• 2 min read
Google launches Lyria 3.5 with richer vocals
Google’s Lyria 3.5 music model promises richer sound, more natural vocals, and finer control over tempo and track length.
Image: ITzine
Google has released Lyria 3.5, the latest version of its music-generation model. The company says it produces denser, more complex sound, understands musical structure more accurately, and responds more precisely to text prompts.
The biggest change is the vocal performance. Google says Lyria 3.5 generates vocals that sound more natural, realistic, and emotional. That could matter particularly for tracks in which the voice is a central element rather than background texture.
What Lyria 3.5 can do
According to Google, the updated model:
- Generates music with richer and more complex sound.
- Better understands musical structure and follows text prompts more accurately.
- Produces more realistic and emotional vocals.
- Provides greater control over a track’s tempo and duration.
Google also says it is now easier to adjust a song’s tempo and length. That is useful when adapting generated music to a specific format, from a short jingle to a full composition. The company did not provide benchmark results or explain how these improvements were measured.
Lyria 3.5 is already available on Google’s Flow Music web platform, where Google has also published an example song made with the new model. The company did not announce pricing or provide additional availability details.
Recommended reading
OpenAI plans free ChatGPT access for 100,000 researchers
The release comes as Suno and Udio gain visibility and major platforms expand generative tools beyond text and images into audio. It also arrives amid continuing questions stemming from last year’s disputes over training data and copyright—issues that affect both the quality of these systems and how their output can be used.
Google presented three models for productivity last week; Lyria 3.5 is the latest addition to its broader set of generative tools.
Marta Barinova is an editor in the news department specializing in software, streaming services, and policy changes at global technology platforms. She has written more than 140 articles covering Windows updates, Spotify and Google feature changes, and antitrust regulation of app stores.
AI Editor
Ava covers the rapidly evolving world of artificial intelligence, from foundational models and research labs to the real-world economics of intelligence. With a background in computational linguistics, she cuts through the hype to find out what actually works. She firmly believes that benchmarks are just marketing until reproduced in the wild.
via ITzine
