The problem with "objective" images
Mixing fine-art documentation with product photography techniques
My role: Digital Imaging Specialist / Photographer
Client: Phuong Nguyen
Scope: Artwork documentation
Tools: Photoshop, Capture One
Deliverables: Publication ready image files
Published: 09/10/25
The project
Over years of working with Phuong, her artwork has shifted from traditional oil paintings to more sculptural art-objects that mix varieties or reflective surfaces. The challenge before us was, how might we represent the artwork in photographs in a way that was accurate to the in person expereince.
The problem
Reflections only when you need them
Phuong's artwork combines multiple surfaces each requiring different approaches. Those being:
Oil paint on panel
Ceramics
Stained/painted wood
Polished metals
Opalescent substrates
High relief sculpture
Each of these materials needs a different lighting treatment and to work in an efficient manner, this meant that a workflow needed to be established to quickly switch between requirements and keep shoots efficient and editing time within project constraints.
Objectives
Develop a streamlined workflow from capture through editing
Reduce time on task for editing to absolute minimum
Accurately represent the artwork as it appears in person
The solution at a glance
Create rigging and lighting setups that allow for shifts between hard polarized light and diffused unpolarized light without moving the camera, or artwork while giving real time previews to the client.
Identifying the right approach
The constraints
After working with Phuong to understand the kind of work that would be documented on an ongoing basis We identified a few key elements that would lead to protraying the work in a way that matched to in person viewing expereince. These elements were:
No shines on the oil panted portions
Soft and difussued shadows on the relief elements
Maintain pearlesncese in the inlays and accents
maintain a sense of depth in the image
The other consideration was that the porcess needed to be reapeatable and compact. It would have to be achaivable in the artists's studio not a photo studio, and should eb replacable on a go forward bassis with shoots every 2-3 months.
Tokens, components, and state logic were adapted rather than adapted. This freed time to focus on flows, content clarity, and accessibility.
In camera vs post-production
Knowing form experience that the process would require compositing it, it was essential to balance the amount of work needed ebtween in camera effects and post-production tricks. Breaking down the image into components it was clear that There would need to be a minimum of 2 exposures needed.
Diffused unpolarized lighting to achieve soft shadows and smooth lighting effects
A hard polarized light to show off textures without shines
The approach then became an exercise in understanding how to quicly get both effects with minimal interaction with the equipment to allow for faster compositing.
Tethered capture for quick feedback
The need ofr quick feedback and accurate images is critical in this context and tethered capture was the only method that would allow for this. For those who aren't nerds about this stuff—"Tethering" refers to the practice of attaching your camera to a computer in order to use i=a program like Capture One, or Lightroom as your camera's user interface.
Tokens, components, and state logic were adapted rather than adapted. This freed time to focus on flows, content clarity, and accessibility.
This process allowed for faster, tighter compositions limiting the need for additional retouching. It also allows for a stremalined workflow in which you are able to quicly qwork with your client to see raw version sof the images on site, eliminating costly back and forths approving images and making revisions. Additionally this process build trust in the client by allowing themn to see rough versions of the deliverables early in the process.
The capture process
With the revisied workflow the shoot days are simplified and usually finished in under 2 hours including set-up and tear down. It folows a simple order of events:
Lighting set up and discussion of the days shot-list
Calibration and begining of tether capture session
Each artwork is brought to a wall photographed 1 polarized and once with difussion
Shots are selected and grouped for composite
Selects are signed off by client
Teardown and and goodbyes
The process is made so smooth by using a single light source and reflectors with the diffusion being achieve thorugh foamcore sheets and a bounce with the reflectors removed. While I could go on for paragraphs about proper lighting placement and so forth, it is enough to say that you can achieve the same effect with 1 light, 10$ worth of foam core and a $20 reflector from Amazon if you were to be on a budget.
Fig 1. This show the direction of the polarized light from the light source. It is reflected off of the artwork and retains it's polarity, minimizing shine.
