modeling a GAN for audio generation

First thing: I am not an expert on Machine Learning, I'm learning it on my own. so I hope I'm asking the right questions.

I am trying to model a GAN, with mlpack library in C++, to try to generate audio data from a dataset audio files.

mlpack has this class:

GAN (   arma::mat &     trainData,
    Model   generator,
    Model   discriminator,
    InitializationRuleType &    initializeRule,
    Noise &     noiseFunction,
    const size_t    noiseDim,
    const size_t    batchSize,
    const size_t    generatorUpdateStep,
    const size_t    preTrainSize,
    const double    multiplier,
    const double    clippingParameter = 0.01,
    const double    lambda = 10.0 
)   

I found a lot of samples for image generation (so starting from matrixes),but nothing for audio generation (vectors). So my questions are:

• What could be good models for generator and discriminator?
• Is there anything else that needs to be known since we are starting from audio data and not images?

(Any links would be helpful)