Jason Kravitz
02-02-2005, 06:00 PM
<img src="http://www.digitalmediathoughts.com/images/jek_nn_logo.jpg" /><br /><br /><b>Product Category:</b> Image Editing Software<br /><b>Manufacturer:</b> <a href="http://www.picturecode.com">PictureCode</a><br /><b>Where to Buy:</b> <a href="http://www.picturecode.com/purchase.php">PictureCode</a><br /><b>Price:</b> $79.95 (professional) $44.95 (home) USD<br /><b>System Requirements:</b> Mac OS X (10.2 or later), Microsoft® Windows® XP/2000/98SE/ME<br />Requires either Adobe Photoshop, Photoshop Elements or other Photoshop Plugin compatible graphics software.<br /><br /><b>Pros:</b><li>Easy and quick to use;<br /><li>Effectively removes digital noise;<br /><li>Seamless integration in processing workflow.<b>Cons:</b><ul><li>Documentation could have more detail.</ul><b>Summary:</b><br />Digital Cameras make photography easier than ever but can have an unsightly side effect, digital noise. Digital noise is more prevalent in low light settings and can often take on the distracting appearance of colored confetti like lines and squiggles. The film equivalent to digital noise would be grain although film grain is often a desired effect where as digital noise typically detracts from the image. Fortunately, there is a solution. Noise Ninja by PictureCode targets noise without effecting the overall image. Noise Ninja was originally released as a stand alone product and is soon to be released as a Photoshop plugin. Read on for a further explanation of digital noise and a preview of the upcoming Noise Ninja plugin.<br /><!><br /><PAGEBREAK><br /><span><b>Patterns and Profiles</b></span><br />Digital noise typically occurs in some sort of predictable pattern based on camera brand and model. It is possible to profile your camera to get a baseline for the pattern of noise. Noise Ninja can use this baseline profile to remove noise while retaining detail and sharpness in the picture. There are many profiles available for download on the Noise Ninja product site. <br /><br />The ISO setting on a digital camera is roughly equivalent to the ISO rating of film. Higher ISO film is also referred to as "fast" because it allows you to use a faster shutter speed to capture action in darker settings. A higher ISO number in digital photography will have the same effect, however, in exchange for shooting in low light, higher ISO values tend to introduce more digital noise.<br /><br />Since the pattern of noise can vary at different ISO settings there is a noise profile for each ISO setting per camera model. When working with an image, you can look at the EXIF data to determine what the ISO value is for that image and then choose the correct profile accordingly. Noise Ninja can also read the EXIF data and load an appropriate profile based on the ISO setting and camera model (note: this feature has not been implemented in the current plugin beta)<br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_ISO_setting.gif" /><br /><i>Figure 1: The ISO Speed Rating can be found by looking at the EXIF data for an image.</i><br /><br />Noise Ninja also has an automatic profiler which can be used if there is not an existing profile for your camera. The auto profiler attempts to find uniform areas with little detail to create a profile for the noise in the rest of the image. The documentation states that the auto profiler can be confused by certain textures and that using a preset profile or manual profile can be more effective. The automatic profiler in the plugin has been improved over the stand alone product. These improvements will be merged back into the stand alone product in a future release. For my uses so far, I've found the automatic profiler does a good job with my test images.<br /><br />For a finer degree of control, there is an option to set up a manual profile. Creating a manual profile is as simple as drawing rectangles over the image preview. Each rectangle area is examined by the program and used to create the profile. The documentation suggests highlighting areas that are similar in color and tone. The manual profile tool can also be used to enhance the results of the auto profiler. <br /><br /><PAGEBREAK><br /><span><b>Getting Around</b></span><br />As a Photoshop plugin, Noise Ninja is installed to be run from the Filters menu. The interface is fairly simple and easy to navigate. The left pane contains a view of the entire image. There are controls to zoom in and out, select a preview area, create manual profile selections and set the color channel view. <br /><br />The right pane contains a detailed view. Sometimes it is difficult to see noise in the larger preview window. The detail view shows a zoomed in subsection of image. After applying a profile, the detail view shows the image with the noise filter applied. There is a toggle button to show the original image and filtered image to see how successful the effect is.<br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_screen_shot_sm.jpg" /><br /><i>Figure 2: The main window is divided into two sections.