Google DeepMind Releases Lyria 3.5 Music Generation Model, Capable of Creating Songs Up to 3 Minutes Long

On July 30, Google DeepMind released the music generation model Lyria 3.5, enhancing capabilities in music structure, lyric quality, instruction adherence, vocal performance, and song duration control, allowing for the direct generation of complete songs up to 3 minutes long. Lyria 3.5 utilizes a latent diffusion architecture for diffusion generation in the time audio latent space, trained on audio data with varying granularity of text annotations, and employs SFT along with reinforcement learning incorporating human and critic feedback for post-training. All generated content is embedded with SynthID watermarks. Google has not disclosed the scale, source, or quantitative benchmarks of the training data, only stating that compared to Lyria 2, significant improvements have been made in audio clarity and lyric instruction adherence. Lyria 3.5 is initially integrated into Flow Music, supporting conversational music creation, stem separation, remixing, music publishing, and playlist generation, and can interact with Veo to generate music videos, supporting the development of audio plugins, music games, and custom DAWs, thereby building a complete ecosystem covering music creation, editing, publishing, and distribution.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.