An example of a digital engagement platform for large scale commu- nity engagement using auralization.

Alex Southern 1 AECOM Ltd 120 Bothwell Street, Glasgow, G2 7JS, UK Brian Bulnes 2 & Alan Oldfield 3 AECOM Ltd 5090 Explorer Drive, Suite 1000, Mississauga, ON, L4W 4X6, Canada

ABSTRACT The global COVID-19 pandemic has resulted in social distance restrictions that have limited the ability for transport authorities to undertake in-person community engagement activities and consult on proposed local infrastructure developments. Potential increases in noise levels or change in acoustic environment can often be a key concern for residents living close to a proposed development. This paper documents the approach taken to engage with local stakeholders regarding a proposed new light rail metro line in Toronto, Ontario, Canada, using an innovative online web-based aurali- zation tool. The tool allows the existing trains and planned new metro trains to be compared inter- actively in an environmental context and with and without acoustic mitigation interventions. The paper discusses the benefits, challenges and limitations associated with the delivery method and pro- vides an overview of the auralization approach of the proposed new metro line.

1. INTRODUCTION

The east segment of the proposed Ontario Line project in Toronto will operate in the existing GO Union Station and Lakeshore East rail corridors (referred to as the Lakeshore East Joint Corri- dor). The north segment will include both a tunneled portion and raised guideways. The transporta- tion authority was looking to enhance communications with community stakeholders during public consultation for the Lakeshore East Joint Corridor and North segments.

The proposed project will improve transportation links for the city and will increase the number of train movements within the Lakeshore East Joint Corridor. The project involves shifting existing GO train railway lines to accommodate the new Ontario Line. The total number of train movements is set to increase, but the use of quieter, shorter Ontario Line trains and targeted placement of noise barriers, means that the new trains may be considered less intrusive than those already operating and the introduction of noise barriers in some areas has the potential to improve the existing situa- tion further and can result in a perceived net beneficial or negligible change for many in terms of noise. Ontario Line North will involve creating a new rail corridor for the Ontario Line, including tunneling and raised guideways.

1 alex.southern@aecom.com

2 brian.bulnes@aecom.com

3 alan.oldfield@aecom.com

worm 2022

AECOM’s technical Noise and Vibration study report, prepared as part of the Lakeshore East Joint Corridor early works impact assessment, includes tables of numerical results and noise con- tour figures, which help to explain the noise impacts of the proposed Ontario Line project. Such materials are considered standard approach in the industry to reporting and communicating planned changes. However, understanding how the different factors described and what the relative change sounds-like compared to that currently experienced is challenging to do when reading a report or verbal discussion.

In addition to this, the social distancing restrictions introduced following the COVID-19 pan- demic had made the usual in-person consultation event practice less reliable as a method of con- necting with communities and discussing the project plans. Web-based presentation of materials was a natural consideration to address this. To help address concerns about changes in noise, AECOM delivered an interactive Immersive Sound Studio website to present sound demonstrations at a number of key locations along the route.

The completed sound demonstrations combined environmental sound propagation modelling, spatial audio recordings, calibrated sound level measurements, audio processing, computer gener- ated visual renderings, animation, video editing, web platform development, and stakeholder en- gagement expertise. Whilst sound demonstrations have been prepared for infrastructure projects in the past, this project is the first of its kind in Canada for major transport infrastructure, with a dedi- cated website with interactive audio and video functionality. This paper documents the technical presentation of the sound demonstrations and the benefits, challenges and limitations of using a web-based platform for stakeholder engagement. 2. SOUND DEMONSTRATIONS When communicating with non-technical stakeholders the term auralization is difficult to interpret without additional explanation and therefore the term sound demonstration is often used as it is more self-explanatory. The term sound demonstration can also be used more broadly, for example it could include demonstrating the relative difference in decibels levels without any spatialized sound com- ponent.

