Principal Scientist and “Mad Man” Eric Chan Discusses His Role in Improving Photoshop
A massive amount of work went into last month’s release of Photoshop CC, and now that the dust has settled, I got a chance to meet with Principal Scientist Eric Chan to talk about what’s new and what role he played in creating the product. Eric’s a brilliant scientist who takes a deep personal interest in the software he helps to develop. I learned a lot from him and I’m stoked to share part of our conversation with you:
Tell us about yourself and your history at Adobe. What features have you helped develop?
I joined the Camera Raw team at Adobe in February 2008. My first project was to improve the color rendition for pictures developed in Camera Raw and Lightroom. Boy, that was a lot of work — and a lot of fun! Since then, I’ve tinkered with improving others aspects of picture quality, including tone mapping, noise reduction, and lens corrections. I also work on raw support for new cameras and I build profiles for lenses.
Photoshop CC includes the newest features in Camera Raw such as Upright. What kind of research went into the new Camera Raw features?
Both Upright and the improved Spot Healing Brush in Photoshop CC are based on research from Adobe’s Creative Technologies Lab (CTL). I’m very privileged to work with those folks: they’re really good at tackling common yet complex photographic problems.
In the case of Upright, the idea was to straighten pictures automatically — fixing not just tilted horizons, but also converging lines (keystone effect). To accomplish this, the research team developed a robust line-estimation method to extract the primary lines within an image, and a system for determining how best to straighten images with several competing goals in mind (e.g., aligned edges vs perspective distortion). The underlying techniques are rather complex, but we tried to make the feature itself easy to use.
How does the Camera Raw team decide which features to add for a new release?
We listen carefully to photographers and take their feedback seriously — that’s the most important part. We receive many more requests than we can implement in a single release, so we have to prioritize and choose which projects to tackle. That’s a fun yet very difficult process! We try to mix some large projects that provide brand new capabilities (e.g., Smart Previews, lens corrections, defringe) with smaller projects (e.g., grid overlay refinements, per-channel RGB curves).
I find that smaller projects are often just as important as the larger projects: they are designed to address specific “pain points” that our users are experiencing. For the larger projects centered around imaging, we will usually try to group related ideas together for a new release: For example, with Process Version 2010 for Lightroom 3, we implemented several “detail-oriented” features: improved demosaicing, sharpening, and noise reduction. For Process Version 2012 for Lightroom 4, we implemented several “tone-oriented” features: revised tone and color controls, and a new tone-mapping method for highlights and shadows. We find that grouping related ideas together makes the features easier to grasp and to use for photographers.
Can you share anything about what you’re working on for future updates to Camera Raw?
After a major product launch (such as Photoshop CC) I find it important to take a step back and examine some of the longer-term projects we have in the pipeline at Adobe. I’m currently investigating several new technologies, many of which were developed by Adobe’s CTL within the last year. These projects are in the research stage (not yet ready for prime time) but they hold tremendous promise, including significant gains in both image quality and rendering performance.
For the past three major releases of Camera Raw I have focused primarily on image quality. While I plan to continue working on image quality advances, I will be spending more time in other areas going forward.
Are you a photographer? What kind of photography do you enjoy the most?
Yes, I love taking pictures, especially landscapes! I also enjoy closeup (macro) work of rocks, flowers, and food.
What’s your favorite feature that you’ve developed?
Probably the revised tone controls in the Basic panel in Process Version 2012. For my own photography, it’s probably made a bigger difference in day-to-day usage than any other feature I’ve worked on.
This was a team effort with Mark Hamburg, Thomas Knoll, Sylvain Paris, Zalman Stern, and many others. I remember being rather apprehensive at the beginning, since it was a pretty open-ended project with many (seemingly conflicting) goals, and frankly I was not convinced that it would make any real difference. Once I took a step back and looked at the big picture, however, I slowly realized that here was an opportunity to revisit many of the assumptions we’d made over the years regarding tone editing: the way we were blending colors for highlight recovery, the way we clipped individual color channels for the Exposure control, the way we handled (or rather, ignored) strong edges for the Clarity control — to name a few examples.
This turned out to be a rewarding and exciting project involving several iterations: I would try an implementation of the controls, users (both internal and external) would try the controls on their images and provide feedback, and then I would attempt to incorporate that feedback into the next iteration of the controls. Every aspect of the controls was reconsidered and revised: from the naming (Whites? Highlights?) to the order (Highlights above Whites?) to the range (defaults 0, -100 to +100) and their underlying implementation. All in all, I think we went through close to a dozen iterations on the controls over the course of a year!
You’re frequently found on our customer forums answering questions with the handle MadManChan. What’s the origin of that name?
Uh oh, I was afraid you’d ask that. Three words: High school basketball.
When testing new camera models, what’s your favorite subject to shoot?
My two cats! Cute, lots of fine detail, and often stationary (I usually shoot test images in the morning, when they’re fast asleep…).