Tag with "Speech Technology"

A powerful new evaluation framework called UltraEval-Audio has emerged from collaboration between Tsinghua University, OpenBMB, and Miga Intelligence. This innovative tool simplifies audio model testing with its one-click evaluation and reproduction features, supporting specialized models like TTS and ASR. Version 1.1.0 brings significant upgrades that could accelerate development across the audio AI landscape.

OpenAI has introduced GPT-Realtime, a groundbreaking multimodal speech model that supports image input and real-time audio processing. The model enhances natural interactions with features like nonverbal signal recognition and language switching while reducing latency and costs. This release intensifies competition in the speech AI market and expands practical applications in customer service and education.