arXiv:2606. 01016v1 Announce Type: cross Abstract: While End-to-End (E2E) Speech-Large Language Models (Speech-LLMs) are rapidly evolving, their evaluation methodologies remain limited to the era of simple transcription.
arXiv:2605. 19266v2 Announce Type: replace-cross Abstract: Automatic speech recognition (ASR) systems are typically optimized for verbatim transcription, which preserves disfluencies, filler words, and informal spoken structures that are often unsuitable for downstream writing-oriented applications.
Recent advances in zero-shot text-to-speech (TTS) have substantially improved speech quality and voice cloning fidelity. However, many zero-shot TTS systems still depend on audio prompt transcripts at inference time.
SEA-SpeechBench is a large‑scale multitask benchmark for speech understanding in 11 Southeast Asian languages, comprising 97,194 samples across 99 evaluation sets and 597 hours of curated audio. It covers nine tasks in three categories—speech processing, paralinguistic analysis, and a novel temporal understanding dimension—using multilingual prompting in both native SEA languages and English. Evaluation of current models shows significant performance gaps, especially in temporal understanding, emotion recognition, and speech translation, with low‑resource languages lagging behind English by up to 41 percentage points.
X-AuT is a progressive framework for compressing the audio encoder of speech large language models. It selects layer combinations via short behavioral probes and restores performance through representation alignment, cross‑scale distillation, scheduled student‑policy supervision, and LoRA finetuning, while keeping the language‑model backbone frozen. On ten Chinese–English benchmarks, reducing Qwen3‑ASR‑0.6B’s encoder from 18 to 16 layers lowers macro‑average error from 5.61% to 5.27%, and a 14‑layer model achieves 5.75% error with 20.7% fewer parameters.
arXiv:2608. 04586v2 Announce Type: replace-cross Abstract: Multimodal large language models (MLLMs) have achieved significant success in speech-to-text translation (S2TT).
The paper introduces a unified phoneme‑based TTS‑to‑ASR augmentation pipeline that uses a multilingual TTS model with language‑ID conditioning and incorporates grapheme‑to‑phoneme conversion, reference‑speech filtering, and candidate‑text selection. It proposes phoneme‑frequency‑guided selection (PFGS) to rank sentences based on phoneme frequencies from real ASR labels, and demonstrates that random augmentation and PFGS both improve ASR performance across Arabic, French, Italian, and Portuguese test sets, with PFGS yielding up to a 19.3% relative WER reduction. The study also shows that filtering reference speech can further lower WER by up to 0.59 points on certain datasets.
arXiv:2508. 05149v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have demonstrated potential in handling spoken inputs for high-resource languages, reaching state-of-the-art performance in various tasks.
arXiv:2608. 04586v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) have achieved significant success in speech-to-text translation (S2TT).
Dual-Form ASR (DF-ASR) is a framework that unifies spoken-form ASR and semantics-aware written-form inverse text normalization (ITN) for Chinese speech recognition. It uses paired spoken- and written-form supervision generated and judged by a large language model, and introduces an ITN-MWER objective to penalize errors on normalization-sensitive spans. DF-ASR also employs a REQUIRE-ITN/FORBID-ITN protocol to separately evaluate required normalization and forbidden-span preservation, achieving superior performance over open-source ASR-ITN systems while maintaining prompt-level control between transcript forms.
arXiv:2606. 03957v1 Announce Type: cross Abstract: Conversational ASR for lower-resource languages and niche domains is limited by the scarcity of domain-matched multi-speaker training data.
The paper introduces methods to improve speech recognition for multilingual video transcription using Whisper-based tools, targeting cross‑cultural understanding. It reports an average transcription error rate of 30% across seven languages, which can be lowered to 20% with modest fine‑tuning. The authors also release associated speech and metadata to aid community refinement of these techniques.