Google Releases Lyria 3.5, an AI-Powered Music Generator with Japanese Vocals

Google has recently released Lyria 3.5, a music generation tool that leverages artificial intelligence to create songs with Japanese vocals up to three minutes in length. This development marks the latest advancement in Google’s efforts to push the boundaries of AI-powered creativity.

To access Lyria 3.5, users can visit Google Flow Music, a service that allows individuals to generate and customize music using various tools, including the new AI-powered generator. The platform is available for free, with optional paid plans offering additional features and increased credits.

Upon logging in to Google Flow Music, users are presented with an explanation of how the tool works and what they can expect from their experience. This includes information on generating music based on input images or text prompts, as well as details about the various file formats available for download.

To get started with Lyria 3.5, users must first create a Google account if they don’t already have one. Once logged in, they will be taken to an explanation of how the tool works and what features are included. This is followed by a prompt to read and agree to the terms of service before proceeding.

The actual process of generating music using Lyria 3.5 involves entering a text prompt that describes the desired song. In one example, the user typed ‘It’s the middle of summer. This is an energetic song by a female idol group that conveys the message, ‘It’s too hot, but let’s do our best,’ and ‘Let’s make sure to stay hydrated.’ It has multiple vocalists and solo parts,’ before clicking the send button.

After waiting for a short period, two songs titled ‘Full Power Hydration’ were generated. The user then attempted to download one of these songs as an audio file in MP3 format, but encountered an error that prevented successful completion.

Despite this setback, the user was able to successfully generate and download another song using Lyria 3.5. This second attempt resulted in a song titled ‘Zenryoku Hydration,’ which featured lyrics that described scenes of summer heat and hydration. The song’s style and tone were reminiscent of those found in Japanese pop music.

The generated songs can be customized further by selecting different file formats or adjusting the tempo, but this feature is only available with paid plans. Users who opt for the free plan will still have access to a range of features, including generating music based on input images or text prompts.

In addition to its AI-powered generator, Lyria 3.5 also includes a function that allows users to generate music based on input images. This feature is particularly useful when combined with other tools available through Google Flow Music, such as the ability to upload and customize images before generating music.

The user attempted this feature by uploading an image of ramen and rice, along with a text prompt describing how beautiful it was. The resulting song, titled ‘Amber Breath,’ featured lyrics that poetically described scenes of eating and enjoying food in a peaceful atmosphere.

Lyria 3.5’s capabilities extend beyond generating music based on images or text prompts. It also includes features such as the ability to adjust tempo, select different file formats, and customize other aspects of the generated song. However, these advanced features are only available with paid plans.

The cost of accessing Lyria 3.5 through Google Flow Music varies depending on whether users opt for a free or paid plan. The annual subscription fees range from $6 per month (approximately 981 yen) for the Starter plan to $48 per month (approximately 7847 yen) for the Member plan.