Training generative adversarial networks to upsample audio

A computing device comprising:

• a memory that includes (i) instructions for training models to produce magnitude spectrograms at sampling rates of interest and (ii) a database of models, each of which has been trained to produce magnitude spectrograms at a different sampling rate,
• wherein the instructions, when executed by a processor, cause the computing device to:
• receive input indicative of a selection of a given sampling rate,
• determine that none of the models included in the database are associated with the given sampling rate,
• obtain multiple magnitude spectrograms that are associated with the given sampling rate,
• providing the multiple magnitude spectrograms to an untrained model as training data so as to produce a trained model, and
• store the trained model in the database.

A non-transitory computer-readable medium with instructions stored thereon that, when executed by a processor of a computing device, cause the computing device to perform operations comprising:

• obtaining a series of magnitude spectrograms that are associated with a given sampling rate;
• training a neural network-based model to convert a first magnitude spectrogram provided as input to a second magnitude spectrogram that is associated with the given sampling rate by providing the series of magnitude spectrograms to the neural network-based model as training data;
• and
• storing the neural network-based model in a database that includes a series of neural network-based models associated with different sampling rates.