Projects that create songs, rhythms, and music in general usually work with small files that are less than a minute long. So, I was surprised by what the new version of Stable Audio 2.0 can do.
With this new version, available on Stableaudio.com, users now have the ability to generate music from descriptive texts and transform existing audio into something completely new, thanks to the audio-to-audio conversion capabilities. We can upload the audio we created and edit it as if we were editing an image.
One of the most exciting new features is style transfer. Imagine being able to take your audio track and, with a few simple commands, give it a classical bolero tone or the energetic rhythm of electronic music. This feature opens up a range of possibilities for experimentation and finding the unique sound you’re looking for. They explain this in this video :
I understand the concerns that may arise regarding copyright in the digital age, and I’m happy to see how Stability AI technology has taken steps to protect creative safety. The integration of Audible Magic technology to prevent copyright infringement is an important step towards respecting the work of artists and ensuring users stay within legal boundaries when creating their content.
They have achieved this by collaborating with AudioSparx to train the model with over 800,000 audio files, ensuring artists can opt out, demonstrating ethical commitment and respect for original creators.
Let’s get to what impresses me the most: the increase in track length that can be created, up to three minutes, with a stereo quality of 44.1 kHz. This is another qualitative leap towards producing complete musical works with the help of artificial intelligence.
For those interested in exploring Stable Audio 2.0, access is easy and free. By simply visiting the Stable Audio website and logging in using a Stable AI or Google account, a world of possibilities opens up for creating music in innovative and personal ways.
This is not the first time we have seen AI-powered music creation websites, as even Google has done it with MusicLM and Microsoft with SUNO, but it’s still surprising to see what appears every day.
– Link Stable Audio 2.0
