Submit to this Journal Review for this Journal Propose a Special Issue

Article Menu

Share Help Cite Discuss in SciProfiles

Open AccessArticle

Peer-Review Record

Sustainable Soundscape Monitoring of Modified Psycho-Acoustic Annoyance Model with Edge Computing for 5G IoT Systems

Sustainability 2023, 15(13), 10016; https://doi.org/10.3390/su151310016

by Jaume Segura-Garcia^1,*

, Juan J. Pérez-Solano¹

, Santiago Felici-Castell¹

, José Montoya-Belmonte²

, Jesus Lopez-Ballester¹

and Juan Miguel Navarro²

Reviewer 1: Anonymous

Reviewer 2:

Pablo Salva-Garcia

Reviewer 3:

Marco Caniato

Sustainability 2023, 15(13), 10016; https://doi.org/10.3390/su151310016

Submission received: 21 April 2023 / Revised: 18 June 2023 / Accepted: 22 June 2023 / Published: 24 June 2023

(This article belongs to the Special Issue Innovations in Urban Acoustics and Sustainable Built Environment)

Round 1

Reviewer 1 Report

This paper proposed a system for calculating annoyance in real time by weighting the tonality of sounds based on Zwicker's psychoacoustic parameters.

I think it is meaningful to develop hardware that measures the level of the soundscape of the space by evaluating the sound field conditions of the outside space. In addition, in terms of sound work, it is reasonable to build an annoyance model that considers the cognitive limitations of raters (You & Jeon, 2008).

However, evaluation considering soundscape, rather than a simple noise model, is greatly influenced by audio-visual interactions, so the previously proposed soundscape model should be reflected in the algorithm. It seems that more functions should be added to the Raspberry Pi.

Please take a look at the following references and set up your measurement parameters:

Hong, J. Y., & Jeon, J. Y. (2015). Influence of urban contexts on soundscape perceptions: A structural equation modeling approach. Landscape and Urban Planning, 141.

Jeon, J. Y., & Jo, H. I. (2020). Effects of audio-visual interactions on soundscape and landscape perception and their influence on satisfaction with the urban environment. Building and Environment, 169.

Jo, H. I., & Jeon, J. Y. (2020). Effect of the appropriateness of sound environment on urban soundscape assessment. Building and environment, 179.

Li, Z., Ba, M., & Kang, J. (2021). Physiological indicators and subjective restorativeness with audio-visual interactions in urban soundscapes. Sustainable Cities and Society, 75.

Jeon, J. Y., Jo, H. I., & Lee, K. (2021). Potential restorative effects of urban soundscapes: Personality traits, temperament, and perceptions of VR urban environments. Landscape and Urban Planning, 214.

Jo, H. I., & Jeon, J. Y. (2021). Overall environmental assessment in urban parks: Modelling audio-visual interaction with a structural equation model based on soundscape and landscape indices. Building and Environment, 204.

Mitchell, A., Erfanian, M., Soelistyo, C., Oberman, T., Kang, J., Aldridge, R., ... & Aletta, F. (2022). Effects of Soundscape Complexity on Urban Noise Annoyance Ratings: A Large-Scale Online Listening Experiment. International Journal of Environmental Research and Public Health, 19(22).

Jo, H. I., Lee, K., & Jeon, J. Y. (2022). Effect of noise sensitivity on psychophysiological response through monoscopic 360 video and stereoscopic sound environment experience: a randomized control trial. Scientific Reports, 12(1).

Author Response

Reviewer 1

COMMENTS

This paper proposed a system for calculating annoyance in real-time by weighting the tonality of sounds based on Zwicker's psychoacoustic parameters.

Please take a look at the following references and set up your measurement parameters:

Hong, J. Y., & Jeon, J. Y. (2015). Influence of urban contexts on soundscape perceptions: A structural equation modeling approach. Landscape and Urban Planning, 141.

Jo, H. I., & Jeon, J. Y. (2020). Effect of the appropriateness of sound environment on urban soundscape assessment. Building and environment, 179.

Li, Z., Ba, M., & Kang, J. (2021). Physiological indicators and subjective restorativeness with audio-visual interactions in urban soundscapes. Sustainable Cities and Society, 75.

ANSWER:

Thank you for this valuable and interesting comment. Although it is true that the evaluation of the environment needs from all senses, our purpose in this paper is to assess the specific sound perception based on ISO 12913 and to explore sound perceptions due to psychoacoustic interactions. Here we don’t use cameras for image transmission. These aspects can be analysed in new research to be developed tackling audio-visual interaction and enabling aspects related to artificial intelligence.

