Jukebox: A Generative Model for Music
arXiv:2005.00341
Abstract
We introduce Jukebox, a model that generates music with singing in the raw audio domain. We tackle the long context of raw audio using a multi-scale VQ-VAE to compress it to discrete codes, and modeling those using autoregressive Transformers. We show that the combined model at scale can generate high-fidelity and diverse songs with coherence up to multiple minutes. We can condition on artist and genre to steer the musical and vocal style, and on unaligned lyrics to make the singing more controllable. We are releasing thousands of non cherry-picked samples at https://jukebox.openai.com, along with model weights and code at https://github.com/openai/jukebox
References in corpus (7)
- MelGAN: Generative Adversarial Networks for Conditional Waveform Synthesis
- Generating Long Sequences with Sparse Transformers
- Axial Attention in Multidimensional Transformers
- SampleRNN: An Unconditional End-to-End Neural Audio Generation Model
- MelNet: A Generative Model for Audio in the Frequency Domain
- Fixup Initialization: Residual Learning Without Normalization
- Counterpoint by Convolution
Cited by in corpus (6)
- POP909: A Pop-song Dataset for Music Arrangement Generation
- A Survey on Self-supervised Pre-training for Sequential Transfer Learning in Neural Networks
- DeepSinger: Singing Voice Synthesis with Data Mined From the Web
- Improved Prosody from Learned F0 Codebook Representations for VQ-VAE Speech Waveform Reconstruction
- Speech-to-Singing Conversion based on Boundary Equilibrium GAN
- Hierarchical Timbre-Painting and Articulation Generation