Embedded Files
This website provides examples of our watermarking scheme. Our watermarks can transfer to text-to-speech models (e.g., YourTTS and PlayHT) when they apply speaker adaptation using our watermarked speech. If there are problems playing the audio, please download and play them locally on your computer. Code and trained models for replicating our experiments are released. Please refer to our paper for more details.
Speaker: Obama
Watermarked:
section_000.wavOriginal:
section_000.wavSpeaker: "p234"
Watermarked:
p234_003_mic1.wavOriginal:
p234_003_mic1.flacSpeaker: "p245"
Watermarked:
p245_003_mic1.wavOriginal:
p245_003_mic1.flacSpeaker: "p261"
Watermarked:
p261_003_mic1.wavOriginal:
p261_003_mic1.flacSpeaker: "p294"
Watermarked:
p294_003_mic1.wavOriginal:
p294_003_mic1.flacSpeaker: "p302"
Watermarked:
p302_003_mic1.wavOriginal:
p302_003_mic1.flacPage updated
Google Sites
Report abuse