Music editing is an important step in music production, which has broad applications, including game development and film production. Most existing zero-shot text-guided editing methods rely on pretrained diffusion models by involving forward-backward diffusion processes. However, these methods often struggle to preserve the musical content. Additionally, text instructions alone usually fail to accurately describe the desired music. In this paper, we propose two music editing methods that improve the consistency between the original and edited music by leveraging score distillation. The first method, SteerMusic, is a coarse-grained zero-shot editing approach using delta denoising score. The second method, SteerMusic+, enables fine-grained personalized music editing by manipulating a concept token that represents a user-defined musical style. SteerMusic+ allows for the editing of music into user-defined musical styles that cannot be achieved by the text instructions alone. Experimental results show that our methods outperform existing approaches in preserving both music content consistency and editing fidelity. User studies further validate that our methods achieve superior music editing quality.

We present SoundLoCD, a novel text-to-sound generation framework, which incorporates a LoRA-based conditional discrete contrastive latent diffusion model. Unlike recent large-scale sound generation models, our model can be efficiently trained under limited computational resources. The integration of a contrastive learning strategy further enhances the connection between text conditions and the generated outputs, resulting in coherent and high-fidelity performance. Our experiments demonstrate that SoundLoCD outperforms the baseline with greatly reduced computational resources. A comprehensive ablation study further validates the contribution of each component within SoundLoCD1.

We propose the stochastic optimal path which solves the classical optimal path problem by a probability-softening solution. This unified approach transforms a wide range of DP problems into directed acyclic graphs in which all paths follow a Gibbs distribution. We show the equivalence of the Gibbs distribution to a message-passing algorithm by the properties of the Gumbel distribution and give all the ingredients required for variational Bayesian inference of a latent path, namely Bayesian dynamic programming (BDP). We demonstrate the usage of BDP in the latent space of variational autoencoders (VAEs) and propose the BDP-VAE which captures structured sparse optimal paths as latent variables. This enables end-to-end training for generative tasks in which models rely on unobserved structural information. At last, we validate the behavior of our approach and showcase its applicability in two real-world applications: text-to-speech and singing voice synthesis. Our implementation code is available at https://github.com/XinleiNIU/LatentOptimalPathsBayesianDP.

We introduce HybridVC, a voice conversion (VC) framework built upon a pre-trained conditional variational autoencoder (CVAE) that combines the strengths of a latent model with contrastive learning. HybridVC supports text and audio prompts, enabling more flexible voice style conversion. HybridVC models a latent distribution conditioned on speaker embeddings acquired by a pretrained speaker encoder and optimises style text embeddings to align with the speaker style information through contrastive learning in parallel. Therefore, HybridVC can be efficiently trained under limited computational resources. Our experiments demonstrate HybridVC's superior training efficiency and its capability for advanced multimodal voice style conversion. This underscores its potential for widespread applications such as user-defined personalised voice in various social media platforms. A comprehensive ablation study further validates the effectiveness of our method.

Acoustic scene classification, where a scene is identified from a sound recording, is a difficult problem that is much less studied than similar problems in computer vision. Re-cent advances in attention-based convolution neural networks (CNNs) can be applied to audio data by operating on two dimensional spectrograms, where frequency and time infor-mation have been separated, rather than a raw audio signal. Typical CNNs have difficulty coping with this problem due to the temporal aspects of acoustic data. In this research we propose a novel and intuitive CNN-based architecture with attention mechanisms called the spatial-temporal-class attention network (STCANet). The STCANet consists of a spatial-temporal attention and a class attention which extracts in-formation along with frequency, temporal, and the class di-mension of spectrograms. In our experiments, the STCANet achieved 75.6%, 95.4%, and 97.0% accuracy on TUT 2018, TAU 2020, and ESC-I0 datasets that are competitive results compared with previous works. Our contributions include this novel network design and a detailed analysis of how attention allows these results to be achieved.