About four months after unveiling its text-to-music generator MusicLM, Google has officially released access to the AI-powered tool.
Google announced the expanded availability of MusicLM (and provided details on how users will test and improve the AI offering) in a text note. News. Interested people can now join a waiting list via the overarching platform to try MusicLM (on the web and on iOS and Android devices). AI test kitchen.
Meanwhile, those who have already gained access to the music generator are sharing their experience on YouTube and social media. “Be very descriptive,” MusicLM recommends in its instructions for creating a “good prompt” for the AI, which currently outputs two song snippets for each of those prompts and then asks users to identify the better option.
Predictably, the debut tool isn’t without its problems – a YouTuber The music prompt for “African bongos with Irish fiddle” was answered with an error message – but initial evidence suggests the resulting non-singing audio is broadly consistent with user input.
Building on the lack of vocals in MusicLM creations and the broader problem of unauthorized tracks using artificial intelligence, Google wrote on the AI platform’s main page that “certain requests” will be denied immediately. “MusicLM is an experimental technology that allows you to generate your own synthetic music for inspiration,” reads the relevant text. “Certain requests that mention specific artists or include vocals will not be generated.”
The latter constraint seems quite severe, as MusicLM accepts requests not only for Katy Perry’s sound-like works, but also for “classical music like Mozart,” as well as requests involving people at all (but not protected media like war of stars), test video shows. In addition, users can download MP3s of MusicLM songs, although the resulting (mono) audio clips appear to be only 20 seconds each and currently have a sample rate of 24,000 Hz.
Needless to say, it’s worth watching the development and spread of MusicLM in the future – especially given the unprecedented impact artificial intelligence is already having on the music world, although it’s still in its infancy.
publications (incl several whole albums) by AI Drake continue to make waves on video sharing services and social platforms, and AI music generator Boomy now announces that its users alone have “created 14,699,511 songs, around 14.04% of the music recorded worldwide”.
With this in mind, some artists are turning to AI, and in various cases are challenging fans to find out how their work compares to sound-like artificial intelligence efforts. Overshadowing these and other initiatives are concerns that uncritical listeners, who have been exposed to mainstream releases of questionable artistic merit for years, continue to play AI audio despite the availability of suitable music.