ad
ad

How to Generate AI Music with Mubert Text to Audio Tutorial

Howto & Style


Introduction

Introduction

Hey fellow dreamers! In this tutorial, we will explore how to generate AI music using a text-to-audio model. The tool we're using is developed by Mubert, a company that specializes in AI-generated music. They've recently released a new MVP (Minimum Viable Product) that allows us to experiment with creating music from textual prompts. Let's dive in!

Getting Started

First, I've set up a Google Colab notebook for this tutorial. If you're new to Google Colab, don't worry; it's straightforward. You'll run all the cells by clicking the play button, and you'll need to enter your email address. After that, we'll be ready to start generating some tunes.

First Audio Generation

Bird Soaring Through the Sky

For our first piece, the prompt is "a bird soaring through the sky," and we've set the duration to 20 seconds. Let's see what we get.

Result: [Music] The sound is pleasant and indeed gives the impression of a bird soaring through the sky. You can also activate the option to create a loopable audio track to prevent the sound from fading out at the end.

Experimenting with Different Prompts

Now, let's try something different. How about "Ancient Japanese futuristic vaporwave synthwave"?

Result: [Music] The music starts to embody these keywords, providing a unique blend of the themes.

More Experiments

Next, we'll try creating Lo-Fi music:

Lo-Fi Calm Hip Hop

Result: [Music] The beat is cool, and it's another successful generation.

We'll now increase the duration to 60 seconds and make it loopable:

Result: [Music] It could use a beat, but it's still a nice track.

Trying Different Moods

How about a post-apocalyptic, calm, and ambient setting with rainy effects?

Result: [Music] The sounds of flutes and possible rain provide a soothing ambiance. This is great for an intro or outro of a video.

Using Known Artist and Songs

Let's see how the AI handles known artists. We'll try "Hans Zimmer" and "Cornfield Chase" from the Interstellar soundtrack:

Result: [Music] The piano is in there, which is somewhat reminiscent of Hans Zimmer's style.

Next, we'll try "The Black Keys - Lonely Boy" with some hints like indie rock:

Result: [Music] The result is a bit monotone, and the variations could be improved for longer tracks.

We'll switch to "punk rock" next:

Result: [Music] This doesn't sound very punk rock to me, more like an advertising jingle.

Exploring Different Styles

Let's explore "J-pop":

Result: [Music] This track has more variation, which makes it quite enjoyable.

Adding "psychedelic" to the J-pop track:

Result: [Music] A cool, weird variant that has more layering and subtle effects switching from left to right.

Conclusion

Overall, this AI-generated music tool from Mubert provides a range of interesting and enjoyable musical pieces. While the tool is still in its MVP stage, it's promising and offers a lot of potentials. You can try it out yourself using the Google Colab notebook linked below.


Keywords

  • AI-generated music
  • Mubert
  • Google Colab
  • Text-to-Audio model
  • Loopable audio track
  • Lo-Fi
  • Post-apocalyptic
  • Ambient
  • Hans Zimmer
  • Indie rock
  • Punk rock
  • J-pop
  • Psychedelic

FAQ

What is Mubert?

Mubert is a company specializing in AI-generated music. They offer tools that can generate music from textual prompts.

How do I get started with the tool?

You can get started by using the Google Colab notebook. Run the cells, enter your email address, and you will be able to generate music with textual prompts.

Can I create loopable tracks?

Yes, there is an option to create loopable audio tracks. Simply activate the 'loopable' option when generating your music.

Does the tool support known artists and songs?

While the tool recognizes some attributes related to known artists and songs, the results can vary in accuracy and genre representation.

How can I add more variation to the generated tracks?

For longer tracks, try disabling the loop option to avoid repetitiveness and gain more variation in the music.


If you want to stay updated on Mubert's APIs and other developments, follow them on Twitter. Try out the tool for yourself and unleash your creativity!