Device Noise Profiling
A device noise profile describes the properties of noise produced by a recording device working in a certain mode. Specific flat featureless (noise-only) video frame areas are used to build such profiles. Examples of video frame areas that can and cannot be used to build a profile are located at the bottom of this page.
You may also want to check video tutorials on how to build a good device noise profile.
Building a noise profile
Auto Profile function in Neat Video can automatically find flat featureless areas if such areas exist in the input video frame. You can also select such an area manually.
To build a noise profile:
click the Auto Profile button to automatically find and analyze a flat featureless area in selected frame of the input video clip
- select such an area manually and then click the Auto Profile button to automatically analyze the selected area
Aim to find a large featureless area (up to the maximum of 512x512 pixels). Using a large area is preferred because that allows to more accurately measure the properties of the noise. Using smaller areas is possible too, but the accuracy may be reduced. The absolute minimum size is 32x32 pixels.
If selected video frame does not contain such featureless areas, try to use another frame from the same series. If no frames of the input clip contain featureless areas, you can try to use a frame from another video clip produced by the same video capturing device working in the same or similar mode. Then additionally you can fine-tune the resulting profile by analysing a frame from your current video clip: use Auto Fine-Tune once or use manual Fine-Tune on several featureless areas.
Alternatively, you can use the Calibration Target to build a new noise profile for the input video clip. The Calibration Target is specifically designed to easily build reusable profiles for various recording devices. It can be used to build a single profile for a certain device mode or a set of profiles for different modes.
Finally, you can obtain a ready-made profile made from a similar video clip produced by another capturing device of the same model and additionally fine-tune the resulting profile by analysing a frame from your current video clip by using Auto Fine-Tune once or using manual Fine-Tune on several featureless areas. This is the least desirable approach: whenever possible, try to build noise profiles using the other methods listed above.
These two images demonstrate several flat featureless areas (marked with white frames) that can be used for building noise profiles.
As you can see, some areas are marked with black frames. It means there are some important detalis in those areas, they are not featureless and cannot be used to build a profile.