How Look Match Works

Look Match uses a custom machine-learning model to suggest likely Lightroom edits when a photo has no metadata.

An estimate for photos without Lightroom metadata

PixelPeeper normally reads the Lightroom settings saved inside a photo. When those settings are present, it can show the actual embedded values. But websites, messaging apps, and export settings often remove that metadata. Once it is gone, there is nothing for PixelPeeper to recover.

That is where Look Match comes in. Unique to PixelPeeper, the model was trained on hundreds of thousands of images and Lightroom recipes from more than nine years of PixelPeeper data. It analyzes how the finished photo looks and suggests three likely recipes you can inspect, compare, and download as an XMP preset.

Look Match does not restore deleted information or reveal the photographer’s original choices. It is making an informed guess from the visible result.

What a match includes

Each match fills in controls that tend to describe the overall style of an edit, including tone, presence, color, curves, calibration, and grain.

Some controls are intentionally left blank: exposure, white balance, highlights, and shadows. These depend heavily on the original capture and the individual image. Copying guessed values for them would often make a match less useful.

The downloaded XMP is therefore a starting point, not a finished one-click edit. After applying it to your photo, you’ll need to adjust the basic light and color controls for your own image.

Why a match can be wrong

There are multiple ways to achieve the same look in Lightroom. Look Match cannot know which settings were actually used, but it can often get you very close to the original edit. The original light, camera, lens, profile, subject, and starting exposure all affect what you see, and Look Match only sees the final image. A warm result, for example, might come from white balance, color grading, a camera profile, the light in the scene, or a mixture of all four.

This creates a few important tradeoffs:

  • Some photos will match much better than others.
  • A recipe can look plausible while still differing from the edit that was actually used.
  • Compression, screenshots, social-media filters, and edits from other apps can make the suggestions less reliable.
  • Look Match will still attempt a guess for a photo that was never edited in Lightroom. In that case, any suggested Lightroom recipe is a false positive rather than evidence of an edit.
  • The three matches are alternatives to explore, not proof that any one of them is correct.

Look Match is currently in beta. Its suggestions should not be used to determine authorship, prove that a photo was edited, or make claims about a photographer’s workflow.

How to get the most useful result

Compare all three matches instead of relying only on the first. Look at the overall direction of the curve, color, and presence controls, then choose the recipe that gives you the most useful base for your own photo. Treat the sliders as clues and adjust them by eye.

For more about metadata, uploads, and account limits, see the PixelPeeper FAQ.