Unfortunately, Roll doesn't talk much. Therefore, I had to use Tortoise-TTS, to generate extra speech for Roll. I am NOT going to sit through hours of footage searching for ONE characters voice. If you would like a better model, you go on ahead, and cut the sections of Roll's voice using audacity, or any other audio editing software, and upload the .wav file here. I will remove the background noise, enhance the audio, and train the model all myself, but I will NOT spend hours looking for Roll's voice. I need at least 6 minutes of Roll's speech to train the model.