Auralization is long established as a potential design and communication tool in architectural acoustic design practice [1][2]. In more recent years the techniques have been adapted to inform the design and/or communicate planned developments within communities i.e. [3]. The application is typically used where transportation, industrial and commercial sources of sound have the potential to cause annoyance or disturbance to residents or other sensitive receptors such as schools, hospitals and heritage sites. While planning processes are not typically mandating the use of auralization when designing or communicating plans with stakeholders, there is generally a broad requirement on the developer to make noise impact assessment information accessible and understandable; and in our experience planning authorities have been known to recommend sound demonstrations as examples of best practice when communicating predicted noise impacts at public consultation events with non- technical stakeholders.

While there is no single standard method to creating sound demonstration content for public con- sultation, the goal of the consultant always remains to communicate changes or differences between scenarios and provide context to the listener so they may arrive at an informed opinion about how they feel about the plans. The context can be provided in various ways such as, visually using maps or virtual models, photos, videos or computer-generated imagery, verifiable viewpoints or compo- sites, and through audio using relatable sounds in some way.

worm 2022

The presented sound demonstrations should therefore be used to compliment and not contradict the standard methods of reporting predicted noise impacts. This means that a sound demonstration used at consultation should be prepared and engineered with reference to the same methods used to measure and predict sound levels. This is an important part of preparing a credible sound demonstra- tion and it is what differentiates a sound demonstration as a reliable and informative decision-making tool from a creative/artistic sound designer’s impression of how a development might sound.

Sound demonstrations presented in-person can be delivered using a calibrated system where the headphone level has been measured and corrected to produce a sound level representative of the in- tended absolute sound level. However, in the context of the Ontario Line project, restrictions imposed by social distancing rules led to the development of a web-based platform with a view to facilitating engagement with stakeholders without delaying the project program or comprising on the quality of the information being provided and instead enhancing it. 3. WEB-BASED SOUND DEMONSTRATION TOOL A sound demonstration web-tool (see Figure 1) was developed comprising the following technical features:  Interactive map of the project, highlighting:

 project route; and  sound demonstration locations.  Audio/Video (AV) player component:

 capable of switching between multiple AV feeds instantly i.e. existing, with project and with

project and mitigation;  able to jump forward and backwards to key moments while keeping audio/video streams in

sync;  works in all major internet browsers running on any common device; and  designed to work natively in the browser and does not require the user to install third party

plugins.  Designed to be accessible and offer blind, partially sighted, various types of colorblind stakehold-

ers the ability to use the site in its entirety.  User interface designed to adapted to mobile viewports while maintaining accessibility features.

4. PROJECT DEMO LOCATIONS

When determining where along the project route the sound demonstrations should be located, it was important to take account of the questions being asked by residents earlier in the engagement process such as:  How much noise will come from passing trains once the Ontario Line opens? And  How will this compare to the sound of today’s GO trains, currently passing through the neighbor-

hood?

worm 2022

worm 2022

The project sound demonstration locations were selected primarily to represent the closest sensi-

Figure 1: Labelled screenshot of the web-tool in an internet browser.

tive receptors and closest public amenity areas to the route. These were identified as being of most benefit to the stakeholders as locations further from the rail corridor are less likely to be disturbed by rail vehicle movements due to the prevailing ambient sound levels. Being generally on publicly ac- cessible land also means that local stakeholders can contextualize the sound demonstrations based on their experience of real current conditions at the sound demonstration locations. One location (Lo- cation 12) was selected to be indoors to represent a residential basement in proximity to the tunneled segment. At this location, a visual rendering of a basement scene accompanies the sound demonstra- tion to help ‘place’ the listener in the scene.

The table below lists the locations and viewpoints selected at each, generally looking toward the rail corridor or the location it was planned for. Table 1 Table of photos showing the existing situation at all 12 locations. First six are Lakeshore East Joint Corridor section and the second set of six are in the North section.

Location 1 Location 2 Location 3

Location 4 Location 5 Location 6

Accessibility tab pop-out Cy Video viewport Map of location —F with respect toroute Jump to key events inthe demo. LY cunt enenseaton = Pc stream selector buttons Navigation Without Project a With Project With Project & Mit al User feedback option. Play/Pause

worm 2022

Location 7 Location 8 Location 9

Pre

Location 7 Location 8

5. SOUND DEMO PREPARATION OVERVIEW

The following tables provide an overview of the elements that make up the sound demonstra- tions at each location.

