/ /

Motion Blur: How to Correct It

Posted on

No one is immune to motion blur, despite sensor or lens stabilization. Curiously, few software programs offer effective remedies, except for Topaz Labs.

Processing motion blur in an image depends on the performance of the application’s algorithms. The corrections obtained by Photo AI overcome it.

Stabilized lenses, initially designed for SLR cameras to improve image sharpness, allowed shooting at speeds as low as ⅛ s with success rates. The stabilized sensors of mirrorless cameras do not suffer from the vibrations of an SLR mirror. They allow sharp shots at half a second or even a second, of course, provided the subject is static. But sometimes, the best series shot falls victim to slight motion blur.

Decomposition and Recomposition

The blur of an image caused by movement is a decomposition of the subject. To regain its sharpness, it must be recomposed using specific algorithms. The blur is analyzed as a convolution, and its correction is obtained by deconvolution. While the principle is known to engineers specializing in digital imaging, few software programs remedy this problem. They settle for an improvement in sharpness that plays on increasing edge contrast, which does not solve the problem.

Once Upon a Time, Photoshop

However, Photoshop offered a specific motion blur reduction for several years thanks to a shake reduction filter (Filter > Sharpen > Shake Reduction…). However, due to the program’s rewrite, this filter has disappeared since version 23.3. As for Lightroom, while it offers a lens blur adjustment, nothing is currently offered to eliminate motion blur.

AI to the Rescue

Artificial intelligence is infiltrating everywhere and has not yet massively invested in correcting motion blur. For now, two applications integrate this type of adjustment powered by AI, which works by analyzing thousands of images: Luminar Neo and Photo AI. Until now, Topaz Labs’ software’s performance is the most convincing solution. Photo AI is not cheap ($199) and is only available in English. A trial version allows for evaluating its performance.

Photo AI Presets

Photo AI offers several settings: Denoise, Sharpen, Adjust lighting, Balance color, Recover faces, and Preserve text. Sharpen is the most relevant for motion blur, with its Motion blur function. The sharpness adjustment can be general or targeted to a part of the image, for example, an entire person or their face. With motion blur, correcting the sharpness on a background that would have been blurry anyway with a non-blurred shot is rarely useful.

Photo AI offers several settings: Denoise, Sharpen, Adjust lighting, Balance color, Recover faces, and Preserve text.
The parameters of the blur reduction settings are reduced to two modifiable sliders on a scale from 0 to 100: Strength and Minor denoise. They are automatically adjusted based on a preliminary analysis of the image. This initial processing is a good starting point. The personalized adjustment depends on each person’s taste. Our advice is to avoid an artificial rendering.

Face Blur

If the photograph contains faces, the motion blur correction on the faces will often be imperfect. Recover faces combined with Motion blur provide a more modelled, more realistic recovery of sharpness. Nevertheless, the combined use of the two adjustments in Photo AI does not have the flexibility of Photoshop layers. Since Photo AI is installed as a Photoshop plug-in, the two types of processing can be separated, and one can work on two layers in Photoshop.

If the photograph contains faces, the motion blur correction on the faces will often be imperfect. Recover faces combined with Motion blur provide a more modelled, more realistic recovery of sharpness.

Back to Photoshop

In the beginning, two layers are created in Photoshop. The first will be named Sharpen, and the second Recover faces, to use the Photo AI names. The first layer is modified with Filter > Topaz Labs > Topaz Photo AI, then Sharpen. The second is modified with Filter > Topaz Labs > Topaz Photo AI, then Recover faces. A mask associated with the Recover faces layer allows the treatment to be applied only to the face.

Since Photo AI is installed as a Photoshop plug-in, images can be adjusted in Photo AI from Photoshop.
After processing in Photo AI, the image is reintegrated into Photoshop.
Two layers are created in Photoshop. The first is named Sharpen, the second Recover faces, to use the Photo AI names. They will be respectively modified in Photo AI with: Filter > Topaz Labs > Topaz Photo AI then Sharpen and Filter > Topaz Labs > Topaz Photo AI then Recover faces.
The Recover faces layer is associated with a black mask. The adjustment made by Photo AI can be proportioned by painting in white with a brush on the black mask at the face level.

Become a professional photographer with Spéos

Spéos offers various training courses ranging from simple one-week photography workshops (initiation and advanced level) to 3-year courses. The long courses to become professional photographers allow you not only to master all the photographic techniques and its vocabulary (blurs, hyperfocus, sharpness zone, depth of field, backlighting, focal length, shutter release, autofocus, wide-angle, rule of thirds, etc.), but also all the stages of shooting and image processing.

Visiting the school allows you to discover the premises, the studios and the equipment, and is undoubtedly the best way to familiarize yourself with your future way of working. This is why, in addition to the open days, Spéos offers throughout the year personalized visits by appointment to come and discover the school with a member of the team.

Text and photos: Philippe Bachelier, teacher of Printing techniques at Spéos

Similar Posts