collaborators
Showing eess.ASShow all

10 papers · 1 filter

eess.AS2026

Long-Tail Rebalancing for Non-Verbal Vocalization-Aware ASR: A Track 1 System for the NVVSpeech Challenge

Shangyue Jia, Jingru Ma, Yangzhuo Li +9

Non-verbal vocalizations (NVVs) carry important paralinguistic information but are often omitted by conventional automatic speech recognition (ASR) systems. The ISCSLP NVVSpeech Ch…

eess.AS2026

Bridging Data, Reasoning, and Alignment: A Unified Framework for Context-Aware Instruction-Following TTS

Jingbin Hu, Luyu Wang, Wenjie Tian +7

The ISCSLP 2026 CoT-TTS Challenge requires TTS systems to generate Chain-of-Thought (CoT) reasoning from dialogue history before synthesizing contextually appropriate speech. While…

eess.AS2026

Modeling, Scaling, and Decoding: Optimizing Controllable Speech Generation with Nonverbal Vocalizations

Ziyu Zhang, Yun Chen, Taihui Wang +5

Controllable synthesis of nonverbal vocalizations (NVVs) is es- sential for natural and expressive speech, but remains challeng- ing due to their acoustic diversity and imbalanced…

eess.AS2026

Preference Optimization with LALM Feedback for Continuous Autoregressive Non-Verbal Vocalization Generation

Jingbin Hu, Qirui Zhan, Yuang Cao +8

We propose a preference optimization framework with Large Audio-Language Model (LALM) feedback for controllable non-verbal vocalization (NVV) generation in continuous autoregressiv…

eess.AS2026

SpeechAnnotator: A Context-Aware Multi-Agent Framework and Benchmark for Multidimensional Speech Annotation

Qirui Zhan, Shuiyuan Wang, Jingbin Hu +10

Recent controllable speech generation requires training data with fine-grained annotations of speaker traits, prosody, emotion, paralinguistic cues, acoustic scenes, and context. E…

eess.AS2026

HumDial-EIBench: A Human-Recorded Multi-Turn Emotional Intelligence Benchmark for Audio Language Models

Shuiyuan Wang, Zhixian Zhao, Hongfei Xue +5

Evaluating the emotional intelligence (EI) of audio language models (ALMs) is critical. However, existing benchmarks mostly rely on synthesized speech, are limited to single-turn i…