BCNDataset: Description and Analysis of an Annotated Night Urban Leisure Sound Dataset

Vidaña-Vila, Ester; Duboc, Letícia; Alsina-Pagès, Rosa Ma; Polls, Francesc; Vargas, Harold

doi:10.3390/su12198140

Cited by 6 publications

(6 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our previous experiences on (manually) labeling real-world acoustic datasets [52][53][54] taught us that assigning a tag to an acoustic sample is a time-consuming process: contrary to other types of datasets (e.g., images) in which a label can be assigned as soon as the sample is shown, in acoustic data labeling one has to wait for the whole acoustic record to be reproduced before assigning it a label. Typically, this is done with off-the-shelf software alternatives such as Audacity [55] that provide end-users with a spectrogram of the full audio record; thus, it becomes easier to visually identify those time frames in which something anomalous (i.e., potential events of interest) might be happening.…”

Section: Data Labelingmentioning

confidence: 99%

“…To mitigate the potential effects of class imbalance while training, we decided to add more training data and to apply data augmentation techniques to obtain more samples on the poorer classes. Additional data were obtained from the BCNDataset [52], which is a dataset containing real-word urban and leisure events recorded at night in Barcelona. As the BCNDataset was labeled differently than the Eixample Dataset, labels from both datasets were unified.…”

Section: Data Augmentationmentioning

confidence: 99%

See 1 more Smart Citation

Multilabel Acoustic Event Classification Using Real-World Urban Data and Physical Redundancy of Sensors

Vidaña-Vila

Navarro

Stowell

et al. 2021

Sensors

Self Cite

View full text Add to dashboard Cite

Many people living in urban environments nowadays are overexposed to noise, which results in adverse effects on their health. Thus, urban sound monitoring has emerged as a powerful tool that might enable public administrations to automatically identify and quantify noise pollution. Therefore, identifying multiple and simultaneous acoustic sources in these environments in a reliable and cost-effective way has emerged as a hot research topic. The purpose of this paper is to propose a two-stage classifier able to identify, in real time, a set of up to 21 urban acoustic events that may occur simultaneously (i.e., multilabel), taking advantage of physical redundancy in acoustic sensors from a wireless acoustic sensors network. The first stage of the proposed system consists of a multilabel deep neural network that makes a classification for each 4-s window. The second stage intelligently aggregates the classification results from the first stage of four neighboring nodes to determine the final classification result. Conducted experiments with real-world data and up to three different computing devices show that the system is able to provide classification results in less than 1 s and that it has good performance when classifying the most common events from the dataset. The results of this research may help civic organisations to obtain actionable noise monitoring information from automatic systems.

show abstract

Section: Data Labelingmentioning

confidence: 99%

Section: Data Augmentationmentioning

confidence: 99%

Multilabel Acoustic Event Classification Using Real-World Urban Data and Physical Redundancy of Sensors

Vidaña-Vila

Navarro

Stowell

et al. 2021

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…For this purpose, the audio files from the dataset are modified emulating the air channel and physical topology of the streets of Barcelona. Moreover, road traffic noise recorded in the city of Barcelona [68] has been randomly added to each of the audio files so each sensor perceives the event partially masked by traffic noise. This experiment shows how a ubiquitous sensor network can improve the classification results over individual sensors when perceiving the same acoustic data from different locations and masked with traffic noise.…”

Section: Experimental Evaluationmentioning

confidence: 99%

“…For this experiment, the average values in the Barcelona city center have been taken: temperature of 20 • C and 70% relative humidity. • Last, but not least, urban recordings of road traffic noise recorded on the city center of Barcelona [68] have been added (i.e., weighted sum) to each audio sample to further emulate a real-world environment. This is aimed to assess what happens when the background noise partially masks the acoustic event of interest.…”

mentioning

confidence: 99%

Low-Cost Distributed Acoustic Sensor Network for Real-Time Urban Sound Monitoring

