VideoHelp Forum




+ Reply to Thread
Results 1 to 2 of 2
  1. Does someone know the difference between 2 ways of noise removal/reduction:


    1. vertical way 3 pixel

    or

    2. temporal way 2 pixel

    What's the difference and what does these mean, how does it work?
    Quote Quote  
  2. Generally there are two ways to remove noise, spacial and temporal.

    Spacial filtering looks at nearby pixels of the same image. In the most simplistic case you can reduce high frequency noise by making the color/intensity of each pixel the average of that pixel and the pixels right around it. This can have the side effect of making the image blurry. There are techniques to limit the blurring -- involving only smoothing areas where the pixels are nearly the same color, using weighted averages, etc.

    Imagine a small horizontal line of pixels where 0 is black, 9 is white, and and the other digits are shades of gray in between:

    3 3 3 3 3 7 7 7 7 7

    This is a sharp transition from a dark gray to a light gray. Add a little noise and you get:

    2 3 4 2 4 8 7 8 6 7

    Now simply make each pixel the average of that pixel and the pixel on each side of it (at the ends use the average of two pixels) and round off to the nearest integer:

    3 3 3 3 5 6 8 7 7 7

    As you can see, the noise that was added in the second step was largely removed by the filtering, but the sharp transition from 3's to 7's has been blurred across several pixels.

    With video, since most parts of the image don't change from frame to frame, you can smooth across the temporal dimension. That is, you can average a pixel with the corresponding pixel in the previous and/or next frame. This will blur the image less spacially but potentially blurs the image temporally. Again, there are techniques to limit the bluring.
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!