Visual Quality

To understand what visual quality we wanted to provide to our users, I worked with the visual engineer and marketing planner. We identified presets that could be used as starting points for new users to easily set up their webcam. We also identified image settings that could be available to users for and how their expected behavior should work. To further understanding, we paired with an outside vendor who had extensive visual quality experience and could provide an objective evaluation of the image quality and what we could do to improve upon it. We then tested image quality against competitors to iteratively check our progress to make sure our visual quality was premium. Attributes related to image quality were used as parameters (i.e., brightness, clarity, saturation) and participants scored on a "just-right" scale where scores ranged from 1 (too low) to 5 (just right) to 9 (too high). In addition to the parameters, overall scores and rankings were also asked to ensure to track our performance relative to our competitors. Based on the results from user testing, visual quality was tuned accordingly. This was a particularly fun experience to go through as one of our early firmware versions was hilariously oversaturated, resulting in an awkward sunburned look for those with paler skins.

​

Presence Detection

Since this was a premium device, we wanted to investigate ways in which we could differentiate this product from the rest of the competition. User privacy was a core tenet we had identified from our user research. One solution we had was to leverage our Dell ecosystem and provide additional security features for customers who had both a Dell laptop/PC as well as our webcam. To further understand how this would work on a technical level I worked with our Advanced Engineering team to understand the technology behind the architecture and sensor technology to define behavior for detecting user presence (or absence) to trigger PC lock, PC wake, and facial recognition. 

I created storyboards and specifications for how the ideal behavior around this feature would work and iteratively improved it as we learned more and more about this feature. Factors like technological capabilities (e.g., sensor tracking distance, field of view), user behavior (e.g., intention to turn on PC or just walking by), and consideration of various use cases (e.g., multiple presence sensors prioritization, privacy control) all played a part in the eventual design and functionality of this privacy feature. 

​

UX testing was held in which participants discussed and played around with presence detection. Because presence detection is a relatively new feature, many participants were unfamiliar with the technology but had no trouble envisioning the various use cases they would need it for. It should be noted that the POC sample we had for testing was having some issues as this was a rushed prototype and the task performance scores we received were not good. Despite the low score, the concept reception by the participants helped validate the use case, but still reinforced the idea with the team that the implementation of this feature in the actual product needs the performance to be foolproof. It also helped reinforce the idea to upper management that novel ideas and features need time to implement, both from a behavioral and engineering perspective. 

​

Auto Framing Specification

This was a more challenging specification to define and create. To help provide visual understanding of what concepts I wanted our auto framing feature to have, I created storyboards to indicate subject behavior and subsequent camera framing action. Lots of collaboration was required between myself and the firmware engineers to make sure that action followed according to spec and was tuned correctly. Iterative improvements came out on an almost weekly basis to help ensure that the spec was followed, and any areas of uncertainty had to be further included in the spec to make sure it covered all the various use cases (multiple subjects, subject distance, head micromovements, etc.) or technological boundaries. 

​

One interesting question I had during this study was the differences or similarities in perception of this feature from the subject and audience perspective. From the subject perspective,  because a speaker often looks at their face in a collage of faces, I didn't want to present a view of themselves and become distracted by their camera constantly moving and adjusting to their head placement. Similarly from the audience side, I didn't want them to become distracted, or worse, get some kind of dizziness or motion sickness from the presenter's background constantly moving around. Thus tuning the speed and sensitivity of the auto framing was a challenging but fun specification I made.

​

For UX testing, we tested participants in both the subject and audience roles. We found that majority of participants were delighted by this feature and intrigued particularly if they had to do a demo of a product. Areas of improvement were the ability to improve the fluidity of the framing, as well as the ability to allow the user to set their own customizable speed/sensitivity/frame, features that we are currently working on for our Auto Framing 2.0.