We have added the following paragraph to the text (Line 32):

In addition to traditional acoustic parameters, other authors have explored other factors that affect soundscapes. A study by Hong and Jeon [15] concludes that the main function of a venue influences the evaluation of soundscapes and should be taken into account in its design. Other works highlight the impact of visual stimuli on the overall environmental assessment [16]. It should also be noted that several researchers have studied in the past how the sound environment produces physiological alterations [17-19], such as electrodermal activity, EEG and heart rate variations. However, in this work, only auditory stimuli have been taken into account for soundscape evaluation.

Author Response File: Author Response.pdf

Reviewer 2 Report

Summary:

This research proposes the use of psycho-acoustic annoyance models in Next Generation IoT systems for sustainable long-term monitoring of urban soundscapes, utilising two Edge Computing approaches. A modified version of Zwicker's model is used, introducing a term that accounts for the tonal component of captured sound. The effectiveness of these models has been confirmed through a measurement campaign in various sound environments on a university campus, revealing that tonality in a noisy environment increases subjective annoyance. While the research is both relevant and engaging, certain sections could benefit from revisions to improve the flow, structure, and overall comprehension for readers. With this in mind, I would like to suggest the following recommendations:

—-- Introduction —---

—----------------------------------------------

> Comment 1:

Line(s): 47 to 50

Where author say:

“In [19], the authors proposed a modification of the psycho-acoustic annoyance model by Zwicker’s in order to weigh the effect of tonal sounds in the subjective evaluation. The purpose in the current work is to modify the basic implementation established in [10] by adding the Tonality as a weighting parameter.”

The differences between this study and the one mentioned in [19] need to be made clearer. Also, the work done in [10] should be explained in simple terms to better understand the new feature that's been added.

> Comment 2:

Lines(s): 44

The implementation of the suggested models to various architectural designs, as detailed in Section 3, represents a significant contribution of this manuscript. However, the study falls short of emphasising this accomplishment. The introduction asserts strong statements, such as the one in line 44: "Nevertheless, all of them require high computational costs due to the complexity of the analysis and required signal processing." Yet, it does not elaborate on whether the proposed research intends to address this issue through the presented implementation.

In summary, the manuscript's title incorporates terms like "Sustainable", "Edge Computing", and "5G IoT systems", but these concepts are not adequately justified in the introduction. The fundamental questions that should be addressed include: Why is the proposed solution sustainable? What advantages does offloading computational processing to the edge provide? How is the proposed solution compatible with 5G IoT systems?

—-- 2. Materials and Methods —---

—----------------------------------------------

> Comment 3:

Line(s): 79 to 80

Authors say:

“In our implementation, it is used the method defined by Zwicker and Fastl which is the one on which the standard is based [10].”

This sentence could be misleading, as it might imply that the standards are established in [10]. However, [10] is actually a prior publication by the authors, which they aim to build upon by introducing tonality as a parameter.

This could benefit from being rephrased for clarity.

> Comment 4:

Line(s): 135 to 137

In Section 2.2 (Study of JNDs in sound quality metrics), the authors effectively introduce the concept of Just Noticeable Difference (JND), detailing the factors that can determine its magnitude. However, they reference a study concerning psycho-acoustic metrics for a refrigerator, without explaining the relevance of this study to the research at hand. The section concludes with the assertion that "psycho-acoustic annoyance depends on N, S, R, and F, but N is the most complex and contributes most significantly to JND." While it can be inferred that the perceptual value of loudness perceived by the human ear makes the greatest contribution to the JND, this point is not adequately tied to the research objective or context.

The clarity of this subsection could be improved by emphasising why the referenced study is important and how it is connected to the broader research presented in this work.

—-- 3. Architecture implementation —---

—----------------------------------------------

> Comment 5:

Line(s): 170 to 171

The paragraph is confusing, too many strategies, options, and figures that represent the same concept (If the case they are not the same they definitely need to be better clarified).

Please, It would be beneficial to homogenise the subsection by choosing either the word “strategy” or “option” and be consistent during the explanation of the methodology used.

—-- 4. Results and discussion —---

—----------------------------------------------

> Comment 6:

Line(s): -

After calibration, this section presents results from various scenarios on the UCAM Campus. The Psycho-Acoustic Annoyance (PA) generated is calculated using both the fundamental Zwicker model and a modified model incorporating tonality. While the results achieved are certainly interesting, the text doesn't sufficiently guide readers in understanding the context of what is deemed as high or low levels of psycho-acoustic annoyance. Moreover, the differences between the Zwicker model and the modified model are not adequately explained in the main body of the text. This oversight misses the opportunity to underline the unique contributions of this study, as it fails to provide a substantial rationale and discussion.

Further, the single commentary provided in Figure 8 states: "The three cases show a quasi-linear relationship." This leaves the reader with the task of analysing, interpreting, and finding a logical reason for the results. Ideally, the authors should guide the reader to reach this conclusion, rather than leaving it open-ended.

