6 Things You Should Know When Using Neat Video
Neat Video is a plugin that gets rids of noise in videos. To understand how to fully utilize this application, we need to understand what noise is and where it comes from.
Cameras can only capture a fraction of the light and color values that we see with our eyes. Therefore, we must alter the camera’s settings to obtain the values from our camera that we want (ie an image that is neither under or overexposed). One of these settings that we typically alter is iso. Iso is the camera’s sensitivity to light. Basically the higher we set the iso, the harder our camera will try to pick up light.
This means that cameras have the ability to capture images in incredibly low light, but when we do bump up the iso we start to introduce noise into the image. This happens because our cameras start to pick up light that isn’t actually there. The camera may think there are mid tones in the shadows, highlights in the blacks, etc. that aren’t actually there. This creates undesirable patterns of brightness or color where we actually want an area consistency (noise).
Neat Video basically analyzes these areas and says that by finding consistency in all of the area surrounding this specific area then this area must be noise. Neat Video then smooths out these areas, matching them to the consistency of the area around them.
There are a few things (that weren’t immediately obvious to me) to consider when using Neat Video:
- It matters where you draw your rectangle.
Neat Video removes noise by sampling an area. You draw a rectangle in an area that contains noise where you want the color and brightness consistent (ie a blank wall, shirt, sky, etc.). Neat Video says ‘Ok, this area should be consistent, but isn’t… this must be what noise looks like.’ This allows Neat Video to understand what that specific image’s noise looks like so it is then able to remove it from the rest of the image and video frames. It works best if the area where you draw the rectangle to be sampled is not too dark or too bright.
2. You can use “Auto Profile” to draw your rectangle for you, but it isn’t always accurate.
The Auto Profile feature will look at your image and use its best guess to try to find an area that it believes should be consistent, but isn’t. Sometimes, there are images hidden in the noise that Auto Profile may mistake for noise, but you actually want in your image. For instance, Auto Profile may draw a rectangle in the sky that contains a small, powerline wire. You need to make sure that everything contained within the rectangle is all noise and not objects.
3. The larger your noise profile rectangle, the more information it has to pull from and the better Neat Video will work.
4. Go into the “Noise Filter Settings” to preview your noise-removed image and to alter how Neat Video handles the noise.
There are three settings in the Noise Filter Settings… Luminance, chrominance, and sharpening. Speaking simply, luminance is the value of brightness and chrominance is the value of colors. Therefore, bumping up the luminance in the settings will smooth more of the varying levels of brightness where there should be consistent brightness and bumping up the chrominance will smooth more of the varying colors where there should be consistent color. The last function, sharpening, I would leave alone because other color correction plugins or your editing software’s dedicated sharpening tool will do a better job.
5. If your settings are the same, you can copy and paste Neat Video across clips.
It takes a lot of time to build noise profiles for all of your clips. Luckily, however, if your lighting and camera settings are the same in multiple clips then you will have the same noise. Therefore, you can just copy and paste Neat Video for these clips. If you think noise may differ slightly, you can go into Neat Video and use its “Auto-Fine Tune” option and it should cater the noise profile more to your individual clip. However, I would not recommend copy and pasting Neat Video to every clip in your timeline because not every clip will have the same noise so you will lose detail in less noisy shots and not smooth all of the noise in more noisy clips.
6. Temporal Filter Radius, Threshold, and Adaptive Filtering
These are the settings in the the Neat Video menu when you first drop Neat Video onto your clip. Temporal Filter Radius controls how many frames of the clip you want Neat Video to analyze (the lower the number, the faster Neat Video works). Temporal Threshold controls how sensitive the temporal filter radius is to changes in the motion of the objects of the frame (the lower the number, the less objects move in the clip). Adaptive filtering controls the amount that Neat Video will look for noise that moves. Use a lower number when the noise stays in the same place and a higher number when it bounces around.