Category Archives: Relighting

Interactive relighting control for Qwen image creation

AI/ML, Relighting, User Interface

A couple of weeks ago I mentioned a cool, simple UI for changing camera angles using the Qwen imaging model. Along related lines, here’s an interface for relighting images:

GenLit enables animated relighting

AI/ML, Relighting

I’m excited to learn more about GenLit, about which its creators say,

Given a single image and the 5D lighting signal, GenLit creates a video of a moving light source that is inside the scene. It moves around and behind scene objects, producing effects such as shading, cast shadows, secularities, and interreflections with a realism that is hard to obtain with traditional inverse rendering methods.

Vividon promises breakthrough relighting

AI/ML, Relighting

I stumbled across some compelling teaser videos for this product, about which only a bit of info seems to be public:

A Photoshop plugin that brings truly photorealistic, prompt-free relighting into existing workflows. Instead of describing what you want in text, control lighting through visual adjustments. Change direction, intensity, and mood with precision… Modify lighting while preserving the structure and integrity of the original image. No more destructive edits or starting over.

Identity preservation—that is, exactly maintaining the shape & character of faces, products, and other objects—has been the lingering downfall of generative approaches to date, so I’m eager to take this for a spin & see how it compares to other approaches.

Adobe’s “Light Touch” promises powerful, intuitive relighting

AI/ML, Relighting

Almost exactly 19 years ago (!), I blogged about some eye-popping tech that promised interactive control over portrait lighting:

I was of course incredibly eager to get it into Photoshop—but alas, it’d take years to iron out the details. Numerous projects have reached the market (see the whole big category here I’ve devoted to them), and now with “Light Touch,” Adobe is promising even more impressive & intuitive control:

This generative AI tool lets you reshape light sources after capture — turning day to night, adding drama, or adjusting focus and emotion without reshoots. It’s like having total control over the sun and studio lights, all in post.

Check it out:

If nothing else, make sure you see the pumpkin part, which rightfully causes the audience to go nuts. 🙂

SynthLight promises state-of-the-art relighting

AI/ML, Photography, Relighting

Here’s a nice write-up covering this paper. It’ll be interesting to dig into the details of how it compares to previous work (see category). [Update: The work comes in part from Adobe Research—I knew those names looked familiar :-)—so here’s hoping we see it in Photoshop & other tools soon.]

Shedding new light with LumiNet

AI/ML, Relighting

Diffusion models are ushering in what feels like a golden(-hour) age in relighting (see previous). Among the latest offerings is LumiNet:

Relighting via Midjourney

AI/ML, Relighting

Check out this impressive use of the new “retexture” feature, which enables image-to-image transformations:

Here’s a bit more on how the new editing features work:

Project Perfect Blend promises game-changing compositing in Photoshop

AI/ML, Relighting

Oh man, for years we wanted to build this feature into Photoshop—years! We tried many times (e.g. I wanted this + scribble selection to be the marquee features in Photoshop Touch back in 2011), but the tech just wasn’t ready. But now, maybe, the magic is real—or at least tantalizingly close!

Being a huge nerd, I wonder about how the tech works, and whether it’s substantially the same as what Magnific has been offering (including via a Photoshop panel) for the last several months. Here’s how I used that on my pooch:

But even if it’s all the same, who cares?

Being useful to people right where they live & work, with zero friction, is tremendous. Generative Fill is a perfect example: similar (if lower quality) inpainting was available from DALL•E for a year+ before we shipped GenFill in Photoshop, but the latter has quietly become an indispensible, game-changing piece of the imaging puzzle for millions of people. I’d love to see compositing improvements go the same way.

Magnific magic comes to Photoshop

AI/ML, ControlNet, Relighting

I’m delighted to see that Magnific is now available as a free Photoshop panel!

For now the functionality is limited to upscaling, but I have to think that they’ll soon turn on the super cool relighting & restyling tech that enables fun like transforming my dog using just different prompts (click to see larger):

Day & Night, Magnific + Luma Edition

AI/ML, Relighting

Check out this striking application of AI-powered relighting: a single rendering is deeply & realistically transformed via one AI tool, and the results are then animated & extended by another.

Meanwhile Krea has just jumped into the game with similar-looking relighting tech. I’m off to check it out!

Relight faces via a slick little web app

AI/ML, Photography, Relighting

Check out ClipDrop’s relighting app, demoed here:

Fellow nerds might enjoy reading about the implementation details.

How Google’s new “Total Relighting” tech works

AI/ML, Photography, Relighting

As I mentioned back in May,

You might remember the portrait relighting features that launched on Google Pixel devices last year, leveraging some earlier research. Now a number of my former Google colleagues have created a new method for figuring out how a portrait is lit, then imposing new light sources in order to help it blend into new environments.

Two-Minute Papers has put together a nice, accessible summary of how it works:

https://youtu.be/SEsYo9L5lOo

VFX & photography: Fireside chat tonight with Paul Debevec

Photography, Relighting

If you liked yesterday’s news about Total Relighting, or pretty much anything else related to HDR capture over the last 20 years, you might dig this SIGGRAPH LA session, happening tonight at 7pm Pacific:

Paul Debevec is one of the most recognized researchers in the field of CG today. LA ACM SIGGRAPH’s “fireside chat” with Paul and Carolyn Giardina, of the Hollywood Reporter, will allow us a glimpse at the person behind all the innovative scientific work. This event promises to be one of our most popularas Paul always draws a crowd and is constantly in demand to speak at conferences around the world.

