Lately i started working a little bit of audio, so i needed something like torchaudio but for NX, i haven’t found anything like that, so i started building NxAudio.
Right now, its a pretty simple library, using ffmpeg as a underlying solution, you are able to read several audio types and work with them as NX tensors, generate spectrograms and visualize them.
Soon i’ll implement audio augmentation, some more transformations and maybe pre-implement some models like tacotron and hifigan purely on NX.
