Msanii: High Fidelity Music Synthesis on a Shoestring Budget
This addresses the challenge of high-quality music synthesis for applications requiring long samples, representing a novel advancement rather than an incremental improvement.
The paper tackles the problem of synthesizing long-context, high-fidelity music efficiently by introducing Msanii, a diffusion-based model that generates 190 seconds of stereo music at 44.1 kHz without using concatenative synthesis, cascading architectures, or compression techniques.
In this paper, we present Msanii, a novel diffusion-based model for synthesizing long-context, high-fidelity music efficiently. Our model combines the expressiveness of mel spectrograms, the generative capabilities of diffusion models, and the vocoding capabilities of neural vocoders. We demonstrate the effectiveness of Msanii by synthesizing tens of seconds (190 seconds) of stereo music at high sample rates (44.1 kHz) without the use of concatenative synthesis, cascading architectures, or compression techniques. To the best of our knowledge, this is the first work to successfully employ a diffusion-based model for synthesizing such long music samples at high sample rates. Our demo can be found https://kinyugo.github.io/msanii-demo and our code https://github.com/Kinyugo/msanii .