“Total Relighting” promises to teleport(rait) you into new vistas

AI/ML, Photography, Relighting

This stuff makes my head spin around—and not just because the demo depicts heads spinning around!

You might remember the portrait relighting features that launched on Google Pixel devices last year, leveraging some earlier research. Now a number of my former Google colleagues have created a new method for figuring out how a portrait is lit, then imposing new light sources in order to help it blend into new environments. Check it out:

Photoshop’s new Smart Portrait is pretty amazing

Adobe MAX, Illustration, Photography, Relighting

My longstanding dream (dating back to the Bush Administration!) to have face relighting in Photoshop has finally come true—and then some. In case you missed it last week, check out Conan O’Brien meeting machine learning via Photoshop:

On PetaPixel, Allen Murabayashi from PhotoShelter shows what it can do on a portrait of Joe Biden—presenting this power as a potential cautionary tale:

Here’s a more in-depth look (starting around 1:46) at controlling the feature, courtesy of NVIDIA, whose StyleGAN tech powers the feature:

I love the fact that the Neural Filters plug-in provides a playground within Photoshop for integrating experimental new tech. Who knows what else might spring from Adobe-NVIDIA collaboration—maybe scribbling to create a realistic landscape, or even swapping expressions among pets (!?):

Check out “Light Fields, Light Stages, and the Future of Virtual Production”

Miscellaneous, Relighting

“Holy shit, you’re actually Paul Debevec!”

That’s what I said—or at least what I thought—upon seeing Paul next to me in line for coffee at Google. I’d known his name & work for decades, especially via my time PM’ing features related to HDR imaging—a field in which Paul is a pioneer.

Anyway, Paul & his team have been at Google for the last couple of years, and he’ll be giving a keynote talk at VIEW 2020 on Oct 18th. “You can now register for free access to the VIEW Conference Online Edition,” he notes, “to livestream its excellent slate of animation and visual effects presentations.”

In this talk I’ll describe the latest work we’ve done at Google and the USC Institute for Creative Technologies to bridge the real and virtual worlds through photography, lighting, and machine learning.  I’ll begin by describing our new DeepView solution for Light Field Video: Immersive Motion Pictures that you can move around in after they have been recorded.  Our latest light field video techniques record six-degrees-of-freedom virtual reality where subjects can come close enough to be within arm’s reach.  I’ll also present how Google’s new Light Stage system paired with Machine Learning techniques is enabling new techniques for lighting estimation from faces for AR and interactive portrait relighting on mobile phone hardware.  I will finally talk about how both of these techniques may enable the next advances in virtual production filmmaking, infusing both light fields and relighting into the real-time image-based lighting techniques now revolutionizing how movies and television are made.

Google & researchers demo AI-powered shadow removal

Image Science, Photography, Relighting

Speaking of Google photography research (see previous post about portrait relighting), I’ve been meaning to point to the team’s collaboration with MIT & Berkeley. As PetaPixel writes,

The tech itself relies on not one, but two neural networks: one to remove “foreign” shadows that are cast by unwanted objects like a hat or a hand held up to block the sun in your eyes, and the other to soften natural facial shadows and add “a synthetic fill light” to improve the lighting ratio once the unwanted shadows have been removed.

Here’s a nice summary from Two-Minute Papers:

https://youtu.be/qeZMKgKJLX4

New Adobe tech can relight structures & synthesize shadows

Photography, Relighting

Photogrammetry (building 3D from 2D inputs—in this case several source images) is what my friend learned in the Navy to refer to as “FM technology”: “F’ing Magic.”

Side note: I know that saying “Time is a flat circle” is totally worn out… but, like, time is a flat circle, and what’s up with Adobe style-transfer demos showing the same (?) fishing village year after year? Seriously, compare 2013 to 2019. And what a super useless superpower I have in remembering such things. ¯\_(ツ)_/¯ 

NewImage

[YouTube] [Via]

Begone, lame skies!

Mobile, Photography, Relighting

Does anyone else remember when Adobe demoed automatic sky-swapping ~3 years ago, but then never shipped it… because, big companies? (No, just me?)

Anyway, Xiaomi is now offering a similar feature. Here’s a quick peek:

And here’s a more in-depth demo:

Coincidentally, “Skylum Announces Luminar 4 with AI-Powered Automatic Sky Replacement”:

It removes issues like halos and artifacts at the edges and horizon, allows you to adjust depth of field, tone, exposure and color after the new sky has been dropped in, correctly detects the horizon line and the orientation of the sky to replace, and intelligently “relights” the rest of your photo to match the new sky you just dropped in “so they appear they were taken during the same conditions.”

Check out the article link to see some pretty compelling-looking examples.

NewImage

[YouTube 1 & 2]

Awesome new portrait lighting tech from Google

AR/VR, Photography, Relighting

The rockstar crew behind Night Sight have created a neural network that takes a standard RGB image from a cellphone & produces a relit image, displaying the subject as though s/he were illuminated via a different environment map. Check out the results:

I spent years wanting & trying to get capabilities like this into Photoshop—and now it’s close to running in realtime on your telephone (!). Days of miracles and… well, you know.

Our method is trained on a small database of 18 individuals captured under different directional light sources in a controlled light stage setup consisting of a densely sampled sphere of lights. Our proposed technique produces quantitatively superior results on our dataset’s validation set compared to prior works, and produces convincing qualitative relighting results on a dataset of hundreds of real-world cellphone portraits. Because our technique can produce a 640 × 640 image in only 160 milliseconds, it may enable interactive user-facing photographic applications in the future.

[YouTube]