You can now generate music samples from AI text prompts with TextToSamples plugin

TTS.png

Do you fancy yourself as the next Dr Dre or DJ Khaled? This AI software might help unlock that hidden talent of yours. The same devs behind Samplab are back with a new addition. According to them, the free VST3 plugin could generate samples from text prompts.

The VST3 plugin, also known as the TextToSample, is a plugin that opens inside your DAW (Digital Audio Workstation). It can also work as a standalone tool. For instance, if you type “bubble synth melody”, it will create a bubble synth melody.

Moreover, TextToSample uses MusicGen to achieve this goal. In case you didn’t know, MusicGen is Meta’s open-source AI-based sound generation tool set. Plus, the TextToSample is trained using data provided by the algorithm. On top of that, you can also drag and drop pre-existing sounds into the plugin and have it generate related samples.

UI-wise, it is beginner-friendly and easy to use. The layout for the UI is also minimal and sparse. However, it does have a few limitations. For instance, it does not consider pre-existing music on your track in its thought process. It also has problems recognising BPM, keys and scales. So it will take you a few attempts to get the preferred tone.

As expected, the plugin is just like any other software. It contains bugs. For instance, we didn’t want drum sounds in the music, but the plugin continues to add them anyways. After all, Samplab mentioned:

“The tool is not intended to replace human musicians, which it’s not capable of doing anyway.”

What are your thoughts on this news? Stay tuned for more updates like this at TechNave!