On January 2023, Google launched the MusicLM model which can generate various types of music based on the text prompts given. It can now create 5-minute songs based on a short text description after Google continued to improve its algorithm.
Google has now confirmed that its new algorithm has been trained on 280,000 hours of recordings. It has the ability to create 5-minute musical compositions based on text input, and the resulting content is closer to that of human composers.
You can listen to MusicLM-generated music here.
Google asked some people to try out the MusicLM model. According to the final results, MusicLM was better than other similar algorithms in every way.
However, the algorithm is still not at the stage where it can create complete songs. Currently, MusicLM cannot precisely understand the structure of music and the generated songs can be clearly distinguished.
But it does seem like in the near future we might have such models which will be able to understand the nuances of music and maybe a lot of future producers would just be bored teenagers at their homes who would be champions at prompting such software in order to produce the best sounding results!
Do let us know your honest opinions on the same in the comments below and feel free to subscribe to our push notifications for more such daily content!
So guys, if you liked this post and wish to receive more tech stuff delivered daily, don’t forget to subscribe to the Inspire2Rise newsletter to obtain more timely tech news, updates and more!