Sound Demonstration Scenario

Indoors ground- borne vi- bration Mit-

Rail Corri- dor Exists -

Specific Contextual

New Ele-

Barrier Correction

Loc. Ambient Sound Description

vated Guideway

GO Train

Event

igated &

Pass-by

Unmiti-

gated

1 Distant suburban traffic, streetcars, birds, resident activ- ity, occasional car pass-bys Yes Yes Car No No

Lakeshore East

2 Distant traffic, birds, park users, aircraft above Yes Yes - No No

3 Distant suburban traffic, birds, resident activity, occa- sional car pass-bys Yes Yes Car No No

4 Distant traffic, resident activity, birds, park users Yes Yes - No No

5 Distant traffic, resident activity, vehicle maneuverer in area Yes Yes Aircraft No No

6 Distant traffic, occasional car pass-bys, resident activity, birds Yes Yes - No No

7 Constant traffic No Yes - Yes No 8 Distant traffic, birds, occasional vehicle pass-bys. No Yes - No No

9 Distant traffic, industry related noises, occasional parking lot movement. No No - No No

North

10 Constant traffic, vehicle pass-bys towards science cen- tre. No No - Yes No

11 Distant highspeed traffic No No - Yes No 12 Air conditioning system No No TV Audio No Yes The sequence of each sound demonstration was generally as follows:

A. The ambient sound of existing acoustic environment; B. A train approaches / passes / departs;

i. the GO Train if relevant

ii. the Ontario Line train C. A specific contextual event occurs if relevant. The demonstrations therefore allows stakeholders to understand how the Ontario Line trains will sound in the context of the GO Trains, existing traffic, and the local environment which are already familiar to them.

Where a barrier is planned, the effect of this barrier can be switched on or off for locations L1 to L6. This is because there is value in being able to experience the existing GO train pass-bys with- out and with the proposed barrier. As a brand new rail corridor is planned at L7 to L12 there is no or little value to presenting a “With Project & No Barrier” scenario and so only “With Project & Barrier” is available where barriers are planned. 5.1. General approach To prepare the sequences previously outlined, the following datasets and actions were performed for each chosen location:

 Ambient sound and video recordings and sound level monitoring;  Existing GO train video and sound recordings and sound level monitoring;  Representative Ontario Line train recordings at a range of reference distances;  3D sound propagation modelling and filter determination;  Sound and video clip analysis, selection and editing; and  CGI scene matching – to allow new Ontario Line train and infrastructure to be animated. We provide some additional details of the elements in the following subsections. Ambient sound, GO train and event recordings

These recordings were undertaken using a combination of a first order ambisonic i.e. [4][5] mi- crophone, class 1 sound level meter and video camera. The three devices were generally used to record for 1 to 2 hours at each location simultaneously.

The aim of the ambient recording was to capture the typical day-time ambient acoustic-environ- ment generally avoiding events that don’t happen often or would not be considered to be happening all the time by residents. This is important because selecting higher ambient sound level clips could be viewed as an attempt to mask sound from the new trains in the sound demonstration, while select- ing uncharacteristically the quietest ambient sound level periods could lead stakeholders to perceive the Project as being more impactful than it is. How the ambient sound clip is chosen is important to the robustness of the sound demonstration. In this work, a shortlist of 30 s, 40 s and 50 s clips of the typical ambient sound were selected by analyzing the mode, mean and median L Aeq,T from the 1 s, 30 s, 40 s and 50 s sliding windows. This allowed a typical level to be assigned for the location so that all 30 s, 40 s and 50 s clips that match that level within +/- x dB could be identified and reviewed for selection as the ambient sound and video clip. Some clips were not suitable because something visually distracting was going on, like people walking by. Where GO trains were present, these were also recorded.. The prediction methodology used for the noise impact assessment and the sound demonstrations was a prediction method for railway noise from the United States Federal Transit Administration’s (FTA) Transit Noise and Vibration Impact Assessment Manual, with implementation in the CadnaA acoustic software package. The predictions adopted assumptions about the worst case GO train, being of type Diesel and the longest length in

worm 2022

operation, based on the configuration of locomotive and car units. The GO train recordings used in the final demonstration represent various configurations, but did include a worst case Diesel GO train where possible.

