karn watcharasupat

Cinematic audio source separation is a relatively new subtask of audio source separation, with the aim of extracting the dialogue, music, and effects stems from their mixture. In this work, we developed a model generalizing the Bandsplit RNN for any complete or overcomplete partitions of the frequency axis. Psychoacoustically motivated frequency scales were used to inform the band definitions which are now defined with redundancy for more reliable feature extraction. A loss function motivated by the signal-tonoise ratio and the sparsity-promoting property of the 1-norm was proposed. We additionally exploit the information-sharing property of a common-encoder setup to reduce computational complexity during both training and inference, improve separation performance for hard-to-generalize classes of sounds, and allow flexibility during inference time with detachable decoders. Our best model sets the state of the art on the Divide and Remaster dataset with performance above the ideal ratio mask for the dialogue stem.

Translation of perceptual soundscape attributes from one language to another remains a challenging task that requires a high degree of fidelity in both psychoacoustic and psycholinguistic senses across the target population. Due to the inherently subjective nature of human perception, translating soundscape attributes using only small focus group discussions or expert panels could lead to translations with psycholinguistic meanings that, in a non-expert setting, deviate or distort from that of the source language. In this work, we present a quantitative evaluation method based on the circumplex model of soundscape perception to assess the overall translation quality. By establishing a set of criteria for evaluating the linguistic and psychometric properties of the translation candidates, statistical analyses can be performed to objectively assess specific strengths and weaknesses of the translation candidates before committing to listening tests or more involved validation experiments. As an initial application domain, we demonstrated the use of the quantitative evaluation framework in the context of an English-to-Thai translation of soundscape attributes. A total of 31 participants who are bilingual in English and Thai were recruited to assess the translation candidates. Subsequent statistical analysis of the evaluation scores revealed acoustico-psycholinguistic properties of the translation candidates which were not previously identified by the expert panel and facilitated a more objective selection of the final translations for subsequent usage. Additionally, with specific biases of the final translations determined numerically, mathematical and statistical techniques for corrections of the survey data may be employed in the future to improve cross-lingual compatibility in soundscape evaluation.

The selection of maskers and playback gain levels in an in-situ soundscape augmentation system is crucial to its effectiveness in improving the overall acoustic comfort of a given environment. Traditionally, the selection of appropriate maskers and gain levels has been informed by expert opinion, which may not be representative of the target population, or by listening tests, which can be time- and labor-intensive. Furthermore, the resulting static choices of masker and gain are often inflexible to dynamic real-world soundscapes. In this work, we utilized a deep learning model to perform joint selection of the optimal masker and its gain level for a given soundscape. The proposed model was designed with highly modular building blocks, allowing for an optimized inference process that can quickly search through a large number of masker-gain combinations. In addition, we introduced the use of feature-domain soundscape augmentation conditioned on the digital gain level, eliminating the computationally expensive waveform-domain mixing process during inference, as well as the tedious gain adjustment process required for new maskers. The proposed system was evaluated on a large-scale dataset of subjective responses to augmented soundscapes with 442 participants, with the best model achieving a mean squared error of }{0.122}}mathbf {}pm }{0.005}{ on pleasantness score, validating the ability of the model to predict combined effect of the masker and its gain level on the perceptual pleasantness level. The proposed system thus allows in-situ or mixed-reality soundscape augmentation to be performed autonomously with near real-time latency while continuously accounting for changes in acoustic environments.