[ByteDance Seed Audio 1.0 Launched: Generate Dialogue, Sound Effects, and Ambient Sounds with a Single Prompt] According to monitoring by Beating, ByteDance's Seed has released the audio creation model Seed Audio 1.0. It can generate dialogue, sound effects, and ambient sounds simultaneously with a single prompt, while controlling the timing of sound entry on a timeline. The model supports both text and reference audio inputs, with a timing control precision of 100ms. It can generate approximately 2 minutes of audio in one go, with the ability to extend further while maintaining consistent character voice tones. Seed Audio 1.0 supports over 20 languages. The same voice tone can be transferred across languages, and it can also switch tones and emotions. Official evaluations show that the audio usability rate exceeds 90% in most scenarios. The naturalness MOS for most languages is above 4 points. The model is now available on the Volcano Ark Experience Center and has opened API access. [Original Link]