</i><br /><br />Most cameras store images in the RGB (Red Green Blue) color space. Noise Ninja converts the image to the YCrCb color space. Y is the Luminance channel, Cr is the red-green channel and Cb is the blue-yellow channel. In the YCrCb space, luminance represents the brightness of each color in the image. Because most of the image data is stored in the luminance channel, Noise Ninja ignores the actual color and looks for edge and texture information to remove noise more effectively. The image can be viewed in luminance, red chroma and blue chroma mode. These modes make the image appear in gray scale and each level of gray represents a different brightness value. These views can show noise better than in the RGB view.<br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_channel_view.jpg" /><br /><i>Figure 3: An example at 100% pixel crop of RGB, Luminance, Red Chroma, Blue Chroma views respectively.</i><br /><br /><PAGEBREAK><br /><span><b>Tuning In</b></span><br />Beneath the detailed view is a panel with three tabs. The first tab contains the profile settings. This is where you create an automatic or manual profile. There is a graphical layout called the profile chart which shows the relationship between noise and color and tone. The graph can be a visual indication of whether or not the current profile is effective. <br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_profile_graph.jpg" /><br /><i>Figure 4. The profile tab contains a graphic representation of the current profile.</i><br /><br />The filter tab contains many different controls for fine tuning the image. I found the default settings to be sufficient, but, for those who like to experiment there are many options. The sliders are divided into three groups. The Luminance group allows you to adjust the strength of the overall effect, smoothness and contrast. The Colors group is useful for removing colored noise which can look a bit like confetti in your photograph. The last group of sliders if for adjusting the edge sharpness using an unsharp mask filter. <br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_filter_sliders.jpg" /><br /><i>Figure 5. The filter tab contains controls for fine tuning the effect.</i><br /><br />The properties tab will be used for setting annotation data for the profile such as camera make, model and ISO setting. The properties tab was not implemented in the beta version used for this review.<br /><PAGEBREAK><br /><span><b>Conclusions</b></span><br />Using the Noise Ninja plugin can be extremely quick and simple. It is recommended to run the filter early in the workflow process because it is easier to remove noise right out of the camera than it is after applying other effects and filters. The documentation contains a quick start guide which is useful for getting the gist of using the product. Ultimately it is as easy as clicking the Profile Image button and applying the filter. All of the other settings are used to get a better profile or fine tune. I found the results to be a good mix of eliminating the noise without affecting the overall sharpness of the image. At the time of this writing, the plugin is still available for free as a public beta. I am looking forward to the final release of this plugin as it will have a place in my workflow process.<br /><br />The release version of Noise Ninja can be purchased from the PictureCode Website. The professional version includes 16 bit image support, batch processing and multiprocessor support. The home version only allows 8 bit image editing and no multi processor or batch functionality. The license will work with both the plugin and stand alone version of the product.<br /><br />Below are some examples of the images before and after the filter. For each example, I used the auto profiler with default settings. The original image size is 3504x2336 pixels. The reduced image is 400x267, although shrinking the picture helps reduce the noise so I will also include a 200x400 pixel subsection of the original image. The yellow box on the larger image show which subsection is highlighted.<br /><br />The image on the left is before the filter and the one on the right shows the effect of the filter. All images were taken with a Canon 20D at either ISO 1600 or 3200.<br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image1_full.jpg" /><br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image1_crop.jpg" /><br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image2_full.jpg" /><br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image2_crop2.jpg" /><br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image2_crop.jpg" /><br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image3_full.jpg" /><br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image3_crop.jpg" /><br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image4_full.jpg" /><br /><br /><img src="http://www.digitalmediathoughts.com/images/jek_nn_image4_crop.jpg" />