Google announced at its annual conference Google Cloud Next it will be launching updates to its GenAI models in Vertex AI Studios, now available in preview and general release.
The updates include improvements to its GenAI text-to-music, image, speech and video creation tools.
Lyria and Veo 2 are now in preview for some customers. Lyria, the tech giant’s text-to-music model, is being marketed as an alternative to royalty-free music libraries as it enables users to create music across a range of styles.
Veo 2 enhances video editing with features like being able to remove backgrounds, create time-lapses, adjust camera angles, and convert landscape into portrait and vice-versa.
Google’s Chirp 3 model is now also generally available for use. This tool can bring voice cloning in 35 languages and speaker separation when there are multiple people on a recording.
Imagen 3, Google’s image generation tool, has been upgraded with improved object removal capabilities – enabling users to edit images through text prompts.
Google said that to promote transparency and responsible AI use, all content generated by Imagen, Veo and Lyria will be embedded with SynthID, Google’s own watermarking technology that helps identify AI-generated media.
SynthID can analyse images, audio, text, or video for digital watermarks, allowing users to identify whether content, or elements of it, were created using Google’s AI tools.
