SOUND GENERATOR AND AUDIO SIGNAL GENERATION METHODS

Abstract

The invention relates to a sound generator and a method for generating sound signals. The sound generator (10) can generate sound signals (16) from input signals (14) and target data (12) representing sound signals (16). The target data (12) is inferred from text. The sound generator consists of: first processing block (40, 50, 50a-50h), which receives first data (15, 59a) inferred from input signals (14) and outputs first data (69); second processing block (45), which receives, such as second data, first output data (69) or data inferred from first output data (69). The most interesting processing smoke (50) includes: a tuning set of learnable classes (71,72, 73) configured to process the target data (12) to obtain tuning characteristic parameters (74, 75); and a type element (77), configured to apply the tuning characteristic parameters (74, 75) to the first data (15, 59a) or the normalized first data (59, 76').