While capturing the ambient and GO train recordings, notes were made about other unplanned events that were captured, such as car pass-bys and aircraft flyovers. Where possible additional con- textualizing events were included in the sound demonstrations, like the ambient sound clips, these were edited only to remove wind noise across the microphone if needed. Along with ambient clips, these provide a reference against which the Ontario Line train movements and barrier performance can be understood. Representative light rail (proxy) recordings for Ontario Line

An obvious challenge for a sound demonstration for a future condition is always: How can we auralize a noise source that doesn’t yet exist? In this case, the Ontario Line vehicles were not in operation. We reviewed several existing light rail systems and undertook calibrated sound level mon- itoring and ambisonic audio recordings of representative vehicle operations as a basis for acting as a proxy for the Ontario Line vehicle pass-by. The representative operations were selected based on the similarity in vehicle type, speed and local topography to the scenarios being developed. The vehicle pass-by sounds were then corrected to account for differences between the train pass-by recording distance and the sound demo location listening distance from the planned route. The 3D sound prop- agation model results at the sound demonstration locations included corrections that generally ad- dressed ground effects, air absorption, level reduction due geometric spreading and screening effects. 3D sound propagation modelling to determine frequency dependent corrections.

The 3D sound propagation model was setup in CadnaA, a noise modelling package which imple- ments both the ISO 9613-2 [7] noise prediction method, and a module for FTA railway modelling. This was used as the basis for reporting the noise impact in line with the standards and guidance relevant to the region. The baseline situation was modelled and included the existing GO train move- ments. The noise model incorporated ground topography, track elevations (including any relevant retain- ing walls for the railway), ground absorption, train speeds, throttle settings, volumes, and train types and configurations. Noise barriers were placed within the noise model for the purposes of meeting noise criteria for the project, and the overall barrier reductions were used as the basis for sound demo “with barrier” reductions where applicable. Notes on Editing and Post-Production

The ambisonic recordings were decoded to a virtual loudspeaker array and then rendered to a binaural reproduction of the recorded soundfield looking in the same direction as the camera view- point i.e. [6]. In general, the intention is to do as little sound editing and post-production to the re- cordings as possible, primarily because there are limits to what can be achieved in post, and instead the focus should be on getting the right high quality recordings in the first instance. In addition, this also reduces the amount of technical explanation and justification needed. Editing is typically under- taken on the binaural reproduction as this allows the editor to understand what the end user will hear.

Once the various binaural clips being used to build a sound demonstration have been selected (to minimize the editing needed), any outstanding required editing is performed usually to remove or isolate extraneous sounds, wind noise and atypical events. For example, removing bird sounds from

worm 2022

the proxy train recording was needed because they are contributed by the location specific ambient recordings – however the species may not be a match either. Removing birds from train pass-bys is possible due to the clear difference in temporal and frequency structure.

