Google Photos now lets you make complex HDR photo edits without sacrificing quality

0comments
Google Photos app on a smartphone.
If you've ever edited photos on your smartphone, you know the struggle – sometimes, they just don't look as good after you tweak them, right? Well, if you're a Pixel phone user, you're in for a treat! Google just revealed that you can now make those complex edits, like using Magic Eraser, in Photos without sacrificing Ultra HDR quality.

Google just rolled out a new machine learning model that's set to elevate HDR image editing in Google Photos. This tech lets you make complex edits to HDR images without sacrificing their visual quality.

On the Pixel 8 and newer models, this new model works in the background within Google Photos, ensuring that even after you make intricate edits to HDR images, your final results still retain their HDR quality.



Before, when you edited HDR photos in Google Photos, you often ended up with a downgrade to standard dynamic range (SDR), which meant losing out on brightness, contrast, and overall image quality. But with this new machine learning model, that's no longer a problem. It cleverly predicts any missing HDR metadata after you make your edits, ensuring that even after using features like Magic Eraser, your HDR images maintain their original dynamic range.

The main hiccup lies with the Gain Map — a log-encoded image that sits alongside the SDR version and shows how much to brighten each SDR pixel for that desired HDR look. Tools powered by machine learning, like Magic Eraser, Photo Unblur, Magic Editor, and Portrait Blur, were designed with SDR images in mind, expecting them as inputs and spitting out SDR results.



Take Magic Eraser, for instance. It lets you zap away annoying elements from a photo by filling in the erased spots with new pixels. Now, if you're working with an HDR image and you erase a bright section, chances are the corresponding area on the Gain Map was bright, too. If you don't also adjust the Gain Map, you'll end up with an annoying "ghosting" effect where the original Gain Map overlaps the edited SDR image, which is definitely not what you want.

So, in simple terms, Google Photos tackled this issue with the new ML model that recreates those missing Gain Map areas by predicting a new one and then seamlessly blending it with the original.

I think Google's ML model is a game-changer for HDR image editing. It gives users way more flexibility and control when it comes to tweaking their photos, making the Pixel an even more appealing choice for mobile photography enthusiasts.

Recommended Stories

Loading Comments...
FCC OKs Cingular\'s purchase of AT&T Wireless