Music Generation
This is another really interesting plugin. Think Stable Diffusion but text prompts generate music not images. This AI plugin uses MusicGen, a text-to-music model capable of generating high-quality music samples conditioned on text descriptions or audio prompts.
There are four models available. Let’s try the the musicgen-small-fp16-mono model.
Again I can use the CPU, GPU, or NPU for the generation process. Write a text prompt to describe the music I want to create, set the duration, context length, and more, In the example below, I’m creating a 20 second track.
With the Arc 140T GPU, a 20 second music sample takes around 1 minute 27 seconds to generate. It’s slightly faster using the NPU taking around 1 minute 6 seconds.
Audacity lets me export the generated track to a variety of formats including FLAC and mp3. Here’s the exported mp3 track.
This sample could form the basis of a backing track. OK it’s is not destined to break into the charts, but with the right vocals…. who knows?
There are three other models available: musicgen-small-fp16-stereo, musicgen-small-int8-mono, and musicgen-small-int8-stereo.
Here’s a track generated with the musicgen-small-fp16-stereo model using the same text prompt.
Let’s use a different text prompt, “mellow lo-fi beat with a soft electric piano, jazzy guitar riffs, and minimalistic drum patterns”.
Maybe that’s too complicated for what is only a small model. I’d love some other MusicGen models to be supported.
There’s also functionality to generate a continuation of an existing snippet of music.
Summary
For an iGPU, the Arc 140T is fast at generating music samples. And the NPU is even quicker for some tasks.
I’ve just looked at two of the five plugins. The Whisper transcription is equally as impressive as the music separation.
Website: github.com/intel/openvino-plugins-ai-audacity
Support:
Developer: Intel
License: GNU General Public License v3.0
AI Plugins for Audacity is written in C++. Learn C++ with our recommended free books and free tutorials.
Pages in this article:
Page 1 – Introduction and Installation
Page 2 – Music Separation
Page 3 – Music Generation and Summary
Complete list of articles in this series:
| ASRock Industrial NUC BOX-255H | |
|---|---|
| Introduction | Introduction to the series and interrogation of the NUC BOX-255H |
| Benchmarks | Benchmarking the NUC BOX-255H |
| Power | Testing and comparing the power consumption under various workloads |
| Stable Diffusion | Deep Learning with Stable Diffusion |
| Audacity AI Plugins | Let's explore OpenVINO AI Plugins for Audacity |