Vidaña-Vila

Navarro²,

Borda-Fortuny³

et al. 2020

Electronics

Self Cite

View full text Add to dashboard Cite

Continuous exposure to urban noise has been found to be one of the major threats to citizens’ health. In this regard, several organizations are devoting huge efforts to designing new in-field systems to identify the acoustic sources of these threats to protect those citizens at risk. Typically, these prototype systems are composed of expensive components that limit their large-scale deployment and thus reduce the scope of their measurements. This paper aims to present a highly scalable low-cost distributed infrastructure that features a ubiquitous acoustic sensor network to monitor urban sounds. It takes advantage of (1) low-cost microphones deployed in a redundant topology to improve their individual performance when identifying the sound source, (2) a deep-learning algorithm for sound recognition, (3) a distributed data-processing middleware to reach consensus on the sound identification, and (4) a custom planar antenna with an almost isotropic radiation pattern for the proper node communication. This enables practitioners to acoustically populate urban spaces and provide a reliable view of noises occurring in real time. The city of Barcelona (Spain) and the UrbanSound8K dataset have been selected to analytically validate the proposed approach. Results obtained in laboratory tests endorse the feasibility of this proposal.

show abstract

“…Unfortunately, noise maps are not suitable tools to provide information on such critical issues because, as already mentioned, they refer to long-term average noise metrics, such as L den . Some studies have dealt with this topic, e.g., [11]; moreover, different road classifications were obtained when considering noise events only [12] or together with sound energy [13].…”

Section: Introductionmentioning

confidence: 99%

Methods for Noise Event Detection and Assessment of the Sonic Environment by the Harmonica Index

et al. 2021

Self Cite

View full text Add to dashboard Cite

Noise annoyance depends not only on sound energy, but also on other features, such as those in its spectrum (e.g., low frequency and/or tonal components), and, over time, amplitude fluctuations, such as those observed in road, rail, or aircraft noise passages. The larger these fluctuations, the more annoying a sound is generally perceived. Many algorithms have been implemented to quantify these fluctuations and identify noise events, either by looking at transients in the sound level time history, such as exceedances above a fixed or time adaptive threshold, or focusing on the hearing perception process of such events. In this paper, four criteria to detect sound were applied to the acoustic monitoring data collected in two urban areas, namely Andorra la Vella, Principality of Andorra, and Milan, Italy. At each site, the 1 s A-weighted short LAeq,1s time history, 10 min long, was available for each hour from 8:00 a.m. to 7:00 p.m. The resulting 92-time histories cover a reasonable range of urban environmental noise time patterns. The considered criteria to detect noise events are based on: (i) noise levels exceeding by +3 dB the continuous equivalent level LAeqT referred to the measurement time (T), criteria used in the definition of the Intermittency Ratio (IR) to detect noise events; (ii) noise levels exceeding by +3 dB the running continuous equivalent noise level; (iii) noise levels exceeding by +10 dB the 50th noise level percentile; (iv) progressive positive increments of noise levels greater than 10 dB from the event start time. Algorithms (iii) and (iv) appear suitable for notice-event detection; that is, those that (for their features) are clearly perceived and potentially annoy exposed people. The noise events detected by the above four algorithms were also evaluated by the available anomalous noise event detection (ANED) procedure to classify them as produced by road traffic noise or something else. Moreover, the assessment of the sonic environment by the Harmonica index was correlated with the single event level (SEL) of each event detected by the four algorithms. The threshold value of 8 for the Harmonica index, separating the “noisy” from the “very noisy” environments, corresponds to lower SEL levels for notice-events as identified by (iii) and (iv) algorithms (about 88–89 dB(A)) against those identified by (i) and (ii) criteria (92 dB(A)).

show abstract

BCNDataset: Description and Analysis of an Annotated Night Urban Leisure Sound Dataset

Cited by 6 publications

References 31 publications

Multilabel Acoustic Event Classification Using Real-World Urban Data and Physical Redundancy of Sensors

Multilabel Acoustic Event Classification Using Real-World Urban Data and Physical Redundancy of Sensors

Low-Cost Distributed Acoustic Sensor Network for Real-Time Urban Sound Monitoring

Methods for Noise Event Detection and Assessment of the Sonic Environment by the Harmonica Index

Contact Info

Product

Resources

About