CHAPTER 8 — Noise Reduction

Noise is present in all digital capture and it’s a major cause of color distortion. When the signal detected by your camera’s sensors is low the signal to noise ratio will also be low. The result is the noise becomes visible and becomes a source of both color distortion and obvious artifacts in your photos. The noise is most often seen in the low light areas of your photos where a camera has inadequate dynamic range allocation to capture all of the information in the shadows. Removing and reducing this noise requires an in depth understanding of the causes of noise, and a unique solution to automatically detect for noise and to remove without blurring the photo.

bride

There are many sources of Noise in digital photos

Noise in digital photography can be thought of as the equivalent of grain in film photography, therefore it would be unfair to view film photos as noiseless. Noise can be introduced into your digital photos by a variety of factors including:
• your camera having inadequate dynamic range available to capture all of the information available at the time of the photo
• your camera’s sensors inadequate sampling, giving rise to aliased signal,
• the small footprint of your camera resulting in smaller and more tightly packed sensors [this is done to facilitate a smaller camera size and in an attempt to reduce aliasing] reducing accurate sensor response
• small sensors in point and shoot cameras record less signal so the signal to noise ratio of each sensor is much lower than the signal to noise ratio of a digital SLR where the sensors will be roughly twice as large
• the arrangement of your camera sensors requiring a de-mosaicking process to generate a picture.

In addition to noise stemming from mechanical camera limitations, noise arises from the conditions of capture itself. For example, sports, indoor scenes and other situations that require high ISO camera settings [rendering the sensors highly sensitive to light variances] are typically affected by visible amounts of light and color noise which can degrade the picture fidelity of the original image.

Because there’s so many potential causes of noise…

It’s important to note that in any given photo there’ll most likely be multiple causes of noise and therefore the noise will take multiple forms. To deal with noise effectively, noise correction algorithms must be able to distinguish between luminance noise and color noise. Color noise is synonymous with color distortion. Luminance noise is corrected by methods that generally create color distortion. So whether the noise is color noise or luminance noise the order of the day will be color distortion.

Perfectly Clear incorporates the power of Stoik automatic noise reduction

Perfectly Clear incorporates the automatic noise detection and removal solution from Stoik Imaging. In rigorous testing performed by Athentech Imaging the Stoik engine was revealed to be:
1. The fastest solution on the market
2. the best solution for preserving image detail
3. the only solution with true automatic noise deduction
4. the only solution to accurately apply varying noise corrections depending on the ISO noise level

Stoik Noise Autofix is the first photo noise reduction algorithm that allows for fully automatic operation. The algorithm will first analyze the photo to determine if noise exists, and only then apply the proper amount of noise removal. Specific algorithms and settings are applied depending on whether the photo was taken with a digital camera, camera phone, or scanner. The algorithm includes modules of noise detection, noise analysis and noise filtration which are statistically trained to provide optimal balance between photo noise reduction and preservation of image details. Unlike other noise removal algorithms that blur photos after removing noise, the Stoik algorithm is unique in preserving the crisp details. The noise in digital photos is reduced by 2 – 3 stops, so that the noise level of the photo shot at ISO 1600 is effectively reduced to ISO 200 – 400 levels. It is possible to select various automatic noise presets creating special softening of the skin for portrait photos, and extra powerful noise removal for camera phones, plus a special night shot noise removal.

Figure 43 Orig
Original

Figure 43 NW
Noiseware

Figure 43 NN
Noise Ninja

Figure 43 St
Perfectly Clear with Stoik Noise

Noise and the human eye

Noise results from camera limitations. Arguably the human eye has noise because the eye can’t see infinite levels of detail. Having said this the human eye sees the fractal world as being continuous, therefore the eye does not create artifacts. It’s for this reason that noise in your photos is so immediately recognized as egregious. This color distortion and artifacts must be removed effectively to reproduce your original experience.

The Noise Reduction Correction in Perfectly Clear is designed by Stoik Imaging to reduce the noise and reproduce photos with true, accurate color … the true Real Colors. Perfectly Clear overcomes your camera’s limitation of noise creation in your photos, a very common problem of digital cameras because of their fixed dynamic range. The results are Accurate Photos that match the original image in the mind’s eye and serve to preserve Precious Memories perfectly. Accurate Photos are Superior Photos and science shows superior photos have the greatest emotional impact with viewers.