DxO PureRAW3 is an innovative noise reduction processor that promises to transform your noisy images so much it will feel like having a new camera. DxO’s DeepPRIME and DeepPRIME XD algorithms appear to produce stunning results at first glance.
I conducted an experiment to evaluate how well DxO PureRAW3’s DeepPRIME and DeepPRIME XD algorithms could reduce noise on wedding photos shot underexposed in dim conditions at high ISO to see just how effective their noise reduction technology works.
My Canon EOS 7D Mark II crop sensor digital SLR took all of the images featured herein.
What is the Difference between DeepPRIME and DeepPRIME XD in DxO PureRAW3 Processing?
Both options use the same deep learning algorithm to reduce noise and preserve detail, with the latter offering enhanced details over its counterpart – taking longer but yielding superior results over regular DeepPRIME.
Before and After Images Processed with DeepPRIME
DeepPRIME took less than one minute to process this image while DeepPRIME XD took five minutes.
DeepPRIME offers dramatic differences from RAW processing; uniform areas become smooth while facial details are preserved yet sharpened and clarified. DeepPRIME XD excels significantly more in high resolution JPG exports but the difference may not be readily evident in smaller resolution JPGs.
After DeepPRIME XD processing, this photo serves as one final comparison: I edited my original edited version before applying my edits to the DeepPRIME XD processed image. At high resolution it’s clear that DxO PureRAW3’s algorithm dramatically reduced noise – it almost looks as though another camera took this same photo!
Lightroom Classic was my choice for edits. DxO PureRAW3 provides built-in export capabilities directly into Lightroom Classic, Lightroom or Photoshop; additionally there’s an import plugin which enables use of DeepPRIME/XD from within Lightroom Classic without ever leaving your catalog; I really liked how easily it integrated into your workflow!
How Does DxO PureRAW3 Reduce Noise? DxO PureRAW3 utilizes proprietary technology utilizing a neural network of billions of images to run both demosaicing and denoising at once without making any compromises between them.
Demosaicing converts red, green and blue data for each pixel into its true colors while denoising removes incorrect pixels caused by camera processing heat and interference. Noise pixels become particularly prevalent at high ISO settings since sensors must become more sensitive to detect low light levels – leading to internal processes compromising image quality more significantly than usual.
DeepPRIME uses AI to combine demoisaicing and denoising simultaneously, without either process undermining the other. By training on billions of images, DxO’s DeepPRIME and DeepPRIME XD models understand what colors and details should be present for an optimal result.
Compare Color Changes Before and After DeepPRIME Processing in DxO PureRAW3
This image clearly displays my mastery of color using DeepPRIME instead of XD. Noise reduction is evident here as well, while look at all that vibrant hue when I edit my photo according to my usual method and adjust for underexposure!
