Understanding the internal mechanisms of large audio-language models (LALMs) is crucial for interpreting their behavior and improving performance. This work presents the first in-depth analysis of how LALMs internally perceive and recognize auditory attributes. By applying vocabulary projection on three state-of-the-art LALMs, we track how attribute information evolves across layers and token positions. We find that attribute information generally decreases with layer depth when recognition fails, and that resolving attributes at earlier layers correlates with better accuracy. Moreover, LALMs heavily rely on querying auditory inputs for predicting attributes instead of aggregating necessary information in hidden states at attribute-mentioning positions. Based on our findings, we demonstrate a method to enhance LALMs. Our results offer insights into auditory attribute processing, paving the way for future improvements.

With advancements in large audio-language models (LALMs), which enhance large language models (LLMs) with auditory capabilities, these models are expected to demonstrate universal proficiency across various auditory tasks.While numerous benchmarks have emerged to assess LALMs' performance, they remain fragmented and lack a structured taxonomy.To bridge this gap, we conduct a comprehensive survey and propose a systematic taxonomy for LALM evaluations, categorizing them into four dimensions based on their objectives: (1) General Auditory Awareness and Processing, (2) Knowledge and Reasoning, (3) Dialogue-oriented Ability, and (4) Fairness, Safety, and Trustworthiness.We provide detailed overviews within each category and highlight challenges in this field, offering insights into promising future directions.To the best of our knowledge, this is the first survey specifically focused on the evaluations of LALMs, providing clear guidelines for the community.We will release the collection of the surveyed papers and actively maintain it to support ongoing advancements in the field.

Multimodal foundation models, such as Gemini and ChatGPT, have revolutionized human-machine interactions by seamlessly integrating various forms of data. Developing a universal spoken language model that comprehends a wide range of natural language instructions is critical for bridging communication gaps and facilitating more intuitive interactions. However, the absence of a comprehensive evaluation benchmark poses a significant challenge. We present Dynamic-SUPERB Phase-2, an open and evolving benchmark for the comprehensive evaluation of instruction-based universal speech models. Building upon the first generation, this second version incorporates 125 new tasks contributed collaboratively by the global research community, expanding the benchmark to a total of 180 tasks, making it the largest benchmark for speech and audio evaluation. While the first generation of Dynamic-SUPERB was limited to classification tasks, Dynamic-SUPERB Phase-2 broadens its evaluation capabilities by introducing a wide array of novel and diverse tasks, including regression and sequence generation, across speech, music, and environmental audio. Evaluation results show that no model performed well universally. SALMONN-13B excelled in English ASR and Qwen2-Audio-7B-Instruct showed high accuracy in emotion recognition, but current models still require further innovations to handle a broader range of tasks. We open-source all task data and the evaluation pipeline at https://github.com/dynamic-superb/dynamic-superb.

We introduce a new zero resource code-switched speech bench-mark designed to assess the code-switching capabilities of self-supervised speech encoders directly. We showcase a baseline system of language modeling on discrete units to demonstrate how the code-switching abilities of speech encoders can be assessed in a zero-resource manner. Our experiments encompass a variety of well-known speech encoders, including Wav2vec 2.0, HuBERT, XLSR, etc., on three tracks of different code-switched language pairs: Spanish-English, French-English, and Chinese-English. We examine the impact of pre-training languages and model size on benchmark performance. Notably, though our results demonstrate that speech encoders with multilingual pre-training, exemplified by XLSR, outperform monolingual variants (Wav2vec 2.0, HuBERT) in code-switching scenarios, there is still substantial room for improvement in their code-switching linguistic abilities.

In this work, we introduce Speech-Copilot, a modular framework for instruction-oriented speech-processing tasks that minimizes human effort in toolset construction. Unlike end-to-end methods using large audio-language models, Speech-Copilot builds speech processing-specific toolsets by analyzing pre-collected task instructions and breaking tasks into manageable sub-tasks. It features a flexible agent that performs tasks through program generation based on large language models. Our approach achieves state-of the-art performance on the Dynamic-SUPERB benchmark, demonstrating its effectiveness across diverse speech-processing tasks. Key contributions include: 1) developing an innovative framework for speech processing-specific toolset construction, 2) establishing a high-performing agent based on large language models, and 3) offering a new perspective on addressing challenging instruction-oriented speech-processing tasks. Without additional training required by end-to-end approaches, our method provides a flexible and extendable solution for a wide range of speech-processing applications.

Speech Integrated Large Language Models (SILLMs) combine large language models with speech perception to perform diverse tasks, such as emotion recognition to speaker verification, demonstrating universal audio understanding capability. However, these models may amplify biases present in training data, potentially leading to biased access to information for marginalized groups. This work introduces a curated spoken bias evaluation toolkit and corresponding dataset. We evaluate gender bias in SILLMs across four semantic-related tasks: speech-to-text translation (STT), spoken coreference resolution (SCR), spoken sentence continuation (SSC), and spoken question answering (SQA). Our analysis reveals that bias levels are language-dependent and vary with different evaluation methods. Our findings emphasize the necessity of employing multiple approaches to comprehensively assess biases in SILLMs, providing insights for developing fairer SILLM systems.

This research explores how the information of prompts interacts with the high-performing speech recognition model, Whisper. We compare its performances when prompted by prompts with correct information and those corrupted with incorrect information. Our results unexpectedly show that Whisper may not understand the textual prompts in a human-expected way. Additionally, we find that performance improvement is not guaranteed even with stronger adherence to the topic information in textual prompts. It is also noted that English prompts generally outperform Mandarin ones on datasets of both languages, likely due to differences in training data distributions for these languages despite the mismatch with pre-training scenarios. Conversely, we discover that Whisper exhibits awareness of misleading information in language tokens by ignoring incorrect language tokens and focusing on the correct ones. In sum, We raise insightful questions about Whisper’s prompt understanding and reveal its counter-intuitive behaviors. We encourage further studies.

The spectrally compact pilot waveform that exhibits small power spectral sidelobes and thus high spectral compactness while achieving accurate channel estimation is desirable for orthogonal frequency division multiplexing systems. However, existing pilot waveforms either suffer large power spectral sidelobes or can achieve high spectral compactness only when the length of the carried pilot sequence is limited to a constrained set of non-prime numbers, particularly those containing a large positive integer power of two as a factor. In this paper, new procedures are proposed for designing more pilot sequences with a broader choice of sequence lengths, so that spectrally compact pilot waveforms can be constructed to achieve optimal channel estimation on quasi-static multipath channels.

In this work, we propose three new static search algorithms and one dynamic search algorithm for fast bit allocation of an MP3 audio encoder. They can substantially reduce the bit allocation operations (for finding the optimal global gain factors) of an MP3 encoder. Specifically, the new dynamic method can reduce the loop iteration counts of the bit allocation block by up to 30% that of the current known O/sup th/-order and 1/sup st/-order predictive search algorithms. It is achieved by taking into account of the statistical properties of the global gain factors. Since generally computation time of the bit allocation loop takes up more than 70% of a complete MP3 coding process, the new algorithm is good for high-speed and low-power realization of an MP3 encoder.