REPRODUCING REVERBERANT SOUND

The reverberant sound field component is effectively “what is left over” after the direct sound and the early reflection components have been considered. It is effectively pure echo.

 

The reverberant sound field component contains contributions from all sound sources smeared over time and with no directional information – it has been destroyed.

 

The reverberant sound field is the result of many reflections from objects and surfaces in the original listening environment providing the overall acoustic behaviour or response of the environment and giving the impression of the capture venue size and nature. This is a desirable effect as it provides a spaciousness to the sound. It is thus often desired to artificially recreate the effect. We now consider how this can be done.

 

Reverberation is characterized by associated decay time constants and spectral acoustic response colouring of the acoustic environment over time. To reproduce this characteristic, all contributing sound sources need to be included with the decay and spectral tailoring characteristics applied.

 

Then the phase consistency of the wave-fronts needs to be destroyed. As there is no coherent distance or direction information in the reverberant sound field.

 

This characteristic can be mimicked by the use of electrical and acoustic treatments that essentially provide a myriad of multiple delayed sound source replicas throughout all frequencies of interest and with the desired overall spectral colouration of the sound.

 

There are three basic techniques for doing this.

1. The signals can be passed through a loudspeaker/microphone combination in a dedicated reverberant room. A variation on this is to pass acoustic waves through solid structures such as springs or plates with acoustic behaviours mimicking the multiple reflections of the reverberant sound field. These approaches suffer from physical impracticality and poor fidelity.
2. The signals can be electronically treated before reproduction, “jumbling” them by use of amplitude and phase treatment including adding multiple delayed, filtered and decaying versions of the original signal. These signals are then fed to one or more conventional loudspeakers in the listening environment. One weakness of this technique is that the loudspeakers being used will tend to disclose their locations as the listener moves because of their individual directional characteristics. This approach can be extended with further electronic processing and multiple loudspeakers to render multiple de-correlated sound images in space. This helps mask the speaker location, but the ability for the listener to locate the loudspeakers is still an issue.
3. The actual acoustic output of the loudspeaker reproducing the reverberant field can be treated. This approach uses acoustic filter structures that effectively “phase shuffle” and spatially control the harmonic related phase information of the source sound. A significant advantage of this approach is that no electronic de-correlation signal processing is needed. The sum of all source signals can be fed directly to the speakers.

Early reverberation mimicking attempts used the diffusive reflection structures of Schroeder and others that relied on stagger tuned quarter wave overtone reflectors to “de-correlate” the waves. These structures are used to control specular reflections from surfaces in rooms but have not been widely used to construct reverberant field loudspeakers. This has in part been because the structures do not lend themselves to compaction and in part because it is difficult to design a reflective structure that treats all the coherent output without some leakage.

 

HuonLabs has developed reverberant field loudspeakers specifically to overcome these limitations. These loudspeakers use acoustic half-wave leaky transmissive/ reflective structures to process and de-correlate their acoustic output. An example is shown in Figure 1. These patented devices offer significant advantages, as quite compact units can be constructed that are very hard to acoustically locate but still achieve very good control over the de-correlation and spectral and polar radiation aspects. The term Whiteroom was coined to describe these devices because of the visual analogy of their acoustic operation.

 

Whatever method of de-correlation is used, the spectral decay characteristics of the recording environment also need to be considered.

 

Active acoustic masking

Reverberant or de-correlated sound can also be useful for actively masking listening environment acoustics.

 

The boundaries of the listening room reflect sound. These reflections are largely specular and provide clues to the general characteristics of the room and the location of the boundaries. The problem with boundary reflections is particularly evident when the listener is free to move about in the listening area. This listening environment reverberant behaviour will be in conflict with the desired recording environment behaviour and if untreated can be quite intrusive.

 

Whiteroom acts as a diffusive source. The specular reflection of a diffusive source remains diffusive. When the listening environment is “illuminated” with diffusive sound having the characteristics of the recording environment any specular boundary reflections can then be masked. The room is acoustically painted “white” with the desired sound source characteristics. This is not the same as using white noise to mask undesired sounds. Active acoustic masking is more useful than white noise because the locations of boundaries and surfaces are masked with the reverberant sound of the source and so full source intelligibility is maintained and used to full advantage and only exactly when needed.

 

The reverberant sound field can thus be correctly reproduced in any listening environment with the bonus of actively masking undesired reverberant listening environment behaviour. The simplest method is to place a Whiteroom loudspeaker somewhere in the reproduction room and feed it with the sum of all channels contributing to the original sound. No other electronic processing is required.

 

The information contained herein is copyright to HuonLabs. No material can be reproduced in its totality or in part or without the express permission of HuonLabs Pty Ltd. Any reference to this material must quote the HuonLabs source. Trade marks and Patent applications apply to most aspects of the work disclosed here. Contact HuonLabs for further details, product information or licensing enquiries.