Fig 2. This shows the additional lighting equipment used to diffuse the light before the second capture.
fig.3 This shows the diffusion of the light for the second capture. Once the light passes through the diffusor, it loses its polarization and the shines return.
Fig 4. The darker purple area shows the polarized light, while the magenta area illustrates the camera's view. The lighter the shade of purple the less impact the polarizer is having on the light. Note how the artwork on teh wall es effectively unchanged by the polarizing gel.
Post-production
Move fast, stay consistent
Leveraging templates
Post work on these is very in depth due to the intricacy of the objects. It was important to find an efficeient and relaible method for masking and way to clean up the background effectively. The editing process has two main phases:
Compositing and cleanup
Color correction and background standardization.
The formulaic nature of the process. along with templates and proper files structures has brought the process down to a highly effective editing workflow, dramatically reducing time on task and shortening turnarounds.
A screenshot of the compositing process in which manual adjustments are made to the AI alignments in Photoshop.
Quiet AI tools to speed up masking
Adobe's machine learning tools have ebeen improving gretaly over the years. The most effective in standard workflows though has been the select subject tool. This tool has reduced the masking time on these works dramatically, even beyond that of the object selection or channel selection tools. By leveraging these powerful tool to reduce time on task, turnaround times on these images has dropped dramatically.
The masking process in photoshop
Quiet AI tools to speed up masking
Adobe's machine learning tools have ebeen improving gretaly over the years. The most effective in standard workflows though has been the select subject tool. This tool has reduced the masking time on these works dramatically, even beyond that of the object selection or channel selection tools. By leveraging these powerful tool to reduce time on task, turnaround times on these images has dropped dramatically.
The prototype
Finding a detour
This video shows the user flow for finding the Bayview Ave detour from the closure at the Riverdale Park Viaduct. It is designed to limit the needed interactions and provide the user with sufficient information as quickly as possible. Considerations to note include:
"Finding a detour" preview animation with an optional skip button
Satellite map view under the "Trail Map" portion of the prototype
Images for quick recall when displaying detours
Transcripts for quick review on the "Audio" screen
Flow shown: Home Page → Find a detour → Bayview Detour → Directions
Accessibility considerations
Baked in, not bolted on
The project focused on accessibility, implementing features like optional audio directions and reduced screen reliance for low-vision users, and non-colour signifiers and clear labels for colour-blind users. Other priorities included scannable copy, WCAG-compliant contrast, and mobility cues. Future plans involve formal usability testing with low-vision participants, screen-reader analysis, and expanded audio cues for better routing and labeling.
Learning and outcomes
By the numbers
12 sec median time‑to‑detour from home to actionable directions (on‑trail pilot tests).
3 taps to reach a usable detour (down from 10+ screens across two sites).
1 site centralising closure info, access points, and detours.
What I delivered
Research synthesis across municipal notices and social posts; 1:1 user interviews.
Information architecture, task/flow diagrams, and an interactive Figma prototype.
Bridge‑based sectioning model aligned to major entry points.
Satellite map UI using customised MDS components; design tokens and UI states.
Audio direction scripts and a basic voice‑guidance prototype.
What I’d do next
Connect to a live closures feed or lightweight CMS for timely updates.
Tag sections on trail map with surface/grade/width to support wheelchair routing and filters.
Localise copy (EN/FR) and expand to additional Toronto ravine corridors.
Run formal usability sessions with low‑vision participants; quantify time‑to‑detour at scale.
Impact & reflection
The solution is implementation-ready with minor adjustments to Toronto’s municipal branding and a simple data pipeline.
Key lesson
Novel visuals can be powerful for insight, but familiar patterns—maps, concise labels, audio—win for speed and inclusion. This project sharpened my ability to pivot quickly, test assumptions, and balance accessibility with real-world usability.
Thanks for reading, let's connect!
info@gregmccarthystudios.com