Fhotoroom has 2 noise reduction filters, Iterations and Bilateral with each having it's pros and cons. Below is a screenshot of a mountain photo with a lot of noise in the sky, which should be a consistent bluish tone instead has red and green pixel level shifts. This type of noise is usually caused by setting the ISO option on the camera too high or a scene is dark and taking a long exposure. This is because digital cameras produce 3 types of noise, random , "fixed pattern", and banding. The example below shows random noise and is usually the hardest to remove.
This is a before screen shot of an image (480% zoom) with random noise on the mountain and in the sky. In addition to this, the noise is both in luminosity and color based.
Bilateral Noise Reduction
Bilateral Noise Reduction is a very powerful fast way of removing random noise in a photo, but can suffer from starting to get a blurry look if the values are set to high. The big advantage of this method over the Iteration method is that can have a larger area of influence and produce smoother results.
Radius = This option determines the range of influence, and although this value can be set rather high usually values between 1 - 4 is more than enough cause image noise on the pixel level.
Threshold = This option determines the amount of influence (which pixels will be affect) within the range of influence determined by the Radius option. If this value was set to the max but the Radius is 0 then nothing will change.
In this example you can see in this photo that all the noise in the sky has been significantly been reduced.
Iteration Noise Reduction
Iteration Noise Reduction is micro noise reduction filter which reduces noise in a small area of influence. The advantage of this method over the Bilateral method is that it better preserves micro detail that can be lost using the Bilateral method.
Iteration = This option controls the number of loops the function will repeat to produce the final results. Increasing this value to high could cause to much blurring depending on the image as well as taking very long. A good initial value is usually 1 or 2. Range (1-10)
Sharpness = This option tries to restore lost detail in the image so the higher the value the great the local sharpness. This variable usually doesn't need a value great than 1, but that is optional. Range (0-5)
Red, Green, Blue, Luminosity = These 4 options all control the amount of correction that is applied to each channel. These options are very soft and can be very hard to notice the changes in some images, so working with a zoomed images is usually essential if changing these values is required. The cleanest results are achieved when these value are higher so 100% is usually the best choice. Range (0-100%)
Even though this image has been reduced from it's original size it is clearly visible that the many different colors in the sky have been reduced and that the mountain has a more consistent tone.