The SoDA project proposes the design and implementation of a sound synthesis engine in real-time for sound design, interfaced through semantic analysis with sound libraries.

• - operability either in real- and non-real-time
• - management of a generic multichannel audio delivery
• - integration of static and interactive audio
• - communication with other software applications

SoDA allows to retrieve audio material from a library, to spatialize it, and to make the sound available to the user in the requested format. The term "spatialization" here and below indicates not only the definition of the sound source position in space but more generally the application of digital signal processing algorithms capable of simulating the overall influence of the diffusion of the sound in space, such as, filtering, reverberation etc.

At the base of SoDA there is the concept of "sound object".
A sound object represents a certain sound material to which a position in a reference space is added (the "scene").

That is, the hypothetical source responsible for the sound object explicitly receives a position in the scene. The scene also includes a listening position (the "listener"), which is the audio equivalent of the point of view of the frame.

SoDA receives in input sound objects: it retrieves the audio material from the library and uses spatial data (position of the objects and of the listener) to output a spatialized version of the same audio material, applying a "space model", that is, a set of processing algorithms that represent the contribution of space to the sound.

SoDA includes a set of sound space models that the user can choose and and configure.

SoDA provides four "working modes" for the sound designer, that can be placed over a seamless continuum real-time audio engine. SoDA can be controlled by an external application such as, for example, a virtual reality engine, in relation to which it works as an audio module; The other working modes can be defined on top of this low-level mode, and couple signal processing with aspects of semantic analysis, a solution that is currently not provided by any other application.

Semantic-aided sound design
SoDA includes a software module for the semantic analysis of the linguistic input by the user; analyzing the input, SoDA queries the connected audio system libraries and returns a collection of possible materials.

Soundscapes generation
SoDA allows the sound designer to enter keywords from which the software selects a set of sound materials; then, the software determines a set of possible placements of the same space and defines a listener. Finally it renders audio and provides in output the audio file in the required format (mono, stereo, 5.1 etc.).

Soundscape delivery
The user specifies via web the semantic constraints that characterize the scene s/he is working on to the service provider; SoDA, once installed on the server side, generates the required soundscape; finally, the server makes available to the client the resulting audio file in the requested multichannel format. SoDa is proposed as a tool for rapid prototyping, making the entire audio workflow more agile, connecting pre- to post-production. To ensure the functionality of SoDA in terms both of sound quality and usability for the sound designer, the design and implementation of the application will be verified at each step according to the requirements actually indicated by industry experts.

SoDa Architecture

Icmc Conference 2014

SoDa Sound Scape Generator

If you have a project you would like to discuss, get in touch with us.