3 articles tagged with #speech-ai. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.
Voxtral TTS is a new multilingual text-to-speech AI model that can generate natural speech from just 3 seconds of reference audio. In human evaluations, it achieved a 68.4% win rate over ElevenLabs Flash v2.5 for voice cloning, demonstrating superior naturalness and expressivity.
Researchers conducted a cross-modal study comparing human preference annotations between text and audio formats for AI alignment. The study found that while audio preferences are as reliable as text, different modalities lead to different judgment patterns, with synthetic ratings showing promise as replacements for human annotations.
Researchers propose Task-Lens, a cross-task survey analyzing 50 Indian speech datasets across 26 languages for nine downstream speech tasks. The study reveals untapped metadata in existing datasets that could support multiple AI speech applications and identifies critical gaps in resources for underserved Indian languages.
