SIGNAL TRANSFORMATION BASED ON UNIQUE KEY-BASED NETWORK GUIDANCE AND CONDITIONING

A method comprises receiving input audio and target audio having a target audio characteristic. The method includes estimating key parameters that represent the target audio characteristic based on one or more of the target audio and the input audio. The method further comprises configuring a neural network, trained to be configured by the key parameters, with the key parameters to cause the neural network to perform a signal transformation of the input audio, to produce output audio having an output audio characteristic corresponding to and that matches the target audio characteristic.