Text-Driven Foley Sound Generation With Latent Diffusion Model

Foley sound generation aims to synthesise the background sound for multimedia content. Previous models usually employ a large development set with labels as input (e.g., single numbers or one-hot vector). In this work, we propose a diffusion model based system for Foley sound generation with text conditions. To alleviate the data scarcity issue, our model is initially pre-trained with large-scale datasets and fine-tuned to this task via transfer learning using the contrastive language-audio pertaining (CLAP) technique. We have observed that the feature embedding extracted by the text encoder can significantly affect the performance of the generation model. Hence, we introduce a trainable layer after the encoder to improve the text embedding produced by the encoder. In addition, we further refine the generated waveform by generating multiple candidate audio clips simultaneously and selecting the best one, which is determined in terms of the similarity score between the embedding of the candidate clips and the embedding of the target text label. Using the proposed method, our system ranks ${1}^{st}$ among the systems submitted to DCASE Challenge 2023 Task 7. The results of the ablation studies illustrate that the proposed techniques significantly improve sound generation performance. The codes for implementing the proposed system are available online.

Latent Diffusion Model Based Foley Sound Generation System For DCASE Challenge 2023 Task 7

Foley sound generation aims to synthesise the background sound for multimedia content, which involves computationally modelling sound effects with specialized techniques. In this work, we proposed a diffusion based generative model for DCASE 2023 challenge task 7: Foley Sound Synthesis. The proposed system is based on AudioLDM, which is a diffusion-based text-to-audio generation model. To alleviate the data scarcity of the task 7 training set, our model is initially trained with large-scale datasets and downstream into this DCASE task via transfer learning. We have observed that the feature extracted by the encoder can significantly affect the performance of the generation model. Hence, we improve the results by leveraging the input label with related text embedding features obtained by a large language model, i.e., contrastive language-audio pretraining (CLAP). In addition, we utilize a filtering strategy to further refine the output, i.e. by selecting the best results from the candidate clips generated in terms of the similarity score between the sound and target labels. The overall system achieves a Frechet audio distance (FAD) score of 4.765 on average among all seven different classes, substantially outperforming the baseline system which achieves a FAD score of 9.7.