OpenAI has released updates to two of its key models, Whisper (an AI for transcription) and Voice Engine (a voice synthesis technology). The new versions offer even better speech recognition accuracy and more realistic voice reproduction, bringing AI closer to the level of natural communication.
OpenAI has released updates to two of its key models, Whisper (an AI for transcription) and Voice Engine (a voice synthesis technology). The new versions offer even better speech recognition accuracy and more realistic voice reproduction, bringing AI closer to the level of natural communication.
OpenAI’s updates make voice technology even more accessible and accurate. This could significantly improve the performance of automatic transcription services, voice assistants, and even systems for narrating videos or audiobooks, reports TechCrunch.
Whisper, OpenAI’s speech recognition model, is now faster and more accurate. It can handle complex accents, background noise, and even corrupted audio recordings better. This makes the technology even more effective for automatically transcribing interviews, conferences, and other conversational formats.
OpenAI has also improved its Voice Engine, its voice generation model. It can now more accurately mimic the human voice based on a short sample. This opens up new possibilities for voice assistants, text-to-speech, and even personalized voice content.
Any user can test new voice synthesis capabilities on the special OpenAIFm website .
Recall that OpenAI announced the launch of o1-pro, the company's most advanced artificial intelligence model.
