Multilingual Multi-Speaker Unit Vocoders: A Systematic Analysis of Discrete Speech Representations

Discrete speech units obtained via k-means clustering of self supervised embeddings entangle phonetic, speaker, and language information, causing speaker mixing and cross-lingual interference in multilingual multi-speaker speech generation. Despite growing use in Audio LLMs and speech to speech systems, unit vocoders remain underexplored. We analyze a BigVGAN based unit vocoder, across four Indian languages. We study the interaction between cluster size and conditioning strategies using WER, speaker similarity, and unit level metrics. Results show that cluster size governs intelligibility by improving phonetic discriminability, while explicit speaker conditioning is indispensable for preventing identity collapse. Language supervision yields further gains mainly at lower cluster sizes where units remain ambiguous. Our analysis shows similar phonemes across languages collapse to the same cluster IDs at smaller inventories, with larger clusters progressively separating them.

Enhancing Audio Captioning with Auxiliary AudioSet Semantics

Automatic Audio Captioning (AAC) seeks to generate natural language descriptions of complex acoustic scenes, bridging auditory perception and language understanding. However, word-selection indeterminacy and increasing reliance on large-scale sequence-to-sequence or LLM-based models limit practical deployment. We propose a resource-efficient AAC framework that explicitly grounds caption generation in auxiliary AudioSet semantics. Frame-level acoustic representations extracted using a ConvNeXt encoder are augmented with top-$K$ predicted AudioSet keywords, providing structured contextual cues for decoding. A compact six-layer BART-style decoder conditions on this joint acoustic-semantic representation, enabling caption generation without LLM-scale decoding. The proposed design balances semantic grounding and computational efficiency within a compact architecture. Evaluations on Clotho V2 and AudioCaps confirm competitive caption quality under practical deployment constraints.

Shared Representation Learning for Reference-Guided Targeted Sound Detection

Human listeners exhibit the remarkable ability to segregate a desired sound from complex acoustic scenes through selective auditory attention, motivating the study of Targeted Sound Detection (TSD). The task requires detecting and localizing a target sound in a mixture when a reference audio of that sound is provided. Prior approaches, rely on generating a sound-discriminative conditional embedding vector for the reference and pairing it with a mixture encoder, jointly optimized with a multi-task learning approach. In this work, we propose a unified encoder architecture that processes both the reference and mixture audio within a shared representation space, promoting stronger alignment while reducing architectural complexity. This design choice not only simplifies the overall framework but also enhances generalization to unseen classes. Following the multi-task training paradigm, our method achieves substantial improvements over prior approaches, surpassing existing methods and establishing a new state-of-the-art benchmark for targeted sound detection, with a segment-level F1 score of 83.15% and an overall accuracy of 95.17% on the URBAN-SED dataset.