Establishing the foundation
The initial 1:1 environment, shader architecture, modelling standards, artist tools, and production pipeline were created while the Toronto foreground expanded.
Case Study / Digital Twin
Toronto Waterfront Digital Twin
Building the technical and artistic foundation for a high-fidelity, cloud-streamed recreation of approximately 2 km² of Toronto at 1:1 scale.
RoleCTO, Technical Director & Lead Technical Artist
ScaleApprox. 2 km²
PeriodEarly 2020 to 2023
TechnologyUnreal Engine 4 and 5
The project
From one development to a city-scale platform
Production began in early 2020 with The HUB at 30 Bay Street and its immediate surroundings. It later grew into an interactive R&D platform combining a hand-modelled foreground, real urban information, multiplayer, cinematic presentation, and browser delivery.
I was the first artist and technical contributor. As the project expanded, I moved into technical direction, production leadership, QA, tooling, and responsibility for the final visual and technical quality bar.
01The HUB
02Surrounding blocks
03Waterfront expansion
042 km² R&D platform
Project overview
A continuous real-time city environment
The environment was built for exploration, data visualization, cinematic presentation, and live browser delivery, not only for isolated architectural images.
Real-time application and cinematic overview. Unreal Engine 5.
Shader-driven architecture
High visual detail without modelling every detail
A city-scale environment cannot treat every room, window, and facade element as unique geometry. I developed a material system that moved much of the visible complexity into shaders while keeping building meshes deliberately efficient.
Engine evolution
Adapting a live production as Unreal changed underneath it
The project began in Unreal Engine 4.26 and 4.27, then moved into Unreal Engine 5. Parts of the environment were rebuilt or adapted around the engine's evolving large-world, rendering, and modelling toolsets.
Rebuilding for scale
World Partition, Nanite, and Lumen changed how the project handled large-world organization, geometry, lighting, and the visual quality target.
Optimizing the urban fabric
Roads, sidewalks, land plots, plazas, and highway structures were inspected and refined directly with Unreal Modeling Tools in lit and wireframe views.
A surviving May 2023 recording documents infrastructure optimization across the waterfront in lit and wireframe views.
Procedural infrastructure
Gardiner Expressway pillars with Houdini Engine
I developed a custom Houdini Digital Asset to reduce repetitive infrastructure work while preserving project-specific control. Exposed parameters allowed pillar dimensions and variations to be adjusted directly inside Unreal Engine 4.
The system was designed around constraints including terrain slope, highway curvature, and ground clearance.
This was one component of a broader modular and procedural R&D effort. The Toronto foreground remained primarily hand modelled and should not be described as a procedurally generated city.
Custom Houdini Digital Asset running inside Unreal Engine 4.
Geospatial workflow
Reference alignment evolved with the project
The 1:1 foreground used LiDAR, photogrammetry, OpenStreetMap, satellite imagery, Google Earth reference, and municipal datasets. The workflow changed repeatedly as the environment grew and better sources became available.
01 / EARLYOSM 3DInitial massing, location, and urban context.
02 / REFERENCEPhotogrammetryVisual and spatial reference during foreground production.
03 / MUNICIPALToronto 3D dataCity models and open datasets informed later production.
04 / LATERCesium 3D TilesBackground context beyond the hand-modelled area.
ForegroundHand-modelled Toronto buildings, roads, infrastructure, public spaces, and project-specific assets.
BackgroundCesium 3D Tiles provided wider city context outside the primary production area.
Google-derived geometry was used only as temporary production reference and was not intended as shipped content.
Evidence in context
Inspect the result, not only the beauty shots
Drag each control to compare the Unreal environment against its real-world reference and to move between the final lit view and the underlying geometry.
Real-world reference versus Unreal
The comparison demonstrates close visual and spatial correspondence. It is not presented as a formal survey-grade accuracy certification.
Google imagery is shown as production reference only. Google and Google Earth are trademarks of Google LLC.
Lit result versus wireframe
The building silhouette remains efficient while materials, glazing, lighting, and interior systems carry much of the perceived complexity.
Connected city systems
More than a visual model
The R&D application was designed to be explored, queried, and connected to real city information. Urban data was translated into an in-world system rather than presented as disconnected overlays.
Building intelligence
Hover, highlight, selection, and building metadata sourced from Toronto Open Data.
Urban information layers
Schools, parks, public art, and other city points of interest.
Live traffic cameras
Camera feeds displayed at their corresponding real-world intersections.
Weather and forecast
Current Toronto conditions and forecast information from Weather.com.
Environmental presentation
Dynamic weather, vegetation response, lighting, and a complete day/night cycle.
Exploration and presentation
Free roaming, multiplayer, browser access, and cinematic capture.
I developed the in-world weather, lighting, material, vegetation, and environmental presentation systems. The lead programmer implemented Weather.com API parsing, forecast data structures, and related interface integration.
Cloud delivery
A city-scale Unreal application delivered through a browser
The application was delivered with PureWeb running on CoreWeave infrastructure. I optimized the Unreal application for the streaming environment while the third-party providers supplied the underlying platform.
7.2 to 7.8 GBPackaged application size
12 to 18 secCloud provisioning and initial entry
30 to 40 FPSReported streamed session
60 FPSReported local shipping build on RTX 3090
These are contemporaneous production measurements from an August 2023 presentation, not a controlled benchmark. Complete resolution and graphics settings were not preserved.
Production leadership
Scaling the team without losing the quality bar
I began as the project's first artist and technical contributor. As production expanded, my role moved toward technical direction, QA, artist enablement, and direct production review.
20 to 25employees and contractors at peak Toronto production
Contributors worked across regions, languages, skill levels, 3ds Max, and Blender. A substantial group in South America had limited English, so visual communication became essential.
01Led interviews and helped select contributors for the production team.
02Established modelling, material, naming, optimization, and delivery standards.
03Created editor tools, visual guides, video tutorials, and project-specific documentation.
04Reviewed incoming work and provided technical and visual QA across the environment.
05Maintained the final visual and technical quality bar while the CEO retained final company authority.
Outcome
400+Toronto-specific assets in a working city-scale platform
According to the company's CEO at the time, the project grew to more than 400 Toronto-specific assets. The finished R&D platform demonstrated how high-fidelity real-time rendering could combine urban information, multiplayer, cinematic presentation, and browser delivery.
The most important result was the production system beneath the imagery: shader-driven detail, infrastructure optimization, procedural R&D, reference alignment, technical QA, documentation, and distributed-team leadership.
Toronto reviewed by Epic's architecture and enterprise teams
In August 2023, members of Epic Games' architecture and enterprise teams reviewed the Toronto project, discussed a possible feature, and requested clean footage for a showcase. The showcase was not completed after 3D CityScapes ceased active operations.
Epic MegaGrant and Unreal Authorized Service Partner
3D CityScapes was named a 2020 Epic MegaGrant recipient before Toronto production began and later operated as an Unreal Authorized Service Partner. These were company distinctions, not a personal award or a Toronto-specific grant.
Project material
Explore the public archive
The videos, project gallery, real-estate context, and municipal data sources below provide additional context for the Toronto Waterfront Digital Twin.