Modeling Music as a Time-Frequency Image: A 2D Tokenizer for Music Generation
A new paper presents BandTok, a 2D Mel-spectrogram tokenizer designed for autoregressive music generation. This tokenizer improves upon existing methods by offering a more independent token structure and better reconstruction quality. The authors demonstrate that BandTok outperforms traditional residual-codebook tokenizers, especially in data-limited scenarios.
- ▪BandTok represents each frame with Mel-frequency band tokens from a single shared codebook.
- ▪The design yields a physically interpretable time-frequency token grid.
- ▪Experiments show that BandTok achieves strong results in a data-limited setting.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Sound arXiv:2605.15831 (cs) [Submitted on 15 May 2026] Title:Modeling Music as a Time-Frequency Image: A 2D Tokenizer for Music Generation Authors:Yuqing Cheng, Xingyu Ma, Guochen Yu, Xiaotao Gu View a PDF of the paper titled Modeling Music as a Time-Frequency Image: A 2D Tokenizer for Music Generation, by Yuqing Cheng and 3 other authors View PDF HTML (experimental) Abstract:Autoregressive music generation depends strongly on the audio tokenizer. Existing high-fidelity codecs often use residual multi-codebook quantization, which preserves reconstruction quality but complicates language modeling after sequence flattening, as the residual hierarchy imposes strong sequential dependencies and can amplify error accumulation.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