The Ontario Line trains were shorter than the GO Train and longer than the proxy trains. Therefore, as the visual animations show the planned length of the Ontario Line trains, the proxy audio record- ings required some work to make the train pass-by sound longer. The repetitive nature of a train pass- by is somewhat forgiving to the layering and delaying copies of the original train pass-by recording with careful editing. The resulting proxy pass-by is then manually synchronized with the visual ani- mation until it looks and sounds natural. 5.2. Guideway (Bridge) Locations The proxy train recordings used to represent the Ontario Line trains were recorded on ballasted track and therefore a correction was required to account for the different track construction planned for new guideways. A correction spectrum was determined using modelled data for ballast track and the elevated guideway structure. This was applied as a constant correction such that sound radiating from the bridge structure is assumed to be unaffected by the trains position on the bridge. 5.3. Location 11 - Minton Place Portal At this location, the railway exits a tunnel portal onto a bridge to cross over a highway in a ravine. The sound demonstration viewpoint looks across the ravine from an elevated position higher than the tunnel. The approach to modelling and determining the correction filters for this was different because a global ballast to bridge correction was needed as described in the previous subsection, but also the partial to no screening transition of the tunnel for a train emerging needed to be considered too. A time-varying level correction was calculated by using a dynamic model of the Ontario Line train emerging from the tunnel. Reverberation from within the tunnel could be scoped out because it was established that the distance and orientation from the tunnel portal, and the prevailing high ambient sound levels due the highway traffic would make it inaudible or negligible to the overall sound at the sound demonstration location. The resulting time-frequency dependent amplitude correction was ap- plied to the proxy Ontario Line train recording. 5.4. Location 12 – Muriel Avenue Basement At this location, trains are expected to travel in a subway tunnel beneath and adjacent to homes, which has the potential for producing ground-borne noise within dwellings. A virtual basement room was created to be representative of a typical basement along this street, for both the purposes of creating an immersive visual for the scene and generating a room response to be applied to ground-borne noise audio within EASE acoustic software. A vibration recording of a proxy light rail subway system was used to create an audio signal representing ground-borne noise. The final ground-borne noise audio with room effect applied was calibrated to meet modelled noise levels from the environmental as- sessment for this segment. COMPLETED DEMONSTRATION EXAMPLES 5.5. Location 3

worm 2022

Without Project

With Project & No Barrier

worm 2022

With Project & Barrier

5.6. Location 10 Without Project

With Project

worm 2022

5.7. Location 11 Without Project (Existing)

With Project – Tree in right foreground is faded to make guideway with train over highway visible.

6. WEB PLATFORM LIMITATIONS 6.1. Web-tool testing Initial prototypes of the web-tool worked reliably on the desktop hardware and ma- jor browsers, such as Google Chrome and Microsoft Edge (Chromium). Videos would be present on the local machine, which made loading of the videos in- stantaneous. Setting up a public webserver in the region where the website would be accessed was therefore an important step towards good device and user experience testing. This testing was un- dertaken continuously through development; first with video placeholders and later with drafts of the final videos. Efforts were made to test the tool on Android, iOS, Windows, Mac with a range of

Ona

popular modern browsers such as Chrome, Edge, Firefox and Safari. Although there was no inten- tion to support Internet Explorer due to its limitations, it was shown to be usable.

The resolution of the video was selected to favor high quality when viewed on mobile devices and reasonable quality when viewed on a desktop/laptop device. The choice was made following user testing where it revealed that slower connection speeds could lead to waiting times of over 5 seconds for full HD videos and it was not clear to the user that they were waiting on the assets downloading to the server. This could have potentially led the user to believe the tool was broken.

User Interface

Care must be taken not to replace a difficult to understand technical subject matter with a different one. That is, without any prior experience of sound demonstrations, a web-tool designed to present sound demonstrations can itself complicate and be a barrier to the stakeholders’ experience and en- gagement with the project. The site’s accessibility requirements were an important guiding factor in determining the user interface functionality as was the ability of the website to adapt to different screen sizes and aspect ratios that might be used. For accessibility it needs to be possible for e-readers to cycle through button options and not rely on the visual layout to inform their function.

We found that when trying to balance these factors that they severely limited our ability to use the initially desired layout and functionality. For example, video players with embedded controls look cleaner and take up less space on the page, but on mobile devices require small un-labelled buttons and therefore a potentially different interface entirely and one that is perhaps not friendly for the visually impaired.

One way to alleviate these issues was to utilize an introductory video to the sound demonstration website that explained, with animated graphics, what to expect on the site and how to use the web- tool.

6.2. Hardware quality control A technical challenge raised early in the project was how to address the lack of control we have of the end user’s audio set up. Users accessing the sound demonstration website would have varying levels of hardware availability and quality, (e.g. loudspeakers not headphones, volume settings). These factors can be avoided when utilizing a purpose-built demonstration facility or public presen- tation at consultation events, but that level of control doesn’t exist with a web-based tool.

