AI now has the capability to generate music samples from text prompts using neural networks. It’s actually great for producing compositions based on provided descriptions like mood or style. These AI models can be used to generate video game soundtracks, or simply to break creative block while relying on quality data and ongoing AI advancements. The innovative tech team Samplab has unveiled TextToSample, a VST3 (Virtual Studio Technology) plugin that allows users to convert text prompts into generated audio samples. The tool, available for free, runs both independently and within a DAW (Digital Audio Workstation), and could be the answer to generating a unique “melody” straight from your thoughts. Imagine the possibilities.

The plugin does use a lot of resources while generating even a medium-sized clip

Drawing upon the potential of Meta‘s open-source MusicGen, an AI-based sound generation suite, TextToSample was developed using the data fed by this advanced algorithm. Adding to its capabilities, the plugin can accept uploaded sounds, analyze them, and generate related audio samples, besides processing text instructions.

TexttoSample

Sporting a very minimalist interface, TextToSample does, however, come with a few challenges. The responses are not always accurate, and you have to grapple and wrestle back and forth with multiple prompts to actually get where you need to be. So, perfection may not be immediate, but there’s a chance you can get there.

Despite these limitations, TextToSample can be a lot of fun. It’s a great example of generative AI‘s capabilities, especially with regard to music production in 2023, although it’s not without its share of bugs. Some users experienced unwanted additions, such as drums when they specifically opted out of percussion. I personally used the application, and it used up a huge amount of resources to generate a 10-second audio clip that was of mediocre quality at best, slowing my machine down in the process.

RELATED:


(Via)

RELATED ARTICLESMORE FROM AUTHOR