Speech Identification and Comprehension in the Urban Soundscape

Abstract

Urban environments are characterised by the presence of copious and unstructured noise. This noise continuously challenges speech intelligibility both in normal-hearing and hearing-impaired individuals. In this paper, we investigate the impact of urban noise, such as traffic, on speech identification and, more generally, speech understanding. With this purpose, we perform listening experiments to evaluate the ability of individuals with normal hearing to detect words and interpret conversational speech in the presence of urban noise (e.g., street drilling, traffic jams). Our experiments confirm previous findings in different acoustic environments and demonstrate that speech identification is influenced by the similarity between the target speech and the masking noise also in urban scenarios. More specifically, we propose the use of the structural similarity index to quantify this similarity. Our analysis confirms that speech identification is more successful in presence of noise with tempo-spectral characteristics different from speech. Moreover, our results show that speech comprehension is not as challenging as word identification in urban sound environments that are characterised by the presence of severe noise. Indeed, our experiments demonstrate that speech comprehension can be fairly successful even in acoustic scenes where the ability to identify speech is highly reduced.

Word Spotting in Background Music: a Behavioural Study

“…In 1953, Colin Cherry [10] introduced the term Cocktail Party effect to indicate the ability of humans, in presence of several signals, to focus on specific stimuli while discarding others, and investigated potential causes influencing such ability. Several experiments confirmed Cherry's initial hypotheses, suggesting that some specific factors could actually help the mental ability of performing this segregation and extracting one specific signal from the environment, such as differences in the features of the competitive stimuli [11], the presence of a task [12], and the spatial location of the involved sound sources [13,14]. This paper focuses on task-driven attention in a simulated noisy scenario, in which speech perception has to cope with the presence of background music.…”

“…The purpose of this is to mitigate the influence of the subjects' fatigue on the results. The sounds were presented to the participants remotely via a web-interface, similarly to [11]. The participants were instructed to use headphones and perform the experiments in silent environments.…”

“…In this paper, we perform a similar analysis in an auditory context, by evaluating the impact of both static aspects (in terms of frequency spectrum similarities between the target and the distractors) and dynamics variations (in terms of temporal changes in the frequency spectrum) of different types of distractors. In [11], the authors carry out analogous experiments, focusing rather on speech identification and comprehension performance, without taking into account attentive aspects. In this work, instead, we focus on the attentive elements involved and, inspired by the concept of motion saliency, expressed in [24,25], we aim to analyse the role played by music dynamics in the distribution of attention.…”

Sound Environments

“…Marchegiani et al explored speech identification and comprehension in the urban soundscape [7]. Listening experiments were performed to evaluate the ability of individuals with normal hearing to detect words and interpret conversational speech in the presence of urban noise.…”