Soundscape Captioning using Sound Affective Quality Network and Large Language Model

We live in a rich and varied acoustic world, which is experienced by individuals or communities as a soundscape. Computational auditory scene analysis, disentangling acoustic scenes by detecting and classifying events, focuses on objective attributes of sounds, such as their category and temporal characteristics, ignoring the effect of sounds on people and failing to explore the relationship between sounds and the emotions they evoke within a context. To fill this gap and to automate soundscape analysis, which traditionally relies on labour-intensive subjective ratings and surveys, we propose the soundscape captioning (SoundSCap) task. SoundSCap generates context-aware soundscape descriptions by capturing the acoustic scene, event information, and the corresponding human affective qualities. To this end, we propose an automatic soundscape captioner (SoundSCaper) composed of an acoustic model, SoundAQnet, and a general large language model (LLM). SoundAQnet simultaneously models multi-scale information about acoustic scenes, events, and perceived affective qualities, while LLM generates soundscape captions by parsing the information captured by SoundAQnet to a common language. The soundscape caption's quality is assessed by a jury of 16 audio/soundscape experts. The average score (out of 5) of SoundSCaper-generated captions is lower than the score of captions generated by two soundscape experts by 0.21 and 0.25, respectively, on the evaluation set and the model-unknown mixed external dataset with varying lengths and acoustic properties, but the differences are not statistically significant. Overall, SoundSCaper-generated captions show promising performance compared to captions annotated by soundscape experts. The models' code, LLM scripts, human assessment data and instructions, and expert evaluation statistics are all publicly available.

No More Mumbles: Enhancing Robot Intelligibility through Speech Adaptation

Spoken language interaction is at the heart of interpersonal communication, and people flexibly adapt their speech to different individuals and environments. It is surprising that robots, and by extension other digital devices, are not equipped to adapt their speech and instead rely on fixed speech parameters, which often hinder comprehension by the user. We conducted a speech comprehension study involving 39 participants who were exposed to different environmental and contextual conditions. During the experiment, the robot articulated words using different vocal parameters, and the participants were tasked with both recognising the spoken words and rating their subjective impression of the robot's speech. The experiment's primary outcome shows that spaces with good acoustic quality positively correlate with intelligibility and user experience. However, increasing the distance between the user and the robot exacerbated the user experience, while distracting background sounds significantly reduced speech recognition accuracy and user satisfaction. We next built an adaptive voice for the robot. For this, the robot needs to know how difficult it is for a user to understand spoken language in a particular setting. We present a prediction model that rates how annoying the ambient acoustic environment is and, consequentially, how hard it is to understand someone in this setting. Then, we develop a convolutional neural network model to adapt the robot's speech parameters to different users and spaces, while taking into account the influence of ambient acoustics on intelligibility. Finally, we present an evaluation with 27 users, demonstrating superior intelligibility and user experience with adaptive voice parameters compared to fixed voice.

Child Speech Recognition in Human-Robot Interaction: Problem Solved?

Automated Speech Recognition shows superhuman performance for adult English speech on a range of benchmarks, but disappoints when fed children's speech. This has long sat in the way of child-robot interaction. Recent evolutions in data-driven speech recognition, including the availability of Transformer architectures and unprecedented volumes of training data, might mean a breakthrough for child speech recognition and social robot applications aimed at children. We revisit a study on child speech recognition from 2017 and show that indeed performance has increased, with newcomer OpenAI Whisper doing markedly better than leading commercial cloud services. While transcription is not perfect yet, the best model recognises 60.3% of sentences correctly barring small grammatical differences, with sub-second transcription time running on a local GPU, showing potential for usable autonomous child-robot speech interactions.

