Ambisonics and Reverb


The sounds in this part of the composition were made in a Max/MSP patch. The sound sources were created using algorithmic reverb – Mitchell Turner’s rev3~ extension. These sources moved through an ambisonic virtual spherical space. The positions of these sources were then encoded and decoded for eight channels using ICST Ambisonic Tools. In this section, the main goal was to avoid the presence of a sweet spot.

In usual setups, the ambisonic technique is expected to be perceived from a sweet spot. All the loudspeakers contribute to achieving correct positioning of the sound for that listening area. Using ambisonics allowed for the sound to move through a space that was warped by the physical placement of the loudspeakers. Additionally, the distribution tool also allows a relationship between the different reverbs to be established. The environment behaved according to unsteady rules. The relations between the channels were present and, at the same time, absent from the physical space. This duality provided a sense to the events, but without localizing the sources completely.

Changing the virtual position of the reverb caused feedback in the virtual space before encoding (see fig. 1). This feedback added tension and a certain level of unpredictability to the sound events. Also, these relations between reverbs were adjusted algorithmically by changing the elevation and the distance from the hypothetical sweet spot and the azimuth, which produced the movement between the loudspeakers.

Fig. 2.1 – Visual representation of the hemispherical ambisonic encoding in Max/MSP; the dots in the left image of the hemisphere represent three different reverb sources

In the final version of the patch, three reverb sources were placed in the virtual space. The virtual loudspeakers were positioned in the circle with no elevation within that space. The phase and amplitude of those virtual loudspeakers was manipulated using Ambisonic Tools. The position of each source was defined by the azimuth, the elevation and the distance in relation to the virtual sweet spot. The first reverb source was slowly circling between the loudspeakers while changing its distance from the sweet spot. The second reverb source changed its distance and elevation in time, but its azimuth was kept the same. One last long reverb source was then placed near the top of the virtual hemisphere, in order to be very subtle and almost non-present in the mix. The first fed the second, the second fed the first and also the third.

Fig. 2.2 – Diagram of relations between reverbs moving through the ambisonic virtual space

Using reverbs in virtual space allows for long, nearly infinite reverberation and feedback to color the sound between the virtual loudspeakers. Besides, using wet sounds in this setup became less significant, as any sound with a short envelope could be used as a dry source for one of the other reverbs. This is mainly due to pitch shifting, which was implemented before this reverb mechanism. The texture thus results from the pitch shifting, rather than depending on the texture of the original dry sound. The focus on long reverberations and feedbacks among the sources led to the creation of a patch in which any slight adjustment could break the balance established in the mechanism. The final result is sonically calm, and yet the system behind it is very fragile.

Sound 2.1 – Ambisonics and Reverb excerpt