software, Audacity, on my laptop. The sampling rate was set to Mono, 384000 Hz, 32 bit float. I recorded a clap burst by clapping my hands together once in an elevator room about 4x4m large in UBC Rose Garden parkade. I positioned myself at one end of the room and placed my laptop on the ground in the middle of the floor, facing the wall in the photo. The total sound time was measured in the recording as t = 0.547s.

Using Praat, another sound measuring tool, I measured the intensity values for RT60 using a spectral slice that shows a 60dB intensity range from the selected 0.547s frame.

Note: I had intended to record in the actual parkade which is a much larger space, but background noise (loud hum) made it difficult to distinguish the reverberation tail. The noise was still evident in the elevator room but at a low enough level so the reflections were perceivable during recording and spectral analysis.



Considerations

Simple 2D Model of Reflected Sound Energy in recording environment. The hand symbol, centre, shows the position of the direct sound source (clap). Direct sound waves are in black; reflected waves are in pink.

I chose the highest possible sampling rate on Audacity (384000 Hz) as the default, 44100 Hz, did not have enough fidelity to produce my desired results. The 384000 Hz picked up enough distinct peaks (unique reflections) so I proceeded with that sampling rate.

Further considerations possibly affecting my results: the left wall (in the photo) was made of mostly glass which is not as reflective as the surrounding cement. The reflections may not have been as strong due to this. I also stood behind my laptop to produce the sound impulse (clap), this means sound would have reflected off me as well as the surrounding walls; I would absorb more sound than a cement wall since fabric (clothing) absorbs more than it does reflect.

Given the very small size of the room I recorded in, I expect to see shorter reverberation times as there are more opportunities for the reflected sound to transfer energy to the surrounding walls. Thus, the reverberations will decay faster and produce a shorter reverberation tail.


Results

Three reflections (white) and the reverberation tail (pink) following the initial burst (black).

I was able to record a decent example of reverberation using my laptop microphone (MacBook Air) and a decently reverberant room. The entire event, from direct sound impulse to reverberation tail measures 0.547s.

The first image on the right can be segmented into 3/4 acoustic features:

The first segment, 'clap', is the direct sound.

The second segment, 'reflections', show two three reflections which I could not perceive as being separate to the direct sound but instead sounded like 'added body' to the direct sound.

Spectral slice demonstrating three reflections and a reverberation tail following a direct sound impulse.

The third segment, reverberation tail, lasts for approximately 0.344s. Considering the small size of my recording space (~4x4m), a shorter reverberation tail is produced since the walls are closer together so there is less time between reflections and thus more opportunities for the reflected waves to lose energy.

The intensity range is 60 dB (50 dB to -10 dB) in this frame. In the spectrum graph of the impulse to reverberation, the maximum recorded intensity = 88.4 dB and the minimum intensity = 60.61 dB.

Note: I assume the negative intensity measurement (-10 dB) is from the background noise (a loud humming) in the whole parkade. The intensity is shown as -10dB because the software assumes it is below the threshold of human hearing. I believe the background noise may have caused it to adjust the threshold level. Thus, the reverberation tail, which was audible, is recorded as being -10dB lower than the background noise. In future analysis I would record in a quiet room that the software can better work with.

The sampled frame here is the same as before, t = 0.547s, this means the impulse in this room has a decay time of RT60(50dB to -10dB) = 0.547s.


References

  1. 1.0 1.1 Traer, J; McDermott, J. H (2016). "Statistics of natural reverberation enable perceptual separation of sound and space". Proceedings of the National Academy of Sciences of the United States of America – via PubMed.
  2. Young, Cory (14/08/2023). "AUDIO FREQUENCY RANGE EXPLAINED". gear4music.com. Retrieved April 2, 2024. Check date values in: |date= (help)
  3. 3.0 3.1 Borgerson, B. (09/10/2021). "Church Acoustics: Four Must-Know Basic Concepts". Churchproduction.com. Retrieved March 5, 2024. Check date values in: |date= (help)
  4. "RT-60 analysis". Acoustic Sciences Corporation. Retrieved March 30, 2024.
  1. ASC (Acoustic Sciences Corporation). (2022, October 31). RT-60 analysis. Acoustic Sciences Corporation. https://www.acousticsciences.com/product/rt-60-analysis/#:~:text=For%20domestic%20listening%20rooms%20and,is%200.4%20to%200.6%20s. Accessed: March 30, 2024.
  2. Reverberation. AkuTEK. (2019, December 8). https://www.akutek.info/research_files/reverberation.htm Accessed: March 30, 2024.
  3. TKACZYK, V. (2015). The Shot Is Fired Unheard: Sigmund Exner and the Physiology of Reverberation. Grey Room, 60, 66–81. http://www.jstor.org/stable/43832231 Accessed: March 5, 2024.
  4. Acoustics First Corporation. (2016, June 13). Reverberation, the invisible architecture. Acoustics First Blog. https://acousticsfirst.info/2016/06/13/reverberation-the-invisible-architecture/ Accessed: March 5, 2024.
  5. Reverberation. (n.d.). https://www.sfu.ca/~gotfrit/ZAP_Sept.3_99/r/reverberation.html  Accessed: April 2, 2024.