Multi-level graph learning for audio event classification and human-perceived annoyance rating prediction

WHO's report on environmental noise estimates that 22 M people suffer from chronic annoyance related to noise caused by audio events (AEs) from various sources. Annoyance may lead to health issues and adverse effects on metabolic and cognitive systems. In cities, monitoring noise levels does not provide insights into noticeable AEs, let alone their relations to annoyance. To create annoyance-related monitoring, this paper proposes a graph-based model to identify AEs in a soundscape, and explore relations between diverse AEs and human-perceived annoyance rating (AR). Specifically, this paper proposes a lightweight multi-level graph learning (MLGL) based on local and global semantic graphs to simultaneously perform audio event classification (AEC) and human annoyance rating prediction (ARP). Experiments show that: 1) MLGL with 4.1 M parameters improves AEC and ARP results by using semantic node information in local and global context aware graphs; 2) MLGL captures relations between coarse and fine-grained AEs and AR well; 3) Statistical analysis of MLGL results shows that some AEs from different sources significantly correlate with AR, which is consistent with previous research on human perception of these sound sources.

AI-based soundscape analysis: Jointly identifying sound sources and predicting annoyance

Soundscape studies typically attempt to capture the perception and understanding of sonic environments by surveying users. However, for long-term monitoring or assessing interventions, sound-signal-based approaches are required. To this end, most previous research focused on psycho-acoustic quantities or automatic sound recognition. Few attempts were made to include appraisal (e.g., in circumplex frameworks). This paper proposes an artificial intelligence (AI)-based dual-branch convolutional neural network with cross-attention-based fusion (DCNN-CaF) to analyze automatic soundscape characterization, including sound recognition and appraisal. Using the DeLTA dataset containing human-annotated sound source labels and perceived annoyance, the DCNN-CaF is proposed to perform sound source classification (SSC) and human-perceived annoyance rating prediction (ARP). Experimental findings indicate that (1) the proposed DCNN-CaF using loudness and Mel features outperforms the DCNN-CaF using only one of them. (2) The proposed DCNN-CaF with cross-attention fusion outperforms other typical AI-based models and soundscape-related traditional machine learning methods on the SSC and ARP tasks. (3) Correlation analysis reveals that the relationship between sound sources and annoyance is similar for humans and the proposed AI-based DCNN-CaF model. (4) Generalization tests show that the proposed model's ARP in the presence of model-unknown sound sources is consistent with expert expectations and can explain previous findings from the literature on sound-scape augmentation.

Joint Prediction of Audio Event and Annoyance Rating in an Urban Soundscape by Hierarchical Graph Representation Learning

Sound events in daily life carry rich information about the objective world. The composition of these sounds affects the mood of people in a soundscape. Most previous approaches only focus on classifying and detecting audio events and scenes, but may ignore their perceptual quality that may impact humans' listening mood for the environment, e.g. annoyance. To this end, this paper proposes a novel hierarchical graph representation learning (HGRL) approach which links objective audio events (AE) with subjective annoyance ratings (AR) of the soundscape perceived by humans. The hierarchical graph consists of fine-grained event (fAE) embeddings with single-class event semantics, coarse-grained event (cAE) embeddings with multi-class event semantics, and AR embeddings. Experiments show the proposed HGRL successfully integrates AE with AR for AEC and ARP tasks, while coordinating the relations between cAE and fAE and further aligning the two different grains of AE information with the AR.

Tactile interaction with a robot leads to increased risk-taking

Tactile interaction plays a crucial role in interactions between people. Touch can, for example, help people calm down and lower physiological stress responses. Consequently, it is believed that tactile and haptic interaction matter also in human-robot interaction. We study if the intensity of the tactile interaction has an impact on people, and do so by studying whether different intensities of tactile interaction modulate physiological measures and task performance. We use a paradigm in which a small humanoid robot is used to encourage risk-taking behaviour, relying on peer encouragement to take more risks which might lead to a higher pay-off, but potentially also to higher losses. For this, the Balloon Analogue Risk Task (BART) is used as a proxy for the propensity to take risks. We study four conditions, one control condition in which the task is completed without a robot, and three experimental conditions in which a robot is present that encourages risk-taking behaviour with different degrees of tactile interaction. The results show that both low-intensity and high-intensity tactile interaction increase people's risk-taking behaviour. However, low-intensity tactile interaction increases comfort and lowers stress, whereas high-intensity touch does not.