Electronic Arts Inc.

Sergi Andreu

Monica Villanueva Aylagas

Albin Jansson

This specification relates to generating variations of in-game sound effects using machine-learned models. According to a first aspect of this specification, there is described a computer implemented method of training a machine-learned generative model to generate sound effect variations. The method comprises: for each of a plurality of training examples in a set of training examples, each training example comprising a waveform of a sound effect: generating a low dimensional representation of the waveform of the sound effect; inputting the waveform of the sound effect and the low-dimensional representation of the waveform of the sound effect into the generative model; processing, by the generative model, the input waveform of the sound effect and the low-dimensional representation of the waveform of the sound effect to generate a sample from an output distribution; and updating parameters of the generative model using an objective function based on a distribution of the input samples.

Growing community of inventors

Barcelona, Spain

Sergi Andreu