To help address this, an introductory video to the project and sound demonstration website also invited the listener to set the volume of the narrator to a natural conversational level. While the se- lected volume will vary from user to user this approach would avoid the volume being set excessively loud or quiet during the sound demonstration. In formal listening tests on a single device suggested the volume setting could be as much as +/- 10dB, this range of course equates to a perceptual halving or doubling in perceived loudness. As the focus of the sound demonstration is on the predicted relative change in sound level and character, the variation in absolute sound levels observed from user to user was not a direct concern for this project. 6.3. Stakeholder connectivity A stakeholder’s connection quality is an important factor in the end user experience and can po- tentially frustrate an individual’s consultation experience. Stakeholders that have not had a satisfac- tory experience, e.g. where the web-tool would not work reliably due video buffering issues, are less likely to leave the site supportive of the project because they could not access the information they wanted to. Unfortunately, as the cause of the connection issues is down to the individual’s connection

worm 2022

quality or Internet Service Provider, there is little that can be done about this other than encourage people with technical issues to get in touch so that alternative engagement arrangements can be dis- cussed as appropriate. 7. SUMMARY

This paper has provided a high-level documentation of the sound demonstration web-tool and content presented as part of the public consultation process for the Ontario Line project in Toronto, Ontario, Canada. The sound demonstration approach to communicating predicted noise impacts with stakeholders is a first in Canada for a significant major infrastructure project. It is also under- stood that the interactive web-tool delivery of the sound demonstrations approach taken was also the first of its kind on an infrastructure project like this one.

In the future, a combined approach could be taken, with a web-based sound demonstration made available prior to in-person events. At the in-person events, an array of devices and level calibrated headphones may be offered to stakeholders wanting increased accuracy in the renderings or to those having connection issues at home and for people who simply prefer in-person events and have con- cerns about noise.

To identify system issues or stakeholder connection issues more pro-actively, a useful addition to the web-tool might have been to ask the stakeholder to rate the quality or technical reliability of the web-tool functionality. Poor scores could have then been followed up with a request for more infor- mation and possibly, eventually, to receive a call from the project stakeholder engagement team.

Overall, the web-tool has made sound demonstrations of predicted relative changes in sound level and sound character more easily accessible and available to a wider demographic on demand. The benefits provided by making project information more convenient to access and easier to understand and interpret are generally thought to outweigh the possible limitations that can arise for some stake- holders with poor internet connections, no or poor quality audio hardware, or those listening at overly high or low volumes. 8. REFERENCES 1. M. Kleiner, B.-I. Dalenbäck, and P. Svensson, “Auralization – An overview,” J. Audio Eng.

Soc. 41, 861–874 (1993). 2. M. Vorländer. Auralization : Fundamentals of Acoustics, Modelling, Simulation, Algorithms

and Acoustic Virtual Reality. Springer-Verlag, Berlin, 2008. 335 pp. ISBN: 978-3-540-48829-3 3. Stienen, Jonas and Vorlaender, Michael. Auralization of Urban Environments – Concepts to-

wards New Applications. Euronoise 2015: 10th European Congress and Exposition on Noise Control Engineering, Maastricht, Belgium (2015). 4. Michael A. Gerzon, Periphony: With-Height Sound Reproduction. J. Audio Eng. Soc. 1973,

21(1):2–10. 5. Christian Nachbar, Franz Zotter, Etienne Deleflie, and Alois Sontacchi: AmbiX - A Suggested

Ambisonics Format Ambisonics Symposium 2011 , Lexington (KY) 2011 6. Noisternig, Markus. et al. "A 3D Ambisonic Based Binaural Sound Reproduction System." Pre-

sented at 24th AES International Conference: Multichannel Audio , The New Reality, Alberta, June 2003. 7. ISO 9613-2:1996 Acoustics — Attenuation of sound during propagation outdoors — Part 2: Gen-

eral method of calculation. Publication date : 1996-12

worm 2022