Please, It would be beneficial to expand and enhance the description of the results provided.

Author Response

Reviewer 2

COMMENTS

Summary:

—-- Introduction —---

—----------------------------------------------

> Comment 1:

Line(s): 47 to 50

Where author say:

ANSWER:

Thank you for your comment. We have clarified this issue in the new version of the manuscript. Actually, the reference [19] is the basis to enhance the Psychoacoustic (PA) Zwicker’s model by taking into account the tonality, considered as the starting point of our work. In reference [10], we already showed the details for the implementation of these complex algorithms for the PA model using low cost and generic hardware, in particular small board computers (SBC). But this prior study carried out in [10] did not include the tonality, which in practice requires a different thorough analysis and study in order to be implemented on these kinds of platforms. This is caused because these algorithms exceed the computational resources of these generic SBC, and they are assisted in our approach on edge computing in a 5G paradigm, as we describe in this article.

We have added the following sentence in blue to the original text (with new numbering in the references)(Line 63):

In [26], the authors proposed a modification of the psycho-acoustic annoyance model by Zwicker's in order to weigh the effect of tonal sounds in the subjective evaluation. The purpose of the current work is to modify the basic implementation established in [10] by adding the Tonality as a weighting parameter, following the implementation by Aures (from [27]) and an improvement by Estreder et al. [25].

> Comment 2:

Lines(s): 44

ANSWER:

Thank you for your comment. We notice that the paper requires to emphasise these issues. We have improved the wording highlighting the high computational requirements needed by these algorithms related to the PA model implementation. We have introduced some values about these computation costs for these algorithms, as a summarized review of paper [10] (as it was suggested before in point 1), in order to highlight, justify and motivate the different aspects covered within the paper, from edge computing and the 5G paradigm.

In our current implementation, for the new algorithm, we performed tests in a Raspberry 4B+ with 4GB RAM, and the results are shown in the following table:

	N	R	F	T
Mean Time (s) per 1 sec.	0.0676	0.4325	0.4907	0.0400
Time Std (s) per 1sec.	0.0337	0.0355	0.0451	0.0015

Once we have defined these tough requirements to process these Zwicker’s PA and Tonality modified PA models, we define several strategies based on the flexibility of the 5G architecture, where virtualization is a key point. In this context and based on this technique, we show different alternatives to split the PA models using tonality in different pieces that can be managed at different points of the 5G architecture. These alternatives are carried out at the edge, named as an edge computing approach, to avoid excessive delays sending data and information to the core of the network for the same goal. Since all these considerations are justified in this paper to adapt this model in this scenario and because this approach facilitates this implementation in a distributed way, based on functional splitting features for this 5G context, thus this new scenario can be considered “sustainable” in a generic way, since this approach allow to meet this requirements with a minimum impact.

—-- 2. Materials and Methods —---

—----------------------------------------------

> Comment 3:

Line(s): 79 to 80

Authors say:

“In our implementation, it is used the method defined by Zwicker and Fastl which is the one on which the standard is based [10].”

This could benefit from being rephrased for clarity.

ANSWER:

The reviewer is correct. We have improved the wording to clarify it. Actually, as we stated before in this arguments, reference [10] just focus on an implementation of the complex PA model that it did not include tonality.

> Comment 4:

Line(s): 135 to 137

The clarity of this subsection could be improved by emphasising why the referenced study is important and how it is connected to the broader research presented in this work.

ANSWER:

Thank you for your comment. We improved the wording to explain this in a better way. In practice, the conclusion is correct, since N in terms of the JND has the main contribution, however we should conclude this better, by emphasising our results in this line.

This paragraph has been improved as follows:

This study is relevant for our study as is everyday sounds the main contribution is generally made in terms of loudness and sharpness, as shown in [34], and also for the JNDs. Our hypothesis here is that the perception thresholds in the referred study [33] will be the same for everyday sounds.

—-- 3. Architecture implementation —---

—----------------------------------------------

> Comment 5:

Line(s): 170 to 171

The paragraph is confusing, too many strategies, options, and figures that represent the same concept (If the case they are not the same they definitely need to be better clarified).

Please, It would be beneficial to homogenise the subsection by choosing either the word “strategy” or “option” and be consistent during the explanation of the methodology used.

ANSWER:

Thank you for your input. We have improved this section accordingly avoiding different terms for the same meaning. In our case, we have remove the word option in the second sentence, as follows:

“... taking into account the two strategies explained in this section, both have been implemented.”

—-- 4. Results and discussion —---

—----------------------------------------------

> Comment 6:

Line(s): -

Please, It would be beneficial to expand and enhance the description of the results provided.

ANSWER:

Thank you for your comment. We agree with it and we have improved these issues trying to ease the reading. Also, we have enhanced this wording providing more details in order to highlight our contribution and the goals of this manuscript.

We have improved the main text by introducing some explanations, as follows:

“The three cases show a quasi-linear relationship, as the main contribution is made with the loudness, as shown in [34], although there are slight differences generated by tonality, roughness, and fluctuation strength, as shown in Figures 5, 6, and 7.”

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper could be of some interest but some major revisions are needed

1) in the introduction you speak about statistics and sound sensitivity. I THINK this is for tipically developed individuals. No discussion (even brief discussion) is included for neuro divergent ones. Accordingly to WHO data they are at about 18% of the general population. I think at least a brief discussion is due when speaking about sound sensitivity... during the discussion of the JND you also say that who has alower JND is more sensitive etc... so please inlcude a brief discussion. Cite relevant references like (as example)
https://doi.org/10.1016/j.egyr.2022.01.011
https://doi.org/10.1016/j.egyr.2022.01.009
https://doi.org/10.3390/app11093942

2) there some organization issues: for example in the materials and methods you describe only hot to design the sensors but NOT how to apply and use them. This is in the results and discussion BUT belongs to materials and methods. Please move it

3) you compare your results with the literature one, modified with tonality. Ok. Good. Then you present results. Did you asked to real people? Did they confirm your results?

4) ehtical issue: you measured and performed test using humans. Did you go through an ethical procedure? did these persons know that they are measured? and tested? if they wanted, Could they not be included in your experiment?

5) in the last figure, we clearly see that there is a strong link between what calculated and what measured in dB(A). a Linear link, so the easiest one. So, WHY do I have to proceed further with this research if aI have evidences that this is linear with what I am directly measuting and I do not have to perfom all the calculations? Please explain

Author Response

Reviewer 3

COMMENTS

The paper could be of some interest but some major revisions are needed

1) in the introduction you speak about statistics and sound sensitivity. I THINK this is for typically developed individuals. No discussion (even brief discussion) is included for neuro divergent ones. Accordingly to WHO data they are at about 18% of the general population. I think at least a brief discussion is due when speaking about sound sensitivity... during the discussion of the JND you also say that who has a lower JND is more sensitive etc... so please include a brief discussion. Cite relevant references like (as example)

https://doi.org/10.1016/j.egyr.2022.01.011

https://doi.org/10.1016/j.egyr.2022.01.009

https://doi.org/10.3390/app11093942

ANSWER:

Thank you for your comment. We have improved the wording in order to introduce the sound sensitivity in the JND discussion. Also, we have included your relevant references in the new version of the manuscript.

Line 55:

It is important to note that this type of models are typically obtained through experiments with healthy hearing people with average psychosocial behaviour. Some authors have found that the sound environment has a great impact on comfort in people with autism spectrum disorder [22, 23]. Autistic individuals with hearing impairment or hypersensitivity to sound need additional acoustic requirements in indoors facilities [24]. These issues are of interest to be addressed in future research by adding other psychosocial aspects to the classical models.

Line 146:

There is no empirical evidence that individuals with developmental disorders have physiological differences in the auditory system [35], but nevertheless, it has been shown that these individuals, for example with autism spectrum disorder, appear more sensitive to sounds perceived as annoying [36].

—-

[22] Caniato, M.; Zaniboni, L.; Marzi, A.; Gasparella, A. Evaluation of the main sensitivity drivers in relation to indoor comfort for individuals with autism spectrum disorder. Part 1: Investigation methodology and general results. Energy Reports, 2022, 8, 1907–192

[23] Caniato, M.; Zaniboni, L.; Marzi, A.; Gasparella, A. Evaluation of the main sensitivity drivers in relation to indoor comfort for individuals with autism spectrum disorder. Part 2: Influence of age, co-morbidities, gender and type of respondent on the stress caused by specific environmental stimuli. Energy Reports 2022, 8, 2989–3001.

[24] Bettarello, F.; Caniato, M.; Scavuzzo, G.; Gasparella, A. Indoor Acoustic Requirements for Autism-Friendly Spaces. Applied Sciences 2021, 11.

[35] Stiegler, L. N., & Davis, R. (2010). Understanding sound sensitivity in individuals with autism spectrum disorders. Focus on Autism and Other Developmental Disabilities, 25(2), 67-75.

[36] Lucker, J. R. (2013). Auditory hypersensitivity in children with autism spectrum disorders. Focus on Autism and Other Developmental Disabilities, 28(3), 184-191.

Article Menu

Sustainable Soundscape Monitoring of Modified Psycho-Acoustic Annoyance Model with Edge Computing for 5G IoT Systems

Further Information

Guidelines

MDPI Initiatives

Follow MDPI