Video is also separated into voice with video and non-voice video. The separated data can be saved in the playlist in the app. You can also export outside the app. DeepDenoise developers do not collect any voice, audio data and video data. The neural network uses the model developed by PsycheTone Inc. We do not reuse the trained models of third parties. *The sound sources that can be read from the Music app are songs purchased from the iTunes Store. Or only music that is not protected by DRM